VC
Value Add VC
⚡HomePulse⚡Helpful Apps📝Blog🤝Partner
Illustration for: OLIX's $312M Bet Against Nvidia's Chip Shortage
Value Add VC/Pulse/FUNDING$312M Series B

OLIX's $312M Bet Against Nvidia's Chip Shortage

OLIX raised $312M to build a photonic inference chip that skips the HBM shortage bottlenecking Nvidia -- backed by Arm, Hudson River Trading and Reed Hastings, and a real signal that the AI chip market is finally getting a credible non-GPU challenger.

By the Numbers

$312M
Round size
$3.3B
Valuation
~2 years
Company age
H2 2027
First shipments
TC
Trace Cohen
Early-stage VC & angel · Founder, New York Venture Partners
August 3, 2026
2 min read
ShareXLinkedInEmail

THE RUNDOWN

1

London-based OLIX raised $312M (€270.5M) in a Series B at a $3.3B valuation, roughly two years after its founding

2

Its DX-1 decode accelerator uses fast on-chip SRAM instead of High-Bandwidth Memory, letting it bypass the HBM supply shortage currently constraining the entire GPU industry

3

Backers include Arm, Hudson River Trading, the UK's Sovereign AI venture fund, and individual investors including Netflix co-founder Reed Hastings

4

The company is targeting first customer shipments of its DX-1 chip by the second half of 2027, betting compute-efficiency-per-dollar, not raw performance, is the next battleground

TC

The VC Read · Trace's Take

Trace Cohen

Every AI-chip pitch for two years has been "we're faster than Nvidia." OLIX's pitch is "we don't need the thing that's actually scarce," which is a much better wedge given HBM allocation is the real bottleneck right now, not raw compute. The 2027 shipment date is the number I'd actually underwrite as a GP -- silicon companies slip, and an 18-month gap between a $3.3B mark and first revenue is a long time for a thesis to stay right.

AI Buildout Tracker →AI Chip Supply, Ranked →

Analysis

OLIX, a London-based AI chip startup founded roughly two years ago, raised $312 million ($270.5 million) in a Series B at a $3.3 billion valuation -- one of the largest rounds any AI hardware challenger has raised this year outside the US. The round adds Professor Nick McKeown, a pioneer of software-defined networking, to its board.

Betting Against HBM, Not Nvidia

The technical bet is specific and, if it works, genuinely disruptive to Nvidia's current moat. OLIX's first chip, the DX-1 decode accelerator, relies on fast on-chip SRAM instead of High-Bandwidth Memory for token generation -- the exact component category currently rationed across the entire semiconductor industry due to supply constraints. By designing around HBM entirely rather than competing for the same scarce packaging capacity as Nvidia and AMD, OLIX is betting it can deliver customers usable inference chips years before HBM supply catches up to demand.

“## Betting Against HBM, Not Nvidia The technical bet is specific and, if it works, genuinely disruptive to Nvidia's current moat.”

The investor list underlines how seriously infrastructure players are taking that bet: Arm and Hudson River Trading both backed the round, alongside the UK government's Sovereign AI venture fund and individual angels including Netflix co-founder Reed Hastings. That's a mix of strategic silicon expertise, a sophisticated trading firm with direct interest in inference economics, and sovereign capital betting on domestic chip capability -- not just financial investors chasing an AI-adjacent story.

The timeline is the honest risk here. OLIX isn't targeting first customer shipments until the second half of 2027, meaning the company is raising at a $3.3 billion valuation roughly 18 months before it proves the architecture at commercial scale. That's standard for frontier silicon -- Cerebras and Groq both ran multi-year gaps between funding and shipping -- but it means today's valuation is pricing execution risk that won't resolve for well over a year.

What to watch: whether OLIX can hold its 2027 shipment timeline as HBM supply either eases or tightens further, since a supply-chain fix from the incumbents would blunt the core argument for OLIX's SRAM-based approach before it ever reaches customers.

ShareXLinkedInEmail

More on

Nvidia →

Analysis and editorial commentary by Value Add Pulse.

← Back to Pulse

THE WIRE in your inbox— Tech, startup & VC news with Trace's take. Free, no spam.

Read Next

FUNDING· Aug 3, 2026

P-1 AI Raises $50M to Automate Engineering Design

Illustration for: P-1 AI Raises $50M to Automate Engineering Design
FUNDING$50M Series A

P-1 AI Raises $50M to Automate Engineering Design

P-1 AI raised a $50M Series A led by NEA for Archie, an AI agent that performs mechanical and electrical design work -- one of the few funding stories this week aimed at industrial engineering rather than software.

FUNDING· Aug 4, 2026

The Physical AI Funding Wave, By The Numbers

Illustration for: The Physical AI Funding Wave, By The Numbers
FUNDING

The Physical AI Funding Wave, By The Numbers

This week's largest venture checks all went to physical infrastructure -- nuclear power, chips, energy -- not software, and the dollar totals now dwarf what's going into application-layer AI startups.

FUNDING· Aug 4, 2026

Defense Tech's Capital Keeps Compounding, Not Cooling

Illustration for: Defense Tech's Capital Keeps Compounding, Not Cooling
FUNDING

Defense Tech's Capital Keeps Compounding, Not Cooling

Defense tech has already pulled in more venture funding in 2026 than all of 2025 combined, and the checks keep landing even as public defense-tech comps wobble -- a split between private conviction and public market patience.

@Trace_Cohen·t@nyvp.com