Practical takes on venture capital, startup building, and the AI economy — from a 3x founder with 65+ investments.
Jeff Bezos's industrial-AI startup Prometheus raised $12B at a ~$41B valuation led by Bezos, JPMorgan, Goldman Sachs, BlackRock, DST Global and Arch — just seven months after launch. With $18B+ raised against ~150 employees, it's the purest bet yet that the next AI trillion is in atoms, not text.
SpaceX hit a ~$350B valuation in a December 2025 tender offer, up from $210B a year earlier — the most valuable private company on earth. How that number gets set without an IPO, why Starlink (60%+ of $15.5B revenue) drives it, and what a public listing would change.
Claude scored 14.9% on the OSWorld benchmark at launch, ~22%+ now, against a ~72% human baseline. The honest breakdown of what Computer Use can do, what it costs per task, and when to use a real API instead.
o3 posted 87.7% on GPQA, 71.7% on SWE-bench, and a record 87.5% on ARC-AGI — but the ARC run cost ~$3,400 per task. The full benchmark table, the o3 vs o1 gap, and where the scores fall short.
Roughly 40% of every new dollar a16z deploys now goes to AI, from a ~$45B AUM base. The full breakdown of the portfolio by stage, sector, and check size — OpenAI, Mistral, xAI, Cursor, Databricks and more.
46% of all Google searches are local and 88% of mobile local searchers act within 24 hours. The exact playbook to rank a Boca Raton, West Palm Beach, or Miami business in the map pack — what moves the needle, what's a waste of money, and real costs.
Stargate commits up to $500B over four years to build ~10GW of US AI data centers — but only $100B is actually funded. The structure, the four backers, how it compares to the Big Four's $300B in 2025 capex, and where the number could break.
GPT-5 scores ~89% on MMLU and a reported ~74% on SWE-bench Verified — up from GPT-4o's ~85% and ~33% — at roughly $1.25 per million input tokens. The real change: one model that decides how hard to think per query. A full breakdown of what it does, what it costs, and who should switch.
a16z raised $15B+ across vehicles in 2025–2026, pushing total AUM past $80B — the largest war chest any pure venture firm has assembled. The fund-by-fund breakdown, why more than half is pointed at AI, and what the mega-firm model signals for the squeezed middle of venture.
Only 15-20% of recent seed companies raise a Series A within 24 months, down from ~30% for the 2018 cohort. The graduation rate collapsed because seed volume jumped ~80% while Series A capacity stayed flat and the bar tripled to ~$1M-$2M ARR. The full breakdown by cohort, the new Series A bar, and the math behind the cliff.
Micro funds under $100M post a ~2.4x median TVPI and a fat tail of 5x+ outcomes; mega funds over $1B cluster at 1.5–1.8x. But micro funds carry far wider dispersion and mega funds return more absolute dollars per LP. The full breakdown by fund size, DPI, IRR, and loss rate — and a verdict on which size actually wins.
$150 trillion moves across borders every year on rails that still charge 6.3% and settle in days. Stablecoin volume already past $27T a year, take rates collapsing toward 0.3–0.7%, and Stripe paying $1.1B for Bridge. A breakdown of Wise, dLocal, Airwallex, Nium, and the crypto rails capturing the spread.
A $400B SpaceX liquidity event could hand venture investors $50–70B in value, and Founders Fund's seed-era stake alone is plausibly worth $8–12B — a 100–150x return. But most of that has been realized through twice-yearly tenders, not an IPO Musk says may never happen. The full cap-table and exit math.
ChatGPT has 900M+ weekly active users in 2026, ~1.2B monthly, and roughly 35–40M paying subscribers — a ~4% paid conversion. Blended revenue per user is just ~$13/year because 95% of users are free, yet the consumer business runs at ~$12B ARR. The full breakdown of users, ARPU by tier, and the growth trajectory.
Anthropic is valued at roughly $350B in 2026 on a reported ~$9B–$14B annualized revenue run-rate — a 25x–39x multiple and almost a 6x jump from $61.5B a year earlier. A breakdown of the funding history, the cap table, the Amazon and Google stakes, and whether the multiple holds.
Microsoft invested ~$13B in OpenAI and now holds a ~27% economic stake worth about $135B after the October 2025 restructuring. A clause-by-clause breakdown of the deal terms, the rewritten AGI clause, the end of Azure exclusivity, and who got the better end.
Elon Musk's xAI is valued at ~$50B on a reported $100M–$200M revenue run-rate — a 250x–500x multiple roughly 10x richer than OpenAI or Anthropic on ARR. A data-driven breakdown of the funding history, the Grok business model, and whether the Colossus-plus-X distribution bet justifies the price.
OpenAI has ~800M weekly users and ~$20B ARR; Google has $400B+ in revenue, its own AI chips, and Gemini embedded everywhere. A data-driven breakdown of GPT-5 vs Gemini 2.5 benchmarks, API pricing, and market share — and an honest call on who's actually winning.
Anthropic hit a $9B+ annualized revenue run-rate by mid-2026 — about 80% from the Claude API and enterprise contracts, not consumer subscriptions. A full breakdown of every way Anthropic monetizes Claude, the unit economics, and how the model compares to OpenAI's.
The four largest US hyperscalers are guiding to more than $380B in combined 2026 capex, up from ~$246B in 2025. A company-by-company breakdown of who spends what, what the money actually buys, and whether the spending pays off.
Starlink is projected to hit ~$15.5B in 2026 revenue on 7M+ subscribers and ~$65 blended ARPU. The full segment breakdown, the free-cash-flow turn, and what it means for SpaceX's ~$350B valuation and a potential IPO.
A $500M fund needs ~$1.5B in proceeds to return 3x net — too big to be moved by a single seed breakout, too small to lead $100M+ growth rounds against $5B megafunds. The barbell math behind why mid-market VC is losing its edge in 2026, with the return tables and where LP capital is actually going.
Claude Opus 4 launched in May 2025 at 72.5% on SWE-bench Verified — the best coding score of any frontier model at the time, now 80%+ on the 4.5/4.8 line. A full breakdown of the release timeline, the Opus/Sonnet/Haiku tiers, $15/$75 per-million pricing, and an honest head-to-head with GPT-5.
Chime went public on the Nasdaq in mid-2026 under ticker CHYM at roughly an $11B valuation — down 56% from its $25B peak in 2021. A full breakdown of the neobank's $864M revenue, 8.6M members, interchange-driven model, and what the listing signals for the rest of the fintech IPO pipeline.
Stablecoins settled $33T in volume in 2025 and crossed $300B in supply. A breakdown of which businesses actually use USDC and Tether for cross-border payments, contractor payroll, and treasury — plus the real cost vs cards and wires (pennies vs $50 per transaction).
Sequoia Capital manages $85B+ across 20+ active funds in 2026, with a portfolio of 400+ companies anchored by OpenAI, Stripe, and AI infrastructure. A full breakdown by sector, stage, and biggest positions — plus why the open-ended fund structure makes its returns behave differently from every peer.
Anthropic now holds ~32% of the enterprise LLM API market vs OpenAI's ~25%, and captures 40%+ of AI coding spend. But OpenAI still leads total revenue at ~$13B ARR vs ~$5B and 800M+ weekly users. A layer-by-layer breakdown of market share, pricing, and benchmarks — and who's actually winning.
Stripe charges 2.9% + 30¢, Adyen runs interchange-plus from ~0.6% + €0.11, and Braintree sits at 2.59% + 49¢. A side-by-side breakdown of fees, processed volume, and which processor actually wins enterprise payments at scale — Adyen on cost, Stripe on build, Braintree on the wallet.
Le Chat Pro costs $14.99/month — 25% cheaper than ChatGPT Plus and Claude Pro — and runs on Mistral Large 3 with ~1,000 words/sec Flash Answers. Backed by a €1.7B Series C at a ~€11.7B valuation, it's Europe's leading AI lab. Here's where it beats ChatGPT and Claude, and where it doesn't.
The median 2026 seed extension is $1.5M–$3M on a SAFE at flat or a 10–15% step-up, and roughly 38% of seed-funded startups now raise one before Series A. With the seed-to-A gap stretched to 24+ months and graduation rates down to 15–20%, the bridge round stopped being a red flag. Here's how to structure one without poisoning your Series A.
A 2026 Series A is ~$12M at a $45M–$60M post-money before $2M ARR, diluting founders 18–25%. A growth round is $40M–$150M+ at $300M+ against $20M+ ARR, diluting just 8–15%. But AI traction curves and mega-fund check sizes are dissolving the line — here's how to read the economics, not the label.
€450M primary (€1B+ with secondary) at a €10B+ valuation led by General Atlantic — up from €2.4B in December 2025. But this isn't a hype mark: €250M+ revenue, €100M+ EBITDA, a €1.5B backlog, and 70+ SAR satellites already in orbit. Why a profitable Finnish radar company became critical national infrastructure for half a dozen governments.
$8.1B implied IPO valuation on $136.4M H1 2025 revenue (59x trailing) — Cerebras filed its S-1 in September 2024, sat through 13 months of CFIUS review on G42's $335M anchor, and now leads a 2026 AI-chip listing wave. The numbers, comparables vs Nvidia/Groq/SambaNova, and what a clean print unlocks for $25B+ in private AI silicon.
$300B vs $61B valuation, $20B vs $4B ARR, $3/$15 vs $2.50/$10 per 1M tokens — Claude wins coding (72.7% SWE-bench) and safety; GPT wins multi-modal, voice, and the cheapest mini-models. The real procurement math on which one to pick in 2026.
$15.1B IPO market cap at $40/share on NYSE September 2025 — one-third of the 2021 peak of $45.6B. $3.3B revenue, 100M+ users, 575K+ merchants, 0.4-0.5% credit loss rate, and Klarna vs Affirm vs Afterpay side-by-side. The math behind whether the BNPL era is actually back in 2026.
92% Nvidia market share, $30-40K H200 vs $12K AMD MI300X, and Google TPU v5p at 459 TFLOPs — head-to-head pricing, memory bandwidth, cloud rental costs across CoreWeave/Lambda/Crusoe, and the honest call on which chip platform wins training vs inference in 2026.
$5B+ enterprise ARR, 3M+ paid seats, 600K business customers, 92% Fortune 500 penetration, and the gap between $60 list pricing and $45 realized — the contracts, SKUs, and per-seat economics behind OpenAI's enterprise business in 2026.
$40B+ Azure AI run-rate, $13.6B Google Cloud Q1 2026, $85-90B vs $75-80B 2026 capex, and the head-to-head on Copilot vs Gemini, OpenAI vs Anthropic, and net-new enterprise wins. Microsoft is winning the AI cloud war on revenue — but Google is the only credible challenger left.
$9B Cursor at $500M+ ARR and $20/mo, 20M+ Copilot users at $10-39/mo, and Google's $2.4B Windsurf at $15/mo Pro. Head-to-head on pricing, agentic mode, model defaults (Claude 4 vs GPT-5 vs Gemini 2.5), and which AI code editor enterprise devs are actually buying in 2026.
10% on Substack, $9-$199/mo flat on Ghost Pro, 0% on Beehiiv up to 2,500 subs. Real ARPU math at 1K, 10K, and 100K subscribers, the $15-$30 CPM Beehiiv Ad Network, Boosts marketplace pricing, and exactly when Substack's 10% fee is still worth paying for its recommendation network.
$18B valuation at ~120x ARR — Perplexity's $150M run-rate, 22M monthly users, and four revenue lines (Pro $20/mo, Enterprise, ads, Comet browser). Full funding history, side-by-side vs ChatGPT Search and Google AI Overviews, plus the bull/bear case on the Comet bet.
$200/month ChatGPT Pro unlocks unlimited Sora 2 generations at 1080p with 20-second clips and no watermark. Full pricing breakdown, blind-test wins vs Runway Gen-4 (71% photorealism), Kling 2.0 (longest at 120s), and Veo 2, plus the real production workflow for B-roll, social, and concept previz.
$300B secondary tender at 23x ARR — $13B run-rate, 700M weekly ChatGPT users, $40B SoftBank lead. The full breakdown of the cap table, revenue mix, and how OpenAI compares to Anthropic at $61B and xAI at $50B.
1099-DIV not K-1, 20% federal long-term capital gains, and pass-through RIC tax treatment. The complete breakdown of how Robinhood Ventures I distributions, return of capital, and share sales get taxed — plus state tax math and how RVI compares to Forge, EquityZen, and SPV K-1s.
87.7% on GPQA Diamond for o3 vs 78.0% for o1, $10/$40 per 1M tokens vs $15/$60, and 2,727 on Codeforces vs 1,891. o3 wins every benchmark and costs 33% less — but o3-mini high beats full o1 at 14x cheaper. The full comparison plus the migration playbook.
$350B SpaceX private valuation now exceeds Boeing ($130B) and Lockheed Martin ($112B) combined. SpaceX flew 134 Falcon launches in 2024 to ULA's eight, Boeing posted an $11.8B loss, and Lockheed quietly delivered 11.4% operating margins. Full revenue, contracts, and how to actually invest in each.
30%+ NAV premium on RVI vs 0.75% expense ratio on ARKK. RVI gives retail investors OpenAI (~12%), Anthropic (~10%), and xAI (~9%) at a premium. ARKK holds 30–35 public names led by Tesla (~10%) at NAV but has underperformed the S&P 500 by 110+ percentage points over 5 years.
$350B SpaceX valuation, $400–460B secondary market price. Six real paths for retail to buy pre-IPO shares in 2026 — Destiny Tech100 (DXYZ), ARK Venture Fund (ARKVX) at $500 minimum, Forge Global at $100K, Hiive at $25K, EquityZen at $20K, and SPVs with 1–2% mgmt fee + 10–20% carry.
ServiceNow leased ~211,000 sq ft at Stephen Ross's One Flagler tower in West Palm Beach — the largest enterprise software lease in South Florida history. 1,000+ planned hires by 2028, CEO Bill McDermott based locally, and the lease that re-rated WPB's tech market from narrative to operating story.
$9.9B Series C in June 2025 led by Thrive Capital, $300M+ ARR by mid-2025, and the fastest path to $100M ARR in SaaS history. Full funding history, ARR ramp, customer base, and how Cursor compares to GitHub Copilot, Windsurf, and Replit.
No. The Robinhood Ventures I fund's 18 disclosed positions as of June 2026 do not include SpaceX. Full holdings table led by OpenAI (~12%), Anthropic (~10%), and xAI (~9%), why SpaceX restricts retail-facing vehicles from its secondary tenders, and how DXYZ and ARKVX compare for direct SpaceX exposure.
$11.8B in 2025 Starlink revenue, 7M+ active subscribers, and no S-1 filed. Elon Musk's signals, the three conditions that trigger the spin-out, and the $150B–$200B valuation analysts are modeling for 2027–2028.
$75 vs $15 vs $5 per 1M output tokens. SWE-Bench Verified 74.5% vs 65.3% vs 48.1%. Full pricing math, benchmark scores, and a workload-by-workload picking rule for the three Claude 4.x models.
$91.5B Stripe valuation as of the February 2025 secondary tender, on $1.4T in 2024 payment volume and ~$4.7B in 2024 net revenue. Patrick Collison still says no IPO is planned — here's the math, the timeline, and what would force a listing.
$1.4T in 2025 Stripe payment volume vs €1.29T at Adyen vs $1.5T+ through PayPal's Braintree. Side-by-side comparison of take rates, EV/revenue multiples, customer mix, growth — and an explicit winner declared.
$61.5B post-money on a $3.5B Series E closed March 2025, led by Lightspeed's $1B check. Amazon's parallel commitment took it to $8B cumulative; Google added another $1B+. The full investor list, the cap table math, and what the terms imply for the reported $170B+ Series F.
71% of NYC summer weekends get rained on. Friday rains 44% of the time. Sunday dumps the most precipitation. We analyzed every rainy day at Central Park from 1988–2025 to settle the debate.
OpenAI hit a $20B annualized revenue run rate by mid-2026 — about $4B per month — with 700M weekly ChatGPT users and 20M paid subscribers. Revenue grew 5.4x in 18 months. The harder question: $115B in projected cumulative losses through 2029.
22% of Yale's $41.4B endowment, 39% of Harvard's $53.2B, and 25% of Stanford's $36.5B sits in private equity and venture capital — over $35B in PE/VC from just three schools. The Ivy League endowments are the anchor LPs in nearly every top-tier VC fund, and that's the gate every other LP has to walk through.
Lightspeed closed roughly $9B across three vehicles in 2026 — a $4.5B flagship, $3.5B opportunity fund, and $1B Select late-stage fund — pushing firm AUM past $35B. 85% of capital came from re-up LPs, and ~55% is earmarked for AI infrastructure, applications, and vertical software.
5,000+ Tesla Optimus units targeted for 2026 production at a $20K–$30K cost target, vs Figure 02 shipping in the low hundreds at $50K+ with a more mature VLA model and a live BMW deployment. Tesla wins on volume; Figure wins on revenue per unit. Figure AI is now reportedly raising at $40B.
Thrive Capital crossed $25B in AUM in 2026 across 8 funds and a $1B permanent-capital vehicle — with under 30 investment professionals. The OpenAI position alone is now marked above $6B. Inside Josh Kushner's quietly-built dynasty.
Ramp's valuation nearly tripled in a year to $44B on a $750M Series F led by ICONIQ. ARR crossed $1B with positive free cash flow. The pitch has flipped from corporate cards to an AI spend operating system for CFOs whose model bills caught them off guard.
Supabase doubled its valuation to $10.5B in eight months after a $500M Series F led by GIC. The driver: a 600% surge in database deployments, with Claude Code now the platform's single largest contributor and AI agents deploying the majority of new instances.
AMD MI300X benchmarks within 10–20% of the NVIDIA H200 on most AI training tasks and costs roughly half as much per chip. NVIDIA still owns 80%+ of AI training deployments. The reason isn't hardware — it's a 15-year CUDA software moat.
NVIDIA trades at 25–30x forward revenue and 40–45x forward earnings. The $300B+ AI capex supercycle from Microsoft, Google, Meta, and Amazon is real — but so is the concentration risk. Here's whether the math holds.
Hyperscalers dominate AI training. Colocation is eating inference at 40–60% lower cost. Edge handles sub-5ms latency workloads. Here's the full breakdown of where AI workloads actually run by cost, latency, and architecture — and the hybrid model most AI companies have settled on.
An NVIDIA H100 generates 700W of heat. A B200 generates 1,000W. Air cooling fails above 30–40kW per rack — and AI clusters routinely hit 120kW. Liquid cooling is a $50B picks-and-shovels market by 2030, and most investors still haven't noticed.
Northern Virginia hosts 35%+ of North America's internet traffic — but it's run out of power. Phoenix is water-stressed. Here's the real map of where $300B+ in AI infrastructure capex is actually going: Ohio, Iowa, Wyoming, and the emerging second-tier markets reshaping the compute landscape.
Microsoft restarted Three Mile Island for a single corporate buyer. Amazon paid $650M for a nuclear campus. Google contracted with Kairos Power for SMRs. AI data centers need 24/7 carbon-free baseload at 100–500 MW per campus — and nuclear is the only scalable source that can deliver it.
OpenAI is at $300B+, Anthropic at $61.5B, and Google's Gemini is embedded in a $2T public company. The valuation gap isn't about model quality — it's about distribution. Here's who's winning the enterprise, how each is priced, and what the revenue trajectories actually look like.
Grok 3 scores 90.9% on MMLU and 93.3% on MATH — beating GPT-4o and nearly matching Claude 3.7 Sonnet. But benchmarks are a marketing tool. Here's what the xAI Grok 3 numbers actually mean for production use, and exactly when to pick Grok over Claude or GPT.
Most startup employees accept the first equity offer they receive and leave six figures on the table. Here's the exact framework: convert your grant to a percentage, ask the 5 critical questions before signing, and use the 83(b) election and QSBS windows that most employees miss entirely.
A search fund lets an operator raise ~$500K, spend 18–24 months finding a profitable small business to acquire, then raise $5–20M to buy and run it. Historical investor IRR averages 33%+ per IESE data — here's exactly how the model works.
Fortune called the NY–Miami corridor 'the next great American tech hub.' Here's the full breakdown: $3.5–4B in VC, 85 active firms, D-Wave HQ in Boca, ServiceNow 211K SF WPB lease, and where each city fits in the ecosystem.
Miami raised $3.5–4B in VC in 2025. 85 active VC firms, up from 32 in 2020. Here's who to know: Fuel Venture Capital, The Venture City, New World Angels, Palm Beach Capital, Mana Ventures — check sizes, focus areas, and how to get in front of them.
D-Wave Quantum relocated its global HQ from Silicon Valley to Boca Raton's 1.7M sq ft Innovation Campus. FAU is installing a $20M quantum computer. Here's a complete breakdown of Boca's five emerging tech sectors and how the city stacks up against Miami.
Florida has no state income tax (saving NY founders 10.9%+ annually), lower cost of living, a growing VC ecosystem, and direct flights to NYC. Here's the honest breakdown of what to expect: investor access, engineering talent, city tradeoffs, and zip codes worth knowing.
Rolling funds (quarterly raises, $2.5K LP minimums), revenue-based financing (non-dilutive, 1.1–1.5x repayment caps), venture studios (40–80% equity, in-house builds), and evergreen vehicles are each solving gaps that traditional 10-year LP funds structurally cannot fill.
RBF gives you $50K–$5M against future revenue with no equity dilution and a 1.1–1.5x repayment cap. VC gives you larger checks and network access in exchange for 15–25% equity per round. The right answer depends entirely on whether you're building a cash-flow SaaS business or a winner-take-all market bet.
The Limited Partnership Agreement governs every dollar that flows through your fund — management fees, carried interest, LP removal rights, and what happens when a key GP leaves. First-time fund managers consistently underestimate how much is negotiated before the first close, and how much it costs to get it wrong.
Dozens of programs claim to help first-time fund managers raise capital. Only about five actually move the LP needle. Kauffman Fellows, VC Lab, and NVCA Venture Forward are ranked alongside every other major platform by what they actually deliver for your fund raise.
Over 60% of first-time VC fund managers never close their fund — not because the thesis is bad, but because of process mistakes. The biggest: starting outreach without an anchor LP, pitching too broadly, a generic thesis, and mismanaging close sequencing.
The anchor LP commits 15–30% of total fund capital and unlocks every subsequent close. Most emerging managers spend 6–12 months on this one conversation. Here's who anchors Fund I raises, what they negotiate for, and how to structure the relationship without giving away carry.
Placement agents for venture capital funds charge 2–3% of capital raised plus a $5–15K/month retainer — that's $1–1.5M in fees on a $50M fund. Here's when they're worth it, who the top agents are, and how to evaluate them before signing.
Most Fund I closes take 12–24 months and land at $10–50M with 10–20 LPs. The complete checklist: Delaware LP formation, LPA with fund counsel ($50–150K), SEC VC exemption, GP commit (1–3%), anchor LP strategy, and quarterly LP reporting. The legal work is the easy part — relationships are what close the fund.
Top funds use Harmonic for AI-powered sourcing (2–3x more qualified deal flow per analyst-hour), Claude and GPT-4o to draft investment memos in under 30 minutes, and custom signal dashboards to flag portfolio problems 3–6 months before they surface in board meetings.
Running a $20–50M fund in 2026 doesn't require a team of 10. The full stack — Juniper Square for fund admin, Affinity or 4Degrees for CRM, Visible for LP reporting, Harmonic for sourcing — costs $34–81K per year and replaces a $300K+ back-office hire.
SpaceX IPO date is June 12, 2026. Ticker SPCX on Nasdaq at $135/share, $1.77T valuation — the largest IPO in history at 3x the size of Alibaba. Full breakdown: Starlink revenue, Musk's 82% voting control, valuation math vs comps, and the key risks.
Foundation models average 37.5x revenue. Public SaaS median is 3.4x. Here's the full tier-by-tier breakdown of AI vs SaaS multiples in 2026 — why the gap exists, where it's compressing, and how to apply it as an investor or founder.
The VC secondaries market hit ~$152B in 2026. LP stakes trade at 78 cents on the dollar. GP-led transactions are 52% of volume. Distributions as % of NAV hit 6.4% — lowest since 2009. Full 2026 pricing data and what it means for LPs, GPs, and founders.
Top-quartile VC funds source 40–60% of their best deals from proprietary channels. The exact playbook: operator networks (30–40% of best investments), platform content that creates thesis-specific inbound, direct founder sourcing before a raise starts, and the $34–81K annual tech stack that makes it compound.
Affinity ($18–25K/year) dominates for relationship-heavy funds. 4Degrees ($12–18K) wins for lean deal-flow teams. Folk ($4–8K) is the right call for emerging managers under $100M AUM. Here's the complete framework for picking and configuring a VC CRM — including the five mistakes that kill adoption.
The modern VC tech stack runs on Affinity or 4Degrees for CRM ($15–25K/year), Harmonic or PitchBook for deal sourcing ($12–36K/year), and Visible or Juniper Square for portfolio monitoring. Total annual software spend for a $50–100M fund runs $60–120K. Here's exactly what each category costs and where most emerging managers get it wrong.
The secondary market for startup equity processes $130B+ in annual volume. Platforms like Forge, Hiive, and Nasdaq Private Market let employees sell vested shares and early investors exit before IPO — typically at 20–40% discounts to the last round valuation. Here's how it actually works.
The IPO isn't the exit — it's the start of a 12–24 month distribution process. VCs face a 180-day lockup, then choose between in-kind share distributions or cash sales. Here's exactly how the money flows, how it affects fund metrics, and what smart GPs do to maximize LP returns post-IPO.
A traditional IPO raises capital at 3.5–7% underwriter fees and takes 6–18 months. A direct listing (Spotify, Coinbase, Palantir) gives existing shareholders liquidity with no dilution or fee drag. SPACs peaked at 613 deals in 2021 and collapsed 87% by 2023 — here's the data on which path to choose.
The IPO process for startups takes 6–18 months: hire underwriters (3.5–7% fee on proceeds), file a confidential S-1, navigate 2–3 rounds of SEC comments over 45–90 days, run a 10–14 day roadshow to 150–250 institutional investors, price shares the night before, and begin trading — with a 180-day insider lockup.
Most startups use OKRs and KPIs interchangeably and wonder why neither works. KPIs monitor operational health; OKRs drive quarterly direction. Companies with structured OKR systems report 2–3x higher goal attainment. Here's the framework that actually works — with the common mistakes that cause 67% of OKR implementations to collapse mid-quarter.
Most startups shouldn't form a formal board until Series A. The standard composition is 2 founder seats, 2 investor seats, and 1 independent — giving founders majority control. Here's the framework on timing, who to put on it, how to run board meetings, and the three traps that cost founders control.
The standard startup equity compensation structure is 4-year vesting with a 1-year cliff, monthly thereafter. Seed engineers get 0.25–1.5%, executives 0.5–3%. Option pools are 10–20% at seed. Here's how to set it up correctly — with real benchmarks, the ISO/NSO decision, and the common mistakes that blow up at Series A due diligence.
NPS = % Promoters minus % Detractors. Ask customers to rate likelihood to recommend on a 0–10 scale, bucket 9–10 as Promoters and 0–6 as Detractors, then subtract. Median B2B SaaS NPS is 32; above 50 is excellent; above 70 is world-class. Here's the formula, industry benchmarks, and what to actually do with the number.
Most founders don't understand their own cap table until a VC asks them to model dilution in a term sheet meeting. Here's how to track founder shares, option pool, SAFEs, and each funding round correctly — with the columns, dilution math, and the mistakes that blow up at due diligence.
Harvey hit $50M ARR, Abridge is in 50+ health systems, Glean crossed $100M ARR — all vertical AI agents. Horizontal platforms like Microsoft Copilot are struggling to show usage despite mass distribution. The data explains why workflow ownership and proprietary feedback loops make vertical agents structurally superior.
The AI orchestration layer is the most contested piece of enterprise AI infrastructure in 2026. LangChain has 10M+ developer users and CrewAI grew to 100K GitHub stars in a year, but AWS Bedrock Agents and Google Agent Builder are winning enterprise production budgets. Full market map, investment data, and what enterprises are actually deploying.
The best Taco Tuesday NYC deals in 2026: $3 birria at Taqueria El Azteca in Hell's Kitchen, $4 tacos at La Palapa in the East Village, and the Jackson Heights spots that beat every Manhattan special on price. Ranked by actual cost and real deal quality.
NYC happy hours in 2026 still deliver — $5 drafts at Rudy's in Hell's Kitchen, $4 beers at Lucy's in the East Village, 2-for-1 margaritas in FiDi. The deals are real but concentrated: Hell's Kitchen, East Village, Williamsburg, and the Financial District are where bars actually compete on price.
True $1 pizza barely exists in NYC anymore — most quality spots charge $2–4 per slice in 2026. These 12 spots deliver the best price-per-bite in Manhattan, Brooklyn, and beyond, ranked honestly from cheapest to most worth the splurge.
Fewer than 0.1% of VC-backed startups IPO — the winning exit is acquisition. Acquirers in 2026 pay 8–30x ARR, and the difference between those multiples is proprietary data moats, NRR above 120%, and a clean cap table. Here's how to engineer the outcome intentionally.
Tech M&A 2025 crossed $100B in announced deals — Google bought Wiz for $32B, Meta took a 49% stake in Scale AI at a $14.8B implied valuation, and OpenAI spent $7.6B+ acquiring io Products and Statsig. The dominant trend: large-cap tech buying AI capabilities that internal teams cannot build fast enough.
Acqui-hires price talent at $1M–$3M per engineer as an asset purchase. Investors get paid first via liquidation preferences, leaving founders with far less than the headline number. Here's the exact deal structure, what employees receive, and how to negotiate a carveout.
Corp dev teams are internal M&A buyers working for the acquirer — investment bankers are external advisors hired to run a sale process for a 1–3% success fee. For founders, confusing the two means misreading deal signals and leaving 25–40% of acquisition value on the table.
Big tech acquisition pricing is not about DCF. Google paid 30x ARR for Wiz, Meta paid $19B for WhatsApp with no revenue, and Microsoft paid $69B for Activision. Three frameworks — revenue multiple arbitrage, strategic option pricing, and talent/IP cost — explain every major tech deal.
RVI (Robinhood Ventures I) is the only listed way to own SpaceX, OpenAI, and Anthropic — but it trades at a 15–30% NAV premium with ~2.5% annual fees. Here's the honest comparison against buying NVDA, MSFT, or a QQQ-style AI index directly, and when the premium is worth paying.
RVI has traded at a 10–35% premium to NAV since its 2024 NASDAQ listing and has never sustained a discount. Here's the full premium history, what drives the swings, and the exact framework for when buying above NAV is — or isn't — justified.
RVI concentrates ~73% of its NAV in five companies: SpaceX (~22%), OpenAI (~18%), Anthropic (~13%), Stripe (~11%), and Databricks (~9%). Here's the complete holdings list, allocation sizes, and what the concentration means for risk and return.
Defense tech startups raised $2B+ in 2024 — Anduril at $14B, Shield AI at $2.7B, Palantir crossing $50B+ market cap. The DoD is the world's largest buyer and it's finally contracting with startups. Here's who's winning and why.
No — New York does not conform to the federal Section 1202 QSBS exclusion. NY residents owe up to 10.9% state income tax plus 3.876% NYC tax on gains that are 100% excluded federally. Here's the full breakdown on what it costs, who it affects, and what you can (and can't) do about it.
A seed round is $1–3M on a $6–12M pre-money valuation to prove product-market fit. A Series A is $10–15M on a $35–50M pre-money to scale repeatable revenue. Only 35–40% of seed-funded startups raise a Series A — the gap is about evidence, not ambition.
The median Series B post-money valuation in 2025 is $150–200M for B2B SaaS — down 40% from the 2021 peak of $280–400M. AI-native companies close at $300–500M post-money. Check sizes average $30–40M with 18–22% dilution. The floor has held for eight consecutive quarters.
The median Series A check size in 2025 is $10–12M on a $35–45M pre-money valuation — down 25–30% from the 2021 peak. AI-native companies command a 2–3x premium. Founders dilute 18–22% before option pool effects.
There were ~547 tech IPOs in 1999 — the all-time US record, with 74% unprofitable at listing. The 2021 SPAC wave added ~613 blank-check vehicles. The AI era 2024–2026 is far more selective at 50–80 tech IPOs annually, but median ARR at listing has risen to $300–500M. Quality has replaced quantity.
VC-backed companies represent ~40% of all US IPOs but generate roughly 63% of total market cap created at offering. First-day returns average 20–25% vs. 10–15% for non-VC-backed listings — but ~45% of VC-backed IPOs trade below offer price after five years.
Approximately 74% of the ~308 tech companies that went public in 1999 were unprofitable at the time of their IPO — the highest concentration of money-losing listings in US history. By 2004, roughly 50% had gone bankrupt, been delisted, or acquired at distressed prices as the NASDAQ fell 78% from its peak.
Complete data on market cap at IPO for every major tech listing from 2020 to 2025. 2022 was nearly frozen — Mobileye at $17B was the only notable listing. 2021 peaked with Coinbase at $85.8B and Rivian at $66.5B. Full dataset and analysis inside.
Approximately 308 technology companies went public in 1999 — the peak of the dot-com bubble and the highest single-year count in US history. The average first-day pop was 73%, VA Linux surged 698%, and 75% of issuers were pre-revenue. Here is the complete data on what happened, why, and what came after.
Public SaaS EV/revenue multiples peaked at 14–17x NTM in November 2021, with hypergrowth companies hitting 30–100x. By end of 2022, medians collapsed to 5–6x after the fastest Fed hiking cycle in 40 years. Here is the complete data on what drove the bubble and where multiples sit today.
Peak XV (Sequoia India), Accel India, Blume Ventures, Nexus Venture Partners and more. India deployed $8B+ in VC in 2025 across 1,000+ deals. Full ranking of the 15 most active Indian VC firms with fund sizes, notable exits, and investment focus.
OpenAI: $5B+ ARR. Anthropic: $2B+ ARR. xAI: $1B+ ARR. Mistral: ~$100M. Cohere: ~$100M. Full revenue rankings of the top private AI companies in 2026 — with revenue multiples, growth rates, and which ones are actually profitable.
Affinity leads on relationship intelligence breadth and firm adoption — 3,000+ VC firms use it. 4Degrees wins on AI-powered warm intro mapping and deal sourcing accuracy. Affinity starts at ~$3,000/user/year; 4Degrees is $2,000–$2,500/user/year. Full feature-by-feature breakdown.
The median time from seed to Series A is 18–24 months in 2026. In 2021 it was 12–14 months. You need $1–2M ARR, 150%+ YoY growth, and 18 months of runway to be competitive. Full timeline data by stage, sector, and fund size.
Notion wins for most startups under 50 people at $8–$16/member/month with flexible docs and databases. Confluence leads for engineering teams already on Jira at $5.75/user/month. Coda is the pick for teams that need doc-plus-app automation — viewers are free, doc makers pay $10/month.
Gusto leads on payroll + benefits simplicity for early-stage startups at $40/month base. Rippling wins on platform depth. Justworks ($59/person/month) is the best PEO for startups that want to avoid HR admin entirely. Full ranked breakdown of 6 options for startups under 50 people.
The best financial modeling tool for startups is Google Sheets until $1M ARR, Runway or Causal for Series A–B companies building integrated forecasts, and Mosaic for Series C+ companies needing board-ready reporting. Here is how 7 tools compare on features, pricing, and fit by stage.
Meta generated $164.5B in revenue in 2025 — but nearly all of it comes from AI-enhanced advertising, not AI products. Meta AI has 700M+ monthly users and zero direct revenue. Advantage+ and Reels recommendations are the real AI story, and direct AI monetization is a 2028 event.
Nvidia's FY2025 revenue hit $130.5B — Data Center at $115.2B (88%), Gaming at $11.4B (9%), Automotive at $1.7B (+55% YoY). Q1 FY2026 reached $44.1B with Data Center at $39.1B. The company has transformed from a gaming GPU business into an AI infrastructure monopoly.
Azure grew 35%, AWS hit 24%, Meta posted $47.6B in revenue, and the hyperscalers committed $300B+ in combined AI capex for 2026. The Q1 2026 earnings season confirmed the AI investment cycle is converting to real revenue — faster than bears expected.
Median public SaaS sits at 8–10x NTM revenue. But Snowflake, Cloudflare, and Datadog command 15–25x. The separators are NRR above 130%, 30%+ growth, and a Rule of 40 score above 50 — not narrative, not category buzz.
Median public SaaS trades at 8–10x NTM revenue in 2026 — down from 20x+ at the 2021 peak. High-growth companies with strong NRR still command 15–25x. The correction is complete, but 2021 is not coming back.
SaaS valuation multiples in 2026 scale dramatically with ARR: sub-$1M ARR gets 3–5x NTM revenue, $1–10M ARR gets 5–8x, $10–50M ARR trades at 6–12x (top-quartile compounders reaching 15–22x), and $50M+ ARR commands 8–20x. The spread between median and elite widens at every band — NRR, growth rate, and burn multiple compound the multiple, not just scale.
The biggest IPOs of 2026 include Klarna (targeting $15–20B, S-1 filed), Chime (~$25B, expected Q3), CoreWeave (already priced at $23B in March), Databricks ($62B last private round), and Cerebras ($8.7B, CFIUS-blocked). Fewer than half of announced filers will list before year-end — here is who makes it and who doesn't.
Klarna, Chime, Cerebras, Discord, and 20+ more are in the 2026 IPO pipeline. Klarna is targeting a $15B valuation (down from $46B peak), Chime ~$25B, Cerebras still blocked by CFIUS. Fewer than half will likely list before year-end.
Most enterprise AI projects fail to show positive ROI in year one — not because the technology fails, but because companies apply the wrong measurement framework. CFOs who get it right separate cost displacement (TEI), infrastructure (NPV/payback), and copilot productivity (multiplier models) into three distinct analyses. McKinsey data shows AI leaders generate 3–5x better returns than laggards with identical technology.
In 1999, approximately 280 technology companies went public in the US — the peak of the dot-com bubble, with average first-day returns exceeding 70%. See the complete year-by-year count from 1980 to 2025, including the 2021 SPAC boom (1,010+ total listings) and the 2022 crash (71 traditional IPOs total).
Stanford SWE-bench showed multi-agent AI achieving 87% accuracy vs 32% for single agents on complex tasks. The orchestration layer — not the underlying models — is where enterprise value is accumulating. Here is what multi-agent systems are, which frameworks dominate, and why vertical agent companies are commanding 40-80x ARR multiples.
Klarna's AI agents handled the equivalent of 700 human agents in their first month. Salesforce Agentforce closed 5,000+ enterprise deals in six months. Agentic AI in enterprise is no longer about task automation — it's full workflow ownership, and the valuation gap between the two is 40–80x ARR vs. 10–15x.
AI agent startups raised over $3B in 2024–2025. Cognition (Devin) hit $2B, Sierra $4.5B, Harvey $1.5B. Gartner projects the agentic AI market at $47B+ by 2026. Here is who is winning, why vertical agents command premium valuations, and what the investment thesis looks like.
Venture debt is typically 25–35% of your last equity round at 11–13% interest plus warrants. Revenue-based financing runs 6–12% flat with zero dilution. Here is when startup debt financing extends runway without giving up equity — and when it backfires.
SAFEs have no interest, no maturity date, and no debt — convertible notes do. Over 80% of YC-backed seed rounds use SAFEs. Here's when each instrument is right and what founders miss about how complexity compounds before a Series A.
RBF costs 6–12% flat with zero dilution. Equity at seed gives up 15–25% permanently. The right answer depends on your MRR, CAC payback cycle, and whether strategic relationships are worth more than ownership.
Venture debt for startups is a non-dilutive term loan at 8–14% interest with warrants, used to extend runway without giving up equity. But covenants, cash sweeps, and MAC clauses make it a tool that blows up companies as often as it saves them.
Amazon committed over $100B in capex for 2025 — the most of any hyperscaler. The majority funds Trainium2 custom chip clusters, new AWS regions, and the infrastructure powering Bedrock, SageMaker, and Amazon Q. Here's where every dollar goes and what it means for enterprise AI buyers.
Meta raised its 2025 AI capex guidance to $64–72B, up 67% from $38.4B in 2024. The money funds NVIDIA GPU clusters, US data centers, and Llama training compute. Here's where every dollar goes and why Meta's bet is structurally different from Microsoft and Google.
Google committed $75B in capex for 2025 — a 43% jump from $52.5B in 2024. The majority funds custom TPU v6 chips, new hyperscale data center campuses, and the compute stack powering Gemini across Search, Cloud, and Workspace. Here's where every dollar is going and whether the math is working.
Microsoft committed ~$80B in AI capex for FY2025 — the largest single-year infrastructure bet in corporate history. Over 50% goes to US data centers. Here's the full breakdown by category, geography, and what Azure AI revenue growth actually justifies it.
The AI model release cycle now averages 2-4 months at major labs — far shorter than enterprise procurement cycles of 4-9 months. OpenAI has deprecated 7+ models since 2022. Here's what enterprise CTOs are actually doing: model abstraction layers, multi-provider routing, and contractual lifecycle SLAs.
Gemini 1.5 Pro offers 2M tokens, Claude 200K, GPT-4o 128K. Processing 1M tokens costs $5–$15 per request vs under $0.05 for equivalent RAG retrieval. Whether to use long context or RAG is a cost architecture decision — here's the framework.
OpenAI and Anthropic dominate the headlines, but Mistral and Cohere are winning the enterprise contracts that matter — in European regulated industries, large-scale RAG deployments, and on-prem environments. Here's how the mid-tier AI model market is actually splitting.
A down round is when a startup raises capital at a lower valuation than its previous round — roughly 25–30% of late-stage deals in 2023–2025 were priced down or flat. Here's what actually happens to founders, employees, and investors when anti-dilution provisions fire, and how companies like Klarna and Stripe survived.
Meta Llama 4 launched in April 2025 with Scout (10M token context window), Maverick (scored 1417 on LM Arena, beating GPT-4o at launch), and Behemoth (~2T params, still training). When open-weight models reach frontier quality, it changes who controls the AI stack — and who doesn't.
Gemini 2.5 Pro scores 97% on MATH vs GPT-4o's 76%, and supports 1M token context vs 128K. GPT-4o still wins on real-time multimodal (live audio/video) and coding (HumanEval ~90% vs ~84%). The decision framework for enterprise AI buyers choosing between the two in 2026.
Claude Sonnet 4.6 scores 72.7% on SWE-bench Verified, 83% on GPQA Diamond, and runs at $3/M input tokens vs $10/M for GPT-5 o3. What the Anthropic Claude 4 benchmarks actually mean for enterprise buyers — and when paying a 70% premium for GPT-5 makes sense.
o3 scores 87.5% on ARC-AGI and ~88% on PhD-level science (GPQA Diamond). o4-mini delivers 80–90% of that performance at $1.10/M input tokens vs $10/M for o3. Here's what the reasoning model shift means for every team building on AI — and why the real cost is hidden in reasoning tokens.
Claude Sonnet 4 leads on coding and safety compliance, GPT-5 wins on ecosystem and multimodal breadth, and Gemini 2.5 Pro dominates on 1M-token context windows. The honest breakdown by use case, API cost (~$3/M vs $10/M input tokens), and enterprise readiness — plus the decision framework to stop overthinking it.
A lead investor sets the terms, writes the largest check (typically 30–60% of the round), and makes every follow-on investor willing to commit. At Series A, leads target 15–25% ownership on $40–60M pre-money valuations, take a board seat, and negotiate pro-rata rights for future rounds. Without a confirmed lead, most rounds don't close.
Net Revenue Retention (NRR) measures how much recurring revenue you keep and grow from existing customers. Best-in-class SaaS hits 120%+ NRR (Snowflake peaked at 158%, Datadog runs ~118%); the median B2B SaaS is ~106%. Above 120% NRR, your EV/revenue multiple is typically 2–3x higher than peers — it's the single most important SaaS valuation driver.
Burn rate is how fast your startup consumes cash. Net burn = monthly expenses minus monthly revenue. At $200K/month net burn with $2M in the bank you have 10 months of runway — and 18 months post-raise is the minimum investors want to see before they fund you again.
MRR (Monthly Recurring Revenue) is the normalized monthly value of all active subscriptions. ARR = MRR × 12. At $1M ARR, median SaaS companies trade at 4–6x revenue; hypergrowth companies with 120%+ NRR trade at 15–25x. Get these definitions wrong in a fundraise and you lose credibility before you get to the deck.
A SAFE (Simple Agreement for Future Equity) is not debt — it's a warrant-like right to receive preferred equity at the next priced round. Invented by YC in 2013 and updated to post-money in 2018, SAFEs have no interest rate, no maturity date, and close in days. Understanding the cap, discount, and pre- vs post-money difference is what separates founders who know what they own from those who find out at closing.
A cap table records every piece of equity in a company — founder common stock, investor preferred shares, employee options, SAFEs, and convertible notes. At founding it fits on one line. By Series A it has a 10–20% option pool, multiple preferred series with liquidation preferences, and determines who gets paid what at exit.
Carried interest is the GP's share of fund profits — typically 20% of returns above an 8% hurdle rate. On a $100M fund returning $300M, the GP earns ~$40M in carry, taxed at long-term capital gains rates (~23.8%). This incentive structure explains why VCs take concentrated, high-risk bets over diversified portfolios.
A term sheet is the non-binding document that sets every economic and governance term of a VC investment. Most founders overfocus on valuation — but liquidation preference (standard is 1x non-participating), board composition, and the option pool shuffle are the clauses that determine actual outcomes at exit.
A unicorn startup is a privately held company valued at $1 billion or more — a term coined by Aileen Lee in 2013 when only 39 companies qualified. As of 2026, roughly 1,200 unicorns exist globally, but fewer than 20% ever exit at or above their peak private valuation.
Venture capital is the engine behind Google, Amazon, and Uber — and almost every transformative tech company of the last 40 years. Top-quartile VC funds return 3x+ TVPI and 20%+ net IRR, but the median fund returns just 1.5–1.8x. Here's how the LP/GP structure works, what VCs look for, and what the return data actually shows.
Early exercise means buying your options before they vest — at today's low 409A strike price — then filing an 83(b) election within 30 days. At seed stage this typically costs $0 in ordinary income tax and starts your capital gains clock immediately, potentially saving employees $500K–$5M+ on a successful exit.
A 409A valuation is the IRS-required appraisal that sets your common stock fair market value before you can legally issue employee stock options. It costs $1,500–$5,000 at seed stage and $5,000–$15,000+ at Series B+. Without one, your employees face an immediate, non-discretionary tax bill on unvested options they can't sell.
RSUs vest automatically and are taxed as ordinary income on the full fair market value. Stock options require an exercise payment — ISOs offer long-term capital gains treatment if held correctly, while NSOs are taxed as ordinary income at exercise. Early employees benefit most from ISO options at low 409A valuations; late-stage and public-company employees should default to RSUs for simplicity.
Standard startup options: 4-year vesting with a 1-year cliff, strike price set by 409A valuation at grant. ISOs save you taxes vs NSOs. The biggest trap: the 90-day exercise window after leaving — most employees can't afford to exercise and forfeit their equity.
Linear wins for most startup engineering teams at $8/seat/month — faster, cleaner UX, built for async-first teams from Seed through Series B. Jira is right at 200+ engineers or deep Atlassian integration. Shortcut at $8.50/user/month is the best middle ground for Series A teams.
Gusto wins for small US-only teams under 50 employees at $46/month base. Rippling is the right call at Series A when you need IT + HR + payroll in one platform. Deel is purpose-built for international hiring — EOR in 150+ countries starts at $599/month per employee.
Stripe is the right default at $500K+ ARR if you have engineering resources and want full control at 2.9% + $0.30 per transaction. Paddle is the best merchant-of-record for global SaaS scaling to $1M–$10M ARR, handling all VAT/GST automatically at ~5% + $0.50. Lemon Squeezy (now Stripe-owned) is the simplest path for early-stage indie SaaS under $250K ARR.
Ramp wins for most startups with 1.5% flat cashback on all spend and free expense management — zero fees, used by 25,000+ companies. Brex leads for VC-backed Series A+ companies that need high credit limits tied to fundraising history. Airbase (now part of Maxio) is the top pick for mid-market teams needing integrated AP automation, POs, and expense management in one platform.
Mercury wins for early-stage startups with zero fees and up to $5M FDIC coverage via sweep. Brex leads for Series A+ companies needing integrated spend management. Arc is best for yield on idle runway — a $3M balance earns ~$147K/year at 4.9% APY. Relay wins for team-based spending controls with up to 20 sub-accounts.
Carta dominates at Series A+ with 40,000+ companies on platform (~$2,500/year base). Pulley wins for seed-stage cost efficiency with a free tier under 25 stakeholders and ~$1,000 409As. AngelList Stack is free and the clear winner if you're running SPVs alongside your cap table.
Africa startup funding hit ~$2.9B in 2024, down from a 2022 peak of $6.5B. Nigeria leads at ~30% of deals, Kenya at ~20%, South Africa at ~15%, Egypt at ~12%. Fintech captures 40%+ of capital, and the investors still writing checks are the ones who understood the fundamentals from day one.
The MENA startup ecosystem raised ~$4.5B in 2025, with UAE (45%) and Saudi Arabia (30%) leading. Fintech dominates at 30%+ of deals. Tabby, Tamara, and Kitopi prove the Gulf can produce unicorns — and Vision 2030 is deploying sovereign capital to accelerate what private markets alone cannot fund.
LatAm VC funding hit ~$5.5B in 2025, recovering from the 2021 peak of $15.7B. Brazil leads at 40%, fintech dominates at 30-35% of deals, and AI-native health and SaaS startups are driving the region's next wave of growth.
SEA VC funding peaked at $25B in 2021, bottomed at $7.5B in 2023, and is recovering to ~$10B. Indonesia, Singapore, and Vietnam drive 80%+ of deals. Peak XV, GIC, Vertex Ventures, and Jungle Ventures are leading the cycle — fintech and AI-native software are where the money is going.
India's startup ecosystem ranks 3rd globally with 108 unicorns and ~$12B in 2025 VC funding, up from the $7B trough in 2023. Peak XV, Accel India, and Elevation Capital lead deal flow. Fintech, B2B SaaS, and AI-native startups are defining the current cycle.
GP commit in VC typically ranges from 1–3% of fund size — a $200M fund GP putting in $2–6M of their own capital. Here's why LPs treat GP commit as the single most credible alignment signal in manager selection, how emerging managers actually fund it, and what happens when you get it wrong.
The lead investor sets price and terms — committing 30–60% of the round — while follow-on investors co-invest at those same terms without renegotiating. Here's how VC syndication actually works, what rights each investor gets, and what founders should demand from their lead before signing.
At VC fund end of life (year 10–12), GPs choose between secondary NAV sales at 60–90 cents on the dollar, GP-led continuation vehicles ($72B in 2024 volume), fund extensions, or write-offs. Here's how the distribution waterfall actually works and what LPs should demand before the clock runs out.
GP-led secondaries hit $72B in 2024, with continuation vehicles accounting for 60–70% of all secondary volume. Here's how CVs work, how they're priced at 90–100 cents on NAV, and the carry reset detail most LPs miss when deciding whether to roll in or take cash.
AngelList's rolling fund model lets GPs raise capital quarterly from LPs committing $25K–$50K per quarter, generating $2B+ in commitments within two years of launch. Here's how the quarterly subscription structure works, who it's best for, and the tradeoffs most GPs underestimate.
Operator angels — ex-founders and senior operators investing personal capital — write $25K–$250K checks in 24–72 hours and outperform traditional VCs by 15–20% on early-stage IRR. Here's why they've become the most coveted check at pre-seed and seed, and what founders should know before taking one.
a16z, General Catalyst, and Lightspeed run the most structured VC internship programs, paying $9K–$12K/month for MBA summer associates. Fewer than 200 spots exist across the entire US industry — and most are filled through warm intros before the role is ever posted.
Less than 30% of working VCs at top funds built a company first. Banking, consulting, big tech product, and content are all viable paths in — what actually matters is sourcing ability, analytical rigor, and a defensible sector thesis. Here's the full playbook.
VC principal salary in 2026 ranges from $130K at emerging managers to $300K base at top-tier funds like a16z and Sequoia. Carry allocation is 0.25–1.0 points per fund — real money at large funds, but it vests over 10 years and most principals never make partner.
VC associate salary in 2026 ranges from $90K–$200K+ base depending on firm tier. Tier 1 funds (a16z, Sequoia, General Catalyst) pay $160K–$200K base with $40K–$80K bonuses. Emerging managers under $150M AUM pay $85K–$130K. Carry is the real upside — but most associates wait 7–10 years to see distributions.
VC analyst salary in 2026 ranges from $85K–$160K base depending on firm tier. Tier 1 firms pay $130K–$160K base with $25K–$40K bonuses. Carry at the analyst level is rare and almost never meaningful — here is what the data actually shows.
Y Combinator has backed 4,000+ companies since 2005, producing 90+ unicorns and $600B+ in combined portfolio value. The top 10 exits — Airbnb, Stripe, DoorDash, Coinbase — account for an estimated 65% of all returns. Here is the full exit data, broken down by batch era.
Y Combinator leads for investor access ($500K for 7%, ~40% Series A close rate), a16z Speedrun for operator mentorship, and NVIDIA Inception for compute credits with no equity taken. Here are 8 programs ranked by what AI founders actually need in 2026.
YC invests $500K for 7% — but the check is the least valuable thing you get. The real value is the alumni network of 80,000+ founders, Demo Day access to 1,000+ investors, and a ~40% Series A close rate within 12 months of Demo Day. Here is the honest breakdown.
YC accepts roughly 1.5% of 12,000–15,000 applicants per batch. Most applications fail not because the idea is bad, but because founders cannot clearly articulate what they are building, who it is for, and why they are the ones to build it. Here is the honest playbook.
YC offers $500K for 7% equity and the most powerful startup network on earth. Techstars runs 90+ vertical and geographic programs at $120K for 6%. 500 Global backs international founders targeting emerging markets. Here is the honest breakdown of how to choose.
Y Combinator accepts approximately 1.5–2% of applicants per batch — roughly 200–250 companies out of 12,000–15,000 applications. The acceptance rate has been falling for a decade as YC's brand compounds. Here is the full funnel breakdown, who actually gets in, and what moves the needle in your application.
The board composition clause is the most consequential paragraph in a term sheet. Founders lose effective board majority at Series A in ~60% of deals — not because they gave up votes explicitly, but because the independent seat ends up investor-aligned in practice. Here is what the standard clause says and what you can actually negotiate.
Redemption rights give preferred investors the contractual right to demand a share buyback — typically at 1x liquidation preference — after 5–7 years. About 10–15% of US VC term sheets include them, rising to 20–30% in corporate venture and cross-border deals. Here is how they work, when they trigger, and how founders can negotiate them out.
An MFN provision in a SAFE gives early investors the right to match any better terms you later offer. About 35–40% of pre-seed angel rounds include MFN provisions, and the dilution shift at Series A can reach 15–40% depending on how much your caps compress between rounds.
A pay-to-play provision requires existing investors to participate pro-rata in a down round or have their preferred shares converted to common stock. Used aggressively in 2001–2003, 2009, and the 2022–2023 correction, it is the most powerful tool for cleaning up a distressed cap table.
A bridge round is short-term financing — typically $500K–$3M via SAFE or convertible note — that extends runway to a defined milestone. About 35–40% of venture-backed startups raise at least one bridge between seed and Series A. Here is when it makes sense, how to structure it, and what it signals to your next lead investor.
Information rights clauses give Major Investors contractual access to monthly financials, cap tables, and books and records. Most founders sign them at Series A without reading the details — here is what the standard package includes, who qualifies, and what you can actually negotiate.
Drag-along rights let a majority of shareholders force all others to approve and participate in a sale. Most founders sign them at Series A and forget they exist — until a distressed sale triggers the clause. Here's how they work, when they bite, and what you can actually negotiate.
The option pool shuffle is a standard VC term sheet tactic that quietly increases investor ownership by 5–8 percentage points. Here's exactly how the math works, a worked example showing the per-share price impact, and how to negotiate pool size and structure before you sign.
Anti-dilution protection adjusts an investor's conversion price when a company raises at a lower valuation. Broad-based weighted average is the 2025–2026 market standard — here is how each structure works, a worked example showing the dilution math, and what founders miss when negotiating these terms.
Pro-rata rights give investors the contractual right to maintain their ownership percentage in future funding rounds. At Series A+, major-investor pro-rata is market standard — here is the mechanics, the negotiation dynamics, and the math every founder and fund manager should know.
SaaS companies are valued using EV/NTM Revenue multiples — 6–8x for median public SaaS in 2025–2026, 12–18x for Rule-of-40 outperformers. Private SaaS carries a 30–50% discount to public comps. Here is the complete framework investors use from Series A to IPO.
A liquidation preference gives investors the right to recover their capital before founders see a dollar at exit. Non-participating 1x is the 2025–2026 market standard — here is what every other structure costs founders and how to negotiate it.
Profitable SaaS companies trade at 20–35x EV/EBITDA in 2025, while high-growth SaaS above 30% YoY commands 50–80x. PE buyouts target 15–22x at close. Here is how saas ebitda multiples work, what drives the spread, and when EV/EBITDA replaces EV/Revenue as the primary valuation lens.
The median public SaaS company trades at 4–5x NTM revenue as of mid-2025, down from 18–20x at the 2021 peak. High-growth SaaS above 40% YoY still commands 8–12x. Here is the full benchmark breakdown by growth tier, what drives the spread between 3x and 15x, and what it means for private company valuations.
Affinity is the best deal flow tool for VC firms that need automated relationship-driven sourcing and pipeline tracking. Grata wins for proprietary middle-market deal origination before companies hit intermediaries. For emerging managers, Visible provides solid pipeline management combined with LP reporting at a fraction of enterprise costs.
Mercury is the best bank account for startups in 2026 — free, API-native, and built for founders with FDIC coverage up to $5M. Brex wins for funded teams needing corporate cards and expense management. Arc leads for earning 4.5–5.2% APY on idle cash. Here is how all five options compare by stage and need.
Visible.vc is the best fundraising CRM for founders — purpose-built with investor pipeline stages and update automation at $49/month. Affinity is best for relationship-graph fundraising at Series A and beyond. Attio is the top free-tier option for pre-seed and seed rounds.
Granola is the best AI meeting notetaker for investors — Mac-native, no meeting bot, enriches your own notes with a full AI summary at $18/month. Fathom is the best free option. Here is how all 6 tools compare by workflow fit, privacy, and what VCs and founders actually use.
Clerky is the best legal platform for VC-backed startups — used by 10,000+ companies and co-developed with YC for SAFE notes and Series A docs from $399. Stripe Atlas ($500) is fastest for Delaware C-Corp. Here is how all 6 options compare by stage, cost, and what investors actually require.
Visible.vc leads as the best dedicated investor update software for most startups with metric integrations and open tracking at $49/month. Cabal is the strongest free option. Here is how all eight options compare by stage, cost, and what investors actually want.
There are hundreds of venture capital podcasts. Most are not worth your time. Here are the 10 that consistently deliver signal — ranked by audience fit, signal quality, and what you will actually learn from each one.
Y Combinator is the best startup accelerator for B2B SaaS — $500K for 7% equity with the strongest alumni network of any program. Alchemist is the only accelerator built exclusively for enterprise, with corporate partner intros at Salesforce, SAP, and Cisco. Full ranked breakdown of all 7 options by outcomes, dilution, and cohort fit.
Tomasz Tunguz leads on data-driven metrics, Jamin Ball's Clouded Judgement is essential for SaaS multiples, and Not Boring produces the deepest company analyses. Full ranked breakdown of the 12 best VC newsletters by signal-to-noise ratio — and which stack fits your role.
QuickBooks Online leads on CPA compatibility, Wave is the best free option for pre-revenue startups, and Pilot is the go-to for VC-backed companies needing GAAP-compliant books without hiring a controller. Here is how all 7 compare by stage and budget.
TVPI looks great on paper but includes unrealized value that may never materialize. DPI — Distributions to Paid-In — is the only metric that tells you if a VC fund actually returned cash to LPs.
The 2019–2020 vintage classes are showing 2.2–2.8x TVPI at the top quartile. The 2021 vintage is underwater for most managers. Here is what the data actually shows.
Current EV/Revenue multiples for public SaaS companies by growth rate bucket. Top-quartile SaaS at 20%+ growth is trading at 8–12x NTM revenue — down sharply from 2021 peaks of 30–40x.
Google click-through rates are falling even at position #1. AI engines now synthesize answers directly — and citations are the new rankings. The complete playbook for winning both channels simultaneously.
Claude and Cursor are the two AI tools every founder should be using daily. Perplexity beats Google for competitive research. Clay is the best AI-powered GTM tool built for the AI era. Full honest rankings with real pricing and what each tool actually replaces.
Carta dominates Series A and beyond with 409A valuations and compliance at $149–599/month. Pulley is free for up to 25 stakeholders and wins on UX for seed-stage startups. AngelList Stack is the right call if you raised your round on AngelList's platform.
DocSend leads for fundraising with page-by-page investor analytics — $45–150/month. Visible.vc wins for ongoing investor relations. Carta's data room is the right call if you're already managing your cap table there. Full breakdown of pricing, features, and fit for pre-seed through Series B.
Pitch.com leads the 2026 rankings for VC-facing investor decks with purpose-built templates and link analytics. Beautiful.ai wins for non-designers who need professional output fast. Canva is the best free option for budget-conscious founders who need design flexibility.
Affinity leads on relationship intelligence and is used by hundreds of institutional VC funds. 4Degrees wins on AI-powered warm intro mapping. Folk is the best option for solo GPs and emerging managers who need structured deal tracking without a five-figure CRM contract.
The KPIs that separate proactive portfolio management from passive check-writing. Top VC funds track 8-12 metrics per company monthly — ARR growth, burn multiple, runway, and NPS at the company level; TVPI, DPI, IRR, and reserve ratios at the fund level.
Most investor updates get skimmed and forgotten. The format that actually generates intros, hires, and bridge checks is monthly, under 400 words, five sections, and ends with one specific ask. Here's the exact investor update template top founders use.
Most startup boards are assembled by accident — whoever led your last round gets a seat. A great board member makes 2-3 high-conviction intros per year, tells you what you don't want to hear, and has seen 10+ companies at your exact stage. Independent directors at 0.1–0.25% equity are often more valuable than your VC board seats.
Most board meetings are information dumps that leave everyone frustrated. A good startup board meeting agenda runs 60–90 minutes in four sections — and sends the deck 48 hours before so the meeting itself becomes a real strategic working session.
Most VCs promise value-add. Top-quartile funds deliver structured talent networks, customer intro pipelines, and operational playbooks that compound into measurable return differences. Here is what VC portfolio acceleration actually looks like — and how founders should diligence it.
Singapore, Canada, France, and the UAE are running the most aggressive government startup policies in history — startup visas, co-investment funds, regulatory sandboxes, and stock option reform. The US has no dedicated startup visa and is losing ground.
Immigrants founded 55%+ of America's billion-dollar startups. From Google to Stripe to Zoom, the data is clear — diaspora founders outperform. Here's why, and how VCs can systematically find them.
MENA startups raised $3.2B in 2025 across 600+ deals. UAE leads deal volume, Saudi Arabia drives the largest rounds. Here's where the money is going — and which government-mandated bets are worth making.
European VC raised €19B in 2023 vs US $170B+. US top-quartile funds return ~20% net IRR vs Europe's ~13%. Here's what actually drives the gap — and where European VC is closing it.
Independent research providers like GLG, AlphaSights, Tegus, and Guidepoint charge $500–$2,000 per expert hour. Large hedge funds run 50–100+ expert sessions monthly. Here is when they work, when the economics break down for smaller funds, and how to build a proprietary research network that outperforms paid platforms.
Alternative data — web traffic, job postings, app store rankings, and credit card transactions — is how top funds like Coatue and a16z evaluate startups before financials exist. Here is every major data source, what it reveals, and what founders should know about being monitored before the first meeting.
GLG, AlphaSights, Tegus, and Guidepoint charge $500–$2,000/hour for access to former executives and domain practitioners. Hedge funds and PE firms run dozens of sessions a month; most VC funds under $200M can get better ROI from a Tegus subscription or building a proprietary expert network instead.
VCs with strong personal brands see 30–40% of deal flow from inbound and raise LP capital 6–9 months faster. Here is the platform-by-platform breakdown — Twitter/X for founder reach, LinkedIn for LP credibility, newsletters for owned distribution — and the content types that actually drive results versus noise.
Most VC scouts earn no salary — compensation is 10–20% of carry on deals they source, illiquid and winner-take-most. A scout who refers a $250K investment returning $5M earns roughly $142K in carry. Here is the full math and which programs are worth joining.
Partner base comp at VC firms runs $250K–$500K+ depending on AUM, but salary is not the negotiation. Carry allocation (5–20% of the fund pool), vesting schedule, GP title, and co-investment rights are where the real value is — and most candidates accept the first offer without touching any of them.
The VC career ladder runs Analyst → Associate → Principal → Partner, with sharply different compensation at each rung. Associates earn $130–180K base plus small carry; Partners earn $250–500K+ plus carry worth millions on a successful fund. Fewer than 20% of Associates reach Partner at the same firm.
There are roughly 800–1,200 new analyst and associate roles in US venture capital each year, against tens of thousands of applicants. The path in is almost always through differentiated dealflow, a relevant operating track record, or a warm introduction from someone already inside.
The three dominant LP types in venture capital allocate very differently: endowments target 15–25% to alternatives (Yale model), pension funds cap at 5–15% due to liability matching, and family offices run 20–30%+ with full discretion. Here's what each type needs from a fund manager.
Institutional LPs fund fewer than 5% of manager pitches they receive. Endowments, pensions, and foundations evaluate on 6 factors: verifiable track record (2x+ TVPI), differentiated sourcing, portfolio construction discipline, team stability, fund economics, and LP-friendly governance. Here's what actually gets through.
Family office tech investing doesn't require standing up your own vehicle. LP positions, co-investments, SPVs, secondaries, direct deals, fund-of-funds, and evergreen vehicles each offer different fee, control, and liquidity profiles. Here's how sophisticated family offices stack them.
A VC scout program gives operators, founders, and angels 5–15% carry in exchange for sourcing deals. 60–70% of top-quartile US venture funds have one. Here is how the programs work, who runs them, and how to get a scout role.
Family offices manage $5.9T+ globally and 35–40% now make direct startup investments. They operate nothing like VCs — longer time horizons, relationship-first diligence, and no deployment pressure. Here's what founders need to know to raise from them effectively.
A zombie startup is a VC-backed company generating enough revenue to survive but growing too slowly to exit. Roughly 30–40% of VC-backed companies become zombies — and the 2021 vintage created more than any prior cycle. Here's what happens to founders, investors, and the fund.
Typical SaaS multiples in 2026 are 4–6x NTM revenue at the median, but companies with 120%+ NRR command 10–15x. Net revenue retention — not growth rate — is what separates a 5x company from a 12x company. Here's the full benchmark table and the math behind the premium.
The median SaaS EV/NTM Revenue multiple in 2026 sits at 4–5x, down from 15x+ at the 2021 peak. Investors price SaaS on next-twelve-months revenue — not ARR — because it normalizes across business models and is harder to game. Here's the full benchmark table by growth tier and notable public comps.
Salesforce (CRM) trades at 5–7x NTM Revenue in 2025–2026, down from 12–14x at the 2021 SaaS peak. With $37.9B in FY2025 revenue, 33%+ FCF margins, and 9% growth, here's exactly how the market prices the world's largest CRM — and what it means for private SaaS valuations.
HubSpot (HUBS) trades at roughly 9–11x NTM Revenue in 2025–2026, down from 30x+ at the 2021 peak. With ~$2.6B FY2025 revenue, 85% gross margins, and NRR above 100%, here's the full breakdown of how the market prices CRM SaaS — and what it means for private company comps.
SaaS companies scoring 40+ on the Rule of 40 command 15–25x NTM revenue multiples vs. 5–8x for sub-40 peers. The formula is revenue growth rate % + FCF margin %. Here's the full benchmark data, what top public SaaS companies actually score, and how to improve your number before a premium exit.
Public SaaS median EV/NTM Revenue is 6–8x in 2025–2026, down from 20–40x at the 2021 peak. High-growth SaaS (30%+) with NRR above 120% still commands 12–18x. Here's the full framework investors use to price software companies — by growth rate, Rule of 40, NRR, and gross margin.
Private SaaS companies trade at a 30–50% discount to public SaaS peers at equivalent growth rates. In 2025, public SaaS median EV/NTM Revenue is 6–8x while private Series B rounds clear at 8–15x ARR. Here's the full breakdown by stage, metric, and what it takes to close the gap.
The next billion-dollar AI infrastructure bet isn't another foundation model — it's the middleware that coordinates them. AI orchestration layers are becoming the load-bearing infrastructure of the enterprise AI stack.
The cloud hyperscalers won the AI training war. They may lose the inference war. Edge AI is growing 25%+ CAGR toward a $60B+ market — driven by latency, privacy law, and chips that run GPT-class models on a device the size of a credit card.
Public SaaS comps in May 2026 show a median EV/NTM Revenue of 6.8x — down from 15–20x at the 2021 peak. Rule of 40 companies trade at 10–14x. Here's exactly how to build the right comps set, apply a private-market discount, and arrive at a defensible SaaS valuation.
TVPI, DPI, RVPI, and net IRR are the four metrics every LP and GP uses to evaluate fund performance. Top-quartile VC funds post 3.0x+ TVPI and 20%+ net IRR — but DPI is the only metric that proves real returns. Here's how to read any fund performance report, including Centurium Capital Partners TVPI data.
Lightspeed's Fund VIII (2011) is estimated at 6–9x TVPI, driven by an ~80x return on Snapchat alone. Here's a full breakdown of Lightspeed fund performance by vintage, how it compares to top-quartile VC benchmarks, and where to find LP-level data.
Venture capital fund performance data is scattered across Cambridge Associates, Preqin, Carta, and public pension filings. Top-quartile funds return 3.0x+ TVPI and 25%+ net IRR by year 10. Here are the actual sources — free and paid — and how to turn raw benchmarks into real decisions.
Top quartile VC funds return 3.0x+ TVPI and 25%+ net IRR by vintage year. The median fund sits at 1.5–1.8x TVPI. Only the top 20% of funds consistently beat public markets — and the gap between top quartile and median is wider than most LPs realize.
Most emerging fund managers report raw IRR and TVPI without context — and lose LP trust during the J-curve. Here is how to benchmark against same-vintage peers using Cambridge Associates data, what top-quartile looks like by vintage year, and how to structure quarterly LP letters that build confidence.
LP match is the process of aligning fund managers (GPs) with investors (LPs) by thesis, check size, and stage. First-time funds close at $15–40M after 12–18 months — most GPs fail not because of strategy, but because they pitch the wrong LPs from day one.
The AI Chief of Staff is emerging as a $200K–$350K cross-functional role at Series B+ companies and large enterprises. Unlike a VP of AI, this role owns internal adoption, workflow automation, and executive leverage — and it's becoming as standard as Head of Finance.
NY currently conforms to Section 1202 QSBS, letting founders exclude up to $10M in capital gains at the state level. Active budget proposals could change that before 2027 — a NYC-based founder with a $10M qualifying gain would owe up to $1.48M in additional state and city taxes if NY decouples.
Only 15–20% of seed-funded startups close a Series A within three years — down from ~25% in 2019. The average Series B funding amount in 2025 is $35–45M on a $150–200M post-money. Here's the full data on why the gap is widening and what it takes to cross it.
Permira manages $80B+ across eight buyout funds. Permira VI (2017 vintage) is tracking 2.0–2.4x TVPI — but Permira IV (2006) is the cautionary tale. Here's the full breakdown of DPI, TVPI, and IRR across fund generations vs. large-cap PE benchmarks.
Vista Equity Partners manages $100B+ focused exclusively on enterprise software PE. Top-quartile buyout funds for 2015–2019 vintages show 22–27% net IRR and 2.0–2.8x TVPI. Here's how Vista's operational playbook, realized exits, and vintage timing translate to real LP returns.
Top-quartile PE buyout funds return 2.3–2.7x TVPI and 18–22% net IRR for 2015–2019 vintages. Median PE consistently outperforms median VC on both TVPI and DPI timing. Full benchmarks by vintage year with the data LPs actually use to evaluate fund managers.
RVI gives retail investors liquid access to SpaceX, OpenAI, and Anthropic via a NASDAQ-listed closed-end fund. The total expense ratio runs ~2.5% annually and the fund persistently trades at a 10–30% premium to NAV. Here's what that means for your actual return.
The ROI of AI in supply chain management is documented and large — 15–20% logistics cost reductions, 10–30% inventory savings, 20–50% demand forecasting improvement. Yet 76% of enterprises are still in pilot mode. The bottleneck isn't technology; it's incentive misalignment and data readiness.
Nvidia grew EPS 17x from FY2023 to FY2025. Meta nearly tripled. Amazon flipped from a net loss to $5.53 EPS. Here is the full big-tech earnings breakdown — AAPL, MSFT, GOOGL, AMZN, META, and NVDA — with data tables, margin analysis, and what the divergence means for investors.
Microsoft, Google, Meta, and Amazon committed over $300 billion to AI infrastructure in 2025 — the largest coordinated technology buildout in history. Here is what each company is buying, why the number keeps rising, and what it means for startups building on AI.
Pre-revenue AI companies are raising at $500M–$10B+ based on team pedigree, model benchmarks, and strategic positioning — not DCF math. Here is the exact framework investors use to price AI startups with zero revenue, and what it means for founders raising today.
Essential AI is valued at $500M+ pre-revenue, founded by co-authors of the Transformer paper. Prime Intellect is betting on decentralized compute infrastructure. Here is how the new class of frontier AI labs is priced — and why team pedigree now drives valuation more than revenue.
OpenAI trades at 46x revenue. Anthropic at 30–40x. Public SaaS sits at 6–8x. AI companies valuation multiples aren't irrational — they reflect infrastructure monopoly bets, strategic capital from hyperscalers, and winner-take-most platform dynamics that make traditional revenue-based pricing irrelevant.
OpenAI is valued at ~$157B on $3.4B ARR (46x revenue). Anthropic at $61.5B. xAI at $50B. AI company valuations operate on fundamentally different logic from SaaS multiples — driven by infrastructure monopoly dynamics, strategic capital, and unprecedented revenue growth rates rather than traditional DCF math.
Defense tech VC hit $29.7B in 2023 — triple 2019 levels. Dual-use startups like Anduril ($28B), Shield AI ($2.7B), and Palantir have mastered selling to both the Pentagon and private markets. Here is how defense tech investments work, how OTA contracts replace traditional procurement, and which VC funds are writing the biggest checks.
New York State does not conform to federal Section 1202 QSBS. NY founders can exclude 100% of federal capital gains on qualifying stock held for 5+ years, but still owe NY state tax at up to 10.9%—13%+ for NYC residents. On a $10M gain, that's up to $1.49M owed to New York even with full federal exclusion.
Defense tech VC investment hit $29.7B in 2023 — triple what it was in 2019. The most promising emerging defense companies include Anduril (~$28B valuation), Shield AI (~$2.7B), Hermeus, Epirus, and Vannevar Labs. Here is who is winning contracts, getting funded, and reshaping national security.
Over 16,000 venture rounds closed in the US in 2024. Most investors and founders have no system to track them. This is the free and low-cost stack — SEC Form D filings, Crunchbase alerts, and NVCA data — that gives you a real information edge without paying $30,000/year for PitchBook.
Over 1,200 new VC funds filed with the SEC in 2025 — most LPs missed 90% of them. A proper VC fund tracker combines SEC Form D alerts, state pension disclosures, and Carta benchmarks to surface emerging managers before they close. Here is the free system that actually works.
Most VCs pay $20,000–$50,000 per year for data subscriptions and still miss half the free benchmarks sitting in public filings, pension disclosures, and open-access reports. Here are the 12 best free VC websites — including Carta, NVCA, SEC EDGAR, and state pension fund disclosures — covering fund performance, deal flow, and market benchmarks.
Every VC says they are value-add. Only about 15% actually are. Top funds like a16z deploy 300+ non-investment staff for talent, BD, and marketing — here is what the data shows founders actually get, and how to evaluate it before you sign.
OTPP's 2023 financial statements show $247.5B CAD in net assets with ~12% in private equity. The Maple 8 model of direct co-investment has produced a 9.7% annualized net return since 1990 — here is exactly how these pension giants allocate to VC and PE, and why it outperforms the US approach.
Ontario Teachers' Pension Plan manages $247.5B CAD in net assets with a 9.7% annualized return since inception in 1990 and a 104.6% funded ratio. Here is how one of the world's best pension funds allocates capital across equities, infrastructure, private equity, and fixed income — and what institutional investors can learn from the Maple 8 model.
SPV formation costs $8K–25K and closes in 2–6 weeks. A fund costs $50K–150K+ in legal alone and takes 6–18 months to raise. Here is the decision framework — and the SPV-to-fund pathway most emerging managers actually follow.
On a $1M SPV with 20% carry and a 10x gross exit, LPs net ~8.1x — not 10x. Here is the complete breakdown of SPV fee structures across AngelList, Carta, Assure, and Sydecar, plus the terms smart LPs negotiate before committing.
A $1M SPV with 20% carry and a 10x gross exit returns ~8x net to LPs — not 10x. Here is the complete framework for modeling SPV fees, carried interest, and net investor returns with real numbers from AngelList, Carta, and Assure structures.
Meta cut 21,000 jobs and posted +163% net income growth in the same year. The tech layoffs graph is not a distress signal — it is the story of pandemic over-hiring being unwound by the most profitable companies in history, with AI making sure those headcounts never come back.
Amazon cut 27,000+ jobs, Meta 21,000+, Google 12,000, Microsoft 10,000+. The full tech layoff count by company from 2020 to 2025 — ranked, sourced, and explained by the macro driver behind each wave.
Over 800K tech workers laid off from 2020 to 2025 — but each year had a completely different cause. Full data table by year, the macro driver behind each wave, and what the 2025 AI rebalancing means for hiring now.
The average Series B in 2025 is $28M on a $130M post-money valuation — but the range by sector runs from $15M in commoditized SaaS to $75M+ for AI-native infrastructure. Full data table by sector, dilution norms, and what investors actually require before leading a B.
340+ companies hit $1B+ valuations in 2021. By 2025, roughly 25–30% have been written down below unicorn status, and the median cohort valuation is down 40–60% from peak. Here is the full data on Klarna, Instacart, Stripe, Gopuff, and the rest.
Median pre-seed is $1M on a $4–6M post-money valuation. Seed is $3M on a $12–15M post-money. Series A is $12M on a $40–50M post-money. The 2025 benchmarks every founder needs before going out to raise.
The US leads the global unicorn index with 650+ companies worth $2.4T — roughly 52% of all unicorns worldwide. China is second with ~170, India third with ~70. AI now drives 40%+ of new unicorn creation globally, and India has the fastest growth rate at +18% YoY.
The median startup takes 7 years to reach unicorn status, but AI-era companies are doing it in 4–5. Data from 1,336 unicorns broken down by sector, geography, and vintage year — plus why 20–25% of the 2021 cohort has since been written down.
New York City has produced 100+ unicorn startups backed by VC funding, making it the #2 startup ecosystem globally. From Ramp ($8.1B) to Chainalysis ($8.6B) to Attentive ($10B+), NYC unicorns cluster in fintech, enterprise software, and healthcare tech — not consumer apps.
A $50M fund generates roughly $1M per year in management fees — barely enough to cover two partner salaries and operating costs. Meaningful carry only materializes on a 3x+ return, which fewer than 20% of funds achieve. The economics are tight, the competitive positioning is awkward, and the path to GP wealth is longer than anyone admits.
Family offices now control over $6 trillion in global assets and are deploying directly into startups at pre-seed through growth stages. With no fund lifecycle pressure and no LP reporting requirements, they operate with a structural patience that traditional VCs simply cannot match.
The global secondary market hit $150B in 2024 — triple its 2019 volume. With LP distributions collapsed 80% and the IPO window shut for three years, secondaries have become the dominant venue for price discovery and liquidity in venture capital.
LPs are done accepting paper gains. After 2021's mark-up bubble left hundreds of funds with inflated TVPI and near-zero distributions, the venture industry is finally reckoning with the only metric that actually matters: cash returned to investors.
Continuation funds now represent ~30% of the $150B secondaries market. GPs use them to hold winners longer instead of forcing exits — but they create real conflicts of interest between fund managers and the LPs who trusted them with capital.
Everyone blames rates, the IPO market, or the 2021 bubble. The real cause is structural: fund sizes grew 10x while power law math stayed the same. Only 6% of deals drive 60% of returns — and deploying 20x more capital into the same number of winners guarantees diluted IRRs.
The median Series A bar jumped from $500K to $2M+ ARR between 2021 and 2025, creating a gap that bridge rounds now fill. Nearly 38% of seed-stage startups raised a bridge in 2024 — here's why that's structural, not cyclical, and what founders should do about it.
AngelList has deployed over $10B through SPVs — and the structure is now a serious challenger to the 10-year blind pool. No management fees, deal-by-deal selection, and 100% of capital going to work: here is why LPs are paying attention.
The median US venture fund returns 1.3x net — below a Treasury bill. The top decile returns 5x or more. Cambridge Associates data shows the top 10% of VC firms generate 80-90% of all industry returns, and the structural advantages that cause this gap only compound over time.
Fewer than 5% of first-time fund managers raise from institutional LPs. The ones who close fast lead with a differentiated sourcing edge, proof of conviction through personal checks, and reference-checkable founder relationships — not deal volume or pedigree.
Spray and pray produced a decade of zombie portfolios. Top-quartile funds in 2026 run 25-35 concentrated positions, reserve 40-50% of capital for follow-ons, and target 8-12% ownership at entry. Here's the math that separates fund-returners from the rest.
Fewer than 5% of first-time fund managers raise from institutional LPs on Fund I. The ones who close fast understand who actually writes first checks, how to build proof before fundraising, and why a hard deadline is the most underrated fundraising tool.
LP distributions collapsed from $220B in 2021 to under $60B by 2024 — an 80%+ drop that almost nobody discusses openly. DPI on 2018-2021 vintage funds averages below 0.3x, and LPs sitting on nearly $3 trillion in paper-valued private assets are running out of patience.
The data is clearer than most mega-fund GPs want to admit: funds under $100M consistently beat funds over $1B on net returns. Top-quartile micro-fund TVPI regularly exceeds 3.5x while mega-fund medians hover near 1.6x — and the structural reasons aren't going away.
While everyone debates AGI timelines and foundation model wars, AI has already quietly taken over advertising technology. $700B+ in global digital ad spend is now optimized, targeted, and measured by AI — and most advertisers don't even realize it happened.
Knowledge workers spend 30% of their day searching for information and fail to find it 44% of the time. RAG is the first architecture to actually fix enterprise search — combining real-time document retrieval with language model reasoning to deliver cited, accurate answers across every tool an enterprise runs.
The global insurance industry writes $7 trillion in premiums annually. AI is creating an entirely new liability category — model hallucinations, algorithmic discrimination, autonomous decision errors — that no existing policy covers. With 85% of enterprises deploying AI but fewer than 15% carrying AI-specific coverage, the underwriting gap is enormous and growing.
Bears cite overvaluation and hype cycles. Bulls cite $500B+ in committed infrastructure and genuine productivity gains. Both are partially right — and the distinction matters enormously for where you put capital.
Everyone benchmarks AI on parameter count. The metric that actually determines enterprise value is context window length — how much a model can hold, reason over, and act on in a single pass. A 500x expansion in three years has quietly changed what AI can do.
The $300B BPO industry was built on labor arbitrage — pay $15/hour offshore instead of $150/hour domestically. AI is collapsing that model by automating the exact tasks that made offshore labor valuable, and replacing the formula with AI-augmented specialists priced on outcomes, not hours.
The companies winning in AI aren't just collecting more data — they're generating it. Gartner estimates 60% of AI training data will be synthetic by 2026, and the startups mastering generation pipelines are building durable moats that raw compute cannot buy.
GPT-4 scored in the 90th percentile on the bar exam. AI outperforms radiologists on specific imaging tasks. But benchmark performance and production reliability are two different things — and understanding the gap is what separates smart AI deployment from costly disappointment.
Traditional SaaS multiples have compressed to 5-8x ARR while AI-native companies still command 15-25x. The market is pricing in displacement risk — and every SaaS company that hasn't rearchitected its core product loop around AI is already on the wrong side of that spread.
The SDR role is being restructured, not eliminated. AI handles list building, personalization, and sequencing at a scale that used to require 10-person BDR teams — companies using AI outbound tools report 3-5x more pipeline coverage per rep. What's left for humans is the judgment work.
78% of enterprises claim to use AI. Fewer than 20% have anything real in production. The bottleneck isn't capability — it's the absence of trust infrastructure: audit trails, liability clarity, and governance frameworks that legal teams can actually sign off on.
Everyone obsessed over training costs. The next trillion-dollar constraint in AI is inference — and whoever controls it sets the economics for the entire industry. At scale, inference spend dwarfs training by 10x or more.
Everyone is chasing text AI. The real enterprise opportunity is voice — 100B+ business calls per year, $400B in global contact center spend, and latency barriers that create genuine moats. Voice is the interface business actually runs on.
Inference costs dropped 95%+ in 18 months. Foundation models are commoditizing fast. But Harvey, Ambience, and Glean are technically wrappers worth billions — because the moat was never the model. Here's what the wrapper debate gets wrong.
Every AI system today starts from zero — no history, no preferences, no institutional knowledge. Persistent memory transforms one-shot chat into compounding intelligence, and the startups that own that memory layer will capture disproportionate enterprise value.
Foundation models just unlocked robotics the same way they unlocked software. Figure AI raised $675M at $2.6B. Physical Intelligence raised $400M at $2.1B. The 20-year science project is now a fundable category with a real inflection behind it.
JPMorgan runs 400+ ML models in production. Klarna replaced 700 support agents with AI. AI-native lenders report 30-50% lower defaults. The race to rebuild the $25 trillion financial services industry is underway — and vertical specialists are winning.
The average top-tier fund screens 1,000–1,500 companies per year and writes checks into 5–10. AI is compressing the research stack between those two numbers — cutting first-pass diligence from weeks to days while leaving conviction, timing, and founder judgment firmly in human hands.
HR manages 60-70% of operating expenses yet runs on a broken stack of 12+ disconnected tools. AI is now collapsing recruiting, performance, and workforce planning into unified systems — and the $32B market is finally ready to be disrupted.
AI in supply chain is delivering 15-30% inventory cost reductions and $1-3M annual savings per major enterprise implementation. But only 23% of projects reach production scale — the implementation gap, not the technology, is the dominant ROI killer.
Bolt.new crossed $40M ARR in 8 months. Lovable hit $10M in 60 days. No-code AI has collapsed MVP build costs by 60-80% and shrunk development timelines from months to days — forcing founders and VCs to find defensibility somewhere other than the ability to write code.
The $1.1 trillion global legal services market is being restructured by AI — contract review that once took 8 hours now takes 12 minutes. Harvey AI, Luminance, and Casetext are compressing due diligence timelines and attacking the billable hour model from the inside.
Google AI Overviews now answer 65% of searches without a click. Organic traffic to content sites is down 30-50% year-over-year. The ten-blue-links era is over and most founders haven't updated their playbook.
Executives waste 30-40% of their time on synthesis, status tracking, and low-value decision prep. AI systems that replace this function cost under $50K per year, operate 24/7, and have zero organizational politics. The AI Chief of Staff is not a role — it's an operating system.
The $300K 'prompt engineer' job title is collapsing as models improve. Job postings declined over 70% from peak. The real LLM value has shifted to fine-tuning, RAG architecture, and multi-agent orchestration — engineering depth, not clever phrasing.
The pilot worked. The demo was impressive. The budget was approved. Then year two arrived. 85% of enterprise AI projects never reach production scale — not because the tech fails, but because data quality, change management, and internal sponsorship collapse.
Model benchmarks are table stakes. The AI companies hitting $100M ARR share four real differentiators: distribution moat, workflow ownership, proprietary feedback loops, and a compounding GTM motion. Everything else is noise.
GitHub Copilot, Cursor, and Claude Code are compressing per-engineer output costs by 30–55%. Companies are choosing not to backfill departing engineers and achieving the same velocity with smaller teams — permanently resetting the unit economics of software development.
Every AI pitch deck claims a data moat. Most are wrong. In an era of synthetic data generation and foundation model fine-tuning, static proprietary datasets are no longer a durable competitive advantage — real moats come from feedback loops, not file storage.
Over 4,000 AI agent startups are competing for the same enterprise budgets. Most are building on sand. The companies that win in the agentic era own the orchestration layer, memory systems, and reliability stack — not just the agents. Here's why the product framing is wrong and what actually survives the next model release cycle.
Everyone is building an AI startup. Almost none will reach $100M in revenue. Fewer than 0.5% of seed-stage companies ever hit that threshold historically — and AI's commoditizing models, narrowing margins, and intensifying competition from model providers make that ceiling even harder to break through.
Most startup marketers look great on paper but can't generate a dollar of pipeline. The wrong hire burns budget on brand awareness while your CAC spikes. Here's the framework for identifying demand marketers over brand marketers — and the take-home assignment that reveals everything.
Salesforce generates 70% of $25B+ in annual revenue through partners. HubSpot's agency program drove 40%+ of new customers for years. Yet fewer than 20% of B2B startups have a structured partnership program. The CAC advantage — 60-80% below direct sales — makes this the most overlooked compounding growth lever in startups.
Most founders track MRR but have never run a cohort analysis. A 3% monthly churn destroys 31% of ARR annually — and the gap between a 10x ARR valuation and a 3x one almost always comes down to net revenue retention and LTV:CAC ratios founders misread or ignore entirely.
Most waitlists are vanity metrics that convert under 5%. The ones that built companies — Robinhood's 1M-signup day, Superhuman's 275K selective queue, Notion's cohort launch — combined referral mechanics, selective admission, and a communication cadence that kept people angry they weren't in yet.
PLG powered Slack to $27.7B, Figma to $20B, and Atlassian past $50B — but most founders who copy the playbook get it wrong. Free-to-paid conversion averages 2–5% for pure self-serve. Here's the framework for knowing if PLG is right for your business.
Generic cold email is dead. Signal-triggered, hyper-personalized outbound is alive — with top quartile senders hitting 9.4% reply rates versus a 2.1% median. Here is the exact playbook for booking meetings in 2026.
Paid ads stop the moment your budget runs out. SEO compounds for years. With organic search driving 53% of all web traffic and content generating 3x more leads at 62% lower cost, here's the data and the 2026 playbook for startups.
Dropbox grew 3,900% with a single referral mechanic. Slack hit $7B without paid ads. Word of mouth is the highest-ROI growth channel — but it doesn't happen by accident. Here's how to engineer it from day one.
Most startup growth tactics plateau after 90 days. True growth loops — viral, content, product-led, and community — compound over time, lowering CAC while increasing retention. Here's how to tell the difference and build the right engine.
Paid ads are getting more expensive and less effective — Google CPCs are up 122% since 2019. Content marketing costs 62% less, generates 3x the leads, and compounds for years after you stop paying. Here's the data and the 2026 playbook.
Strategic investors offer brand validation, distribution, and larger checks — but ROFR clauses, enhanced information rights, and exclusivity traps can quietly kill your exit. Here's how to take strategic money without giving away the company.
Getting a term sheet feels like the finish line. It isn't. Roughly 20-25% of signed term sheets never close — killed by re-trading, cap table problems, syndicate collapse, and reference checks. Here's the full picture of what happens between handshake and wire.
VCs run 15-20 references on every founder before wiring money. Fewer than 30% of founders do the same on their investors. Here's the exact back-channel framework — the five questions that reveal board character and how to find the references investors didn't give you.
Most founders raise ad hoc and drag rounds out for 4-5 months. Structured 6-8 week competitive processes close faster, at 15-25% better valuations, and with far less distraction. Here's the exact architecture — from list-building to deadline-setting to close.
Over-raising at seed doubles your dilution, inflates your Series A target, and creates organizational complexity before you have product-market fit. Here's the burn multiple math and the framework for raising exactly what you need.
Most founders either name a number too early and anchor low, or dodge the question so long they look unsophisticated. Here's how to anchor with market comparables, defend with return math, and use competitive process to close at the price you deserve.
At seed stage, investors fund conviction, not spreadsheets. Your narrative — why now, why you, why this market — matters more than early revenue. Here's the exact story structure that closes seed rounds fast.
Most investor updates are either missing entirely or packed with vanity metrics that tell investors nothing useful. Here's the exact structure that builds trust, accelerates future raises, and turns passive cap-table names into active allies.
Most cold investor emails get deleted in under 5 seconds. Here's the exact framework — based on 65+ investments — that gets founders into meetings with VCs who don't know them.
Vision gets you in the room. Execution gets you to Series A. After 65+ investments, the pattern is clear: founders who ship something new every two weeks beat founders with the most cinematic 10-year story — every time.
Building in public has driven 7-figure ARR for Buffer, Ghost, and Levels.health. But there's a line between strategic transparency and self-sabotage — and most founders cross it without realizing.
Airbnb was rejected by 7 top-tier VCs. Uber seemed like a black car app for rich people. The ideas that change industries always look dumb before they look obvious — and that pattern is not an accident.
Bill Gross analyzed 200 startups and found timing accounted for 42% of the difference between success and failure — more than team, idea, or funding. Ideas don't fail. Execution doesn't fail. Timing kills more startups than either, and almost no one in the industry talks about it honestly.
Every pitch deck claims network effects. Almost none actually have them. There are six distinct types — and the difference between a strong one and a weak one can determine whether your startup compounds or collapses under competitive pressure.
The platform vs point solution debate is one of the most consequential strategic choices a founder makes. Most founders try to build a platform before they've earned the right to — and it kills the company. Here's the framework for knowing when to stay focused and when to expand.
Traditional moats — switching costs, scale economics, feature lock-in — are being eroded faster than any prior technology cycle. When AI can replicate your core product in 90 days, defensibility is no longer about what you built. It is about what competitors cannot take from you.
Most founders undercharge because they fear losing a deal. Price signals value, filters for the right customers, and determines your unit economics before any other business decision does. Get it wrong and no amount of growth fixes it.
Most founders default to funnel thinking because it fits neatly into attribution models — but the companies that compound decade over decade are all built on flywheels. Here's how to diagnose which model your business actually needs, and why the difference shows up in your CAC trajectory long before it shows up in your P&L.
Most founders confuse brand with marketing — but brand is the only business asset that appreciates over time. Apple survived near-bankruptcy in 1997 not because Jobs fixed the products first, but because he rebuilt the brand. In a market where features get copied in months, brand is your last durable moat.
AI companies are being valued at 80-100x revenue on narratives that don't survive basic scrutiny. OpenAI at $300B, Anthropic at $61B, xAI at $50B — the math only works if you assume an impossible market concentration. The correction doesn't have to be dramatic to be painful.
Most VCs won't touch defense AI for ethical or structural reasons. That discomfort is creating one of the most durable, capital-efficient opportunities in venture today. Palantir crossed $100B market cap, Anduril hit a $14B valuation, and private defense tech investment grew from under $1B in 2019 to $20B+ in 2024.
Solo GPs now manage over $100B in AUM across thousands of micro-funds. They close faster, specialize deeper, and align more completely with founders than multi-partner firms.
Meta's Llama 3 hit 350M downloads. DeepSeek R1 matched GPT-4 on reasoning benchmarks at 95% lower cost. When frontier models go free, the entire value chain of AI startups shifts.
Headline numbers claiming $300B+ in undeployed VC capital are deeply misleading. Most of that money is spoken for, locked up in follow-on reserves, or structurally inaccessible to new founders.
Everyone is laser-focused on token pricing and GPU costs. Nobody is talking loudly enough about what AI actually consumes at scale: electricity grids running dry, aquifers being depleted, and data center land that simply does not exist.
Every major AI lab has demonstrated multimodal capabilities. GPT-4o sees and hears. Gemini processes hour-long video. None of it has produced a product people actually use to do their jobs differently.
Why the best companies in an AI-native world look less like fortresses and more like infrastructure. From protecting scarcity to orchestrating movement.
The complete breakdown of how VC funds raise money, make investments, and generate returns — explained by someone who does it.
What each stage actually means, how much to raise, what investors expect, and when to move from one to the next.
The 12 slides every pitch deck needs, common mistakes that kill deals, and real examples from decks that raised millions.
After 65+ investments, here's what actually moves the needle — and what founders waste time on that doesn't matter.
What SAFE notes are, how they work, valuation caps vs. discounts, and the mistakes that cost founders millions in dilution.
The math every founder needs to know — gross burn, net burn, runway, and when to panic.
The signals that tell you it's time, what role to hire first, and how to avoid the mistakes that sink early-stage companies.
How to split equity fairly, set up vesting, and avoid the fights that kill companies.
From incorporation to first revenue — every dollar you'll spend and where to save.
The honest trade-offs between remote, hybrid, and in-person — with actual data from the startup ecosystem.
Where the money is going, what's changed since the ZIRP era, and what it means for founders raising now.
Separating the signal from the noise in the biggest technology shift since the internet.
The real reasons 90% of startups don't make it — and the patterns that separate survivors from statistics.
The IPO window, what it takes to go public, and why most VC-backed companies never will.
Why more investors are going solo, how micro funds work, and what it means for founders.
Why the traditional fundraising ladder is breaking — and what replaces it.
Deal flow follows attention. The firms that figured this out are winning.
AI doesn't replace developers. It makes the distinction between 1x and 10x irrelevant.
Andreessen was right. But the next phase looks nothing like the last one.
The venture capital model has a structural issue at the top of the stack.
Where to look, what to look for, and how to test the relationship before committing.
The playbook for going from zero to 100 — no growth hacks, no shortcuts, just what works.
Pre-money, post-money, revenue multiples, and the art of pricing a company with no revenue.
The minimum viable product framework — what to build, what to skip, and how to ship fast.
The signals that tell you it's time to change direction — and how to pivot without losing everything.
The clauses that matter, the ones that don't, and how to negotiate without killing the deal.
How cap tables work, how they change at each round, and the mistakes that silently kill founder ownership.
Not all money is equal. How to evaluate investors and choose a partner you won't regret.
The honest trade-offs between self-funding and taking venture capital.
Who to recruit, how much equity to give, and how to actually get value from your advisors.
How trade policy is reshaping supply chains, costs, and strategy for technology companies.
Agents are the next platform shift. Here's who captures the value.
The case for going deep instead of wide — and why vertical AI may be the better bet.
Too many funds, not enough returns. The reckoning is here.
Beyond the headlines — what's really happening with tech employment.
Why the most important fit isn't product-market fit — it's founder-market fit. How to evaluate yours before you pitch.
The difference between slapping a UI on GPT-4 and building a genuinely AI-native product — and why it matters for investors and founders.
The SaaS playbook from 2010-2021 is broken. What's working now, what metrics actually matter, and how to build a SaaS company that survives AI.
The exact documents VCs want to see, how to organize them, and the mistakes that slow down — or kill — funding rounds.
Secondary transactions are reshaping venture capital. How they work, why they're booming, and what it means for founders, employees, and LPs.
~$1.3 trillion in value erased. SaaS categories down 30–50%. AI-native companies compounding in the opposite direction.
OpenAI acquired TBPN — not for the media property, but for a repeatable production and distribution machine. Here's what they're actually building.
15 IPOs. $10.4B raised. Only 4 trading above IPO price. Median return -48.7%. $87B in value destruction.
PE is outperforming VC on realized returns — but that misses the story. PE ~8% annual return. VC ~6%. But DPI is near zero for most 2020–2022 vintage funds.
71 companies. 8 sectors. 590 data points. Median EV/Revenue down 52% from peak — but semiconductors tripled.
The data tells a more precise story than the narrative. Public SaaS is still growing at 18–20% with 70%+ gross margins. What changed is how the market values 'good.'
The Brex acquisition isn't just a big exit. It's about where Capital One is repositioning in the trillion-dollar payments stack.
Analyzing performance across 49 funds from Thrive, a16z, Founders Fund, Lightspeed, and more. Scale compresses returns — not in theory, in practice.
The narrative says AI agents plus stablecoins will eliminate interchange. As someone who worked at American Express on B2B payments, I don't think it plays out that way.
Capital is moving again. Large rounds are closing. Yet for many startups and investors, the environment feels more constrained than any point in the last decade. Both are true.
Andreessen Horowitz announced $15B in new funds. The argument that large funds can't produce outsized returns is worth revisiting — Fund III tells a different story.
What has changed is not some abstract leap toward AGI. It's the compression between intent and execution. The limiting factor is no longer code. It's clarity.
Enterprise AI adoption isn't a technology problem. It's an incentives problem. The people closest to the work are being asked to take the most risk for the least reward.
Google did it with CharacterAI. Microsoft with Inflection. Amazon with Adept. Now Nvidia with Groq. Hire & License 2.0 has become the dominant AI acquisition strategy.
VC rebounds 30%. Enterprise AI scales to production. Vertical AI dominates early stage. Big Tech exceeds $500B in infrastructure spend. And one visible AI company fails publicly.
$100B+ in tech M&A in 2025, almost entirely focused on AI, cybersecurity, infrastructure, and vertical software. Enterprises are buying because building is no longer viable at the speed the market demands.
Carta's 2025 Fund Economics Report gives us actual visibility into how funds function today. 75%+ of capital calls paid on time. 2022 vintages only 67% deployed.
Explore 41+ free VC tools and dashboards, plus our recommended startup software.