How LPs Choose VC Funds: DPI, a 3% GP Commitment, and the 70% Re-Up Bar

14-15% of NAV is all the average VC fund has distributed back to LPs since late 2022 — and that stalled distribution rate is why DPI, not IRR or TVPI, is now the first number institutional LPs check before writing a check. That's the short answer. The longer answer is that LPs have quietly rewritten their entire scoring rubric over the last two years, from GP commitment size to what counts as an acceptable re-up rate.

For most of the 2010s, an LP's fund selection process leaned heavily on IRR and TVPI — paper metrics that look great when markups are generous and say almost nothing about whether an investor will ever see cash back. That worked fine during a decade of near-continuous markups. It stopped working once the exit market froze, distributions dried up, and LPs realized they'd been underwriting funds on numbers that never converted to actual returned capital.

How LPs choose VC funds in 2026: the decision framework, ranked

LPs choosing VC funds in 2026 run every manager through DPI first, then track record depth, GP commitment size, and an operational due diligence (ODD) review — with DPI acting as the gatekeeper metric that determines whether the rest of the pitch even gets a serious read. A manager showing 3x TVPI on paper but 0.2x DPI five years into a fund's life now draws immediate scrutiny rather than automatic credibility.

That's a real reversal from the 2015-2021 period, when TVPI markups alone were often enough to secure a re-up. The shift happened because the exit market — IPOs and M&A — went quiet enough for long enough that an entire cohort of LPs sat through multiple fund cycles without meaningful cash coming back, regardless of how attractive the mark-to-market numbers looked in a quarterly performance report.

Why DPI beat IRR and TVPI as the metric LPs check first

DPI (distributions to paid-in capital) measures actual cash returned to LPs, while TVPI and IRR can both be inflated by unrealized markups a GP controls. Industry-wide DPI has been stuck near 14-15% of NAV since late 2022 — meaning the overwhelming majority of reported venture returns still exist only as marks on a cap table, not cash in an LP's account.

Cambridge Associates data still shows top-quartile VC funds producing 3x+ TVPI and 20%+ net IRR, so paper performance hasn't collapsed — but LPs increasingly want a chunk of that 3x expressed as realized DPI, not projected value, before committing to Fund III or Fund IV. Most institutional LPs now require a TVPI of 2x or higher on a prior fund, backed by 3-5 years of data, before writing a first check to a new relationship.

This is the same distribution-drought dynamic we've tracked across the broader VC and PE performance comparison data: private markets generally are holding more unrealized value for longer, and LPs across every asset class are recalibrating around cash-on-cash proof rather than mark-to-market optimism.

GP commitment: the alignment signal LPs now demand at 3%, not 1.5%

The median GP commitment for first-time funds reached 3% of total fund size in 2026, exactly double the 1.5% median LPs accepted in 2020. A GP writing a meaningfully larger personal check into their own fund is the cleanest available proxy for whether a manager's incentives are actually aligned with the LPs underwriting them, and it's become a standard line item in every term sheet negotiation.

Operational due diligence: the checklist that rejects 87% of managers on its own

LP diligence in 2026 runs two parallel tracks — investment due diligence (IDD) on strategy and track record, and operational due diligence (ODD) on the back-office plumbing: fund administrator, auditor, legal counsel, and compliance infrastructure. An estimated 87% of LPs have rejected a manager over ODD concerns alone in 2026, even when the investment thesis and headline returns were strong enough to pass IDD cleanly.

The average due diligence questionnaire (DDQ) now spans 23 sections and more than 280 questions, covering everything from partner succession planning to how carry is calculated on a deal-by-deal basis. For emerging managers, this is often the more expensive hurdle to clear than the fundraising pitch itself, since it requires paying for institutional-grade fund administration and legal infrastructure well before the fund has meaningful AUM to justify the cost.

What LPs actually ask for before a first check gets written

Beyond the DDQ itself, LPs now expect a standard document set before serious conversations start: full net IRR and DPI by fund vintage, deal-level attribution across every prior fund (not just headline aggregate returns), partner and key-team bios, a live reference list, and a breakdown of the back-office stack — who the fund administrator is, which firm audits the books, and who serves as outside counsel. Fund terms disclosure has also become table stakes upfront, meaning management fee, carry, hurdle (if any), and fund life all get laid out before a first in-person meeting rather than negotiated after interest is confirmed.

LPs increasingly cross-reference a manager's stated IRR and TVPI against Cambridge Associates or Preqin vintage-year benchmarks rather than taking a pitch deck's framing at face value, which means a fund claiming "top-quartile" performance now needs to show exactly which benchmark and vintage year that claim is measured against. Managers who can't produce deal-level attribution on request — showing which specific investments drove the fund's returns, not just the blended number — are treated as a diligence red flag on their own, independent of whether the aggregate numbers look strong.

2026 LP fund-selection criteria compared: old standard vs new standard

The table below lays out how each major LP evaluation criterion has shifted, comparing the pre-2022 standard against what LPs actually enforce in 2026.

Figures are 2026 estimates blended from altss.com, VC Lab, Praxis Rock, and PipelineRoad LP survey data. GP commitment and re-up figures reflect reported medians across institutional LP networks, not any single fund.

Are LPs still backing first-time and emerging managers in 2026?

Yes, but the capital is concentrated in small checks and specific sectors: almost 90% of LP commitments to emerging managers in Q1-Q2 2026 went to funds under $15 million, roughly 70% went to seed-stage strategies, and AI, deeptech, and healthcare were the most popular sectors for that capital. Around 75% of individual LP commitments to emerging managers came in under $150,000, per VC Lab's data across more than 900 firms — LPs are testing relationships in small size before scaling up in later funds.

The catch is that this emerging-manager appetite hasn't meaningfully shifted where the dollars actually go: established managers still captured 90.9% of total VC fundraising in Q1 2026. Emerging managers who over-communicate with detailed, honest quarterly updates report meaningfully stronger Fund II and Fund III re-up rates, which tracks with everything else in this piece — LPs are rewarding transparency and demonstrated discipline over pitch-deck promises at every stage of the relationship.

What this means if you're raising a fund right now

If you're a GP heading into a fundraise, the practical takeaway is to lead your fundraising deck with DPI and realized outcomes rather than markup-driven TVPI, and to have your fund administrator, auditor, and compliance stack in place before the first LP meeting rather than scrambling to assemble it once diligence starts. The 87% ODD rejection rate isn't a strategy problem — it's an operational-readiness problem, and it's entirely avoidable with earlier preparation.

For LPs, the framework above is effectively becoming table stakes across the industry rather than a differentiator any single LP is using — which means GPs who can't clear DPI, GP commitment, and ODD screens cleanly are going to see their addressable LP base shrink further in 2027, regardless of how strong the underlying portfolio companies look on paper.