Which VCs are most active in AI startups in 2026?

Per Dealroom data cited by CNBC (April 28 2026), $18.8B has flowed into AI startups founded since the start of 2025. The most active VC firms are Sequoia, Lightspeed, Andreessen Horowitz, Greylock, Index Ventures, DST Global, and Kleiner Perkins. Strategic investors NVentures (Nvidia's venture arm) and Google Ventures are also extremely active. Sovereign AI funds (UK, UAE MGX, Saudi PIF, France) are now first-tier early-stage participants.

What does Nvidia NVentures invest in?

Nvidia's venture arm NVentures focuses on AI startups across the stack — model labs (Mistral, Cohere historically), AI infrastructure (CoreWeave precedent), and vertical AI (Legora at $5.6B Swedish legal AI, April 30 2026). NVentures' edge is GPU allocation alongside capital — portfolio companies get priority access to H200s, B200s, and Vera Rubin compute. For AI startups burning compute, that's often more valuable than the dollars.

How do sovereign AI funds work?

Sovereign AI funds are state-backed investment vehicles deploying patient capital into AI strategic priorities. Examples: UK Sovereign AI Fund (participated in Ineffable Intelligence's $1.1B seed), UAE's MGX (Mubadala Group), Saudi PIF, France's strategic AI investments, Singapore's Temasek. They write large checks ($100M+), are price-insensitive on prestige bets, and bring access to national datasets, GPU allocations, and government customers.

What's the bar to raise AI VC in May 2026?

Hyper-bifurcated. Prestige research labs (DeepMind / OpenAI / Meta alums) raise $300M-$1.1B at concept stage from Sequoia, Lightspeed, NVentures, sovereigns. Application-layer AI companies need 100+ enterprise customers (Decagon's bar) before Series A. The middle is thinner than ever — $5-50M Series A rounds are harder to raise unless you have either prestige team or strong revenue.

Quick Answer

Best AI VCs Backing Startups in May 2026

Published: May 3, 2026

Best AI VCs Backing Startups in May 2026

$18.8 billion has flowed into AI startups founded since the start of 2025, per Dealroom data cited by CNBC on April 28, 2026. The capital is plentiful but selective: prestige research labs raise $1B+ at concept stage; application-layer companies need real revenue; the middle is thinner. Here are the most active AI investors in May 2026 and what they back.

Last verified: May 3, 2026

The active AI VC roster (May 2026)

Firm	AI focus	Notable May 2026 deals
Sequoia Capital	Frontier labs + application AI	Co-led Ineffable Intelligence $1.1B seed (April 27 2026)
Lightspeed Venture Partners	Frontier labs + verticals	Co-led Ineffable; active across AI stack
Nvidia NVentures	AI infrastructure, models, verticals	Backed Legora $5.6B (April 30 2026); Ineffable Intelligence
Andreessen Horowitz (a16z)	Application AI, infrastructure	Active across coding agents, AI customer support, vertical AI
Greylock	Application AI, enterprise	Decagon precedent; B2B AI
Index Ventures	Frontier + applications	Ineffable Intelligence; broad AI portfolio
DST Global	Late-stage frontier	Ineffable; growth-stage AI
Kleiner Perkins	Vertical AI agents	Avoca AI $1B (Texas-rooted vertical agent for trades)
Google Ventures (GV)	Strategic AI	Multiple frontier and vertical bets
Craft Ventures	Visual AI, infrastructure	Led ComfyUI $30M (April 2026)
UK Sovereign AI Fund	National AI priorities	Ineffable Intelligence (April 2026)
UAE MGX	Strategic + frontier	Active across frontier labs
Saudi PIF / Saudi VCs	Strategic AI, infra	Active in compute and labs

What’s getting funded (May 2026 categories)

1. Frontier research labs ($300M-$1.1B at concept)

Prestige bets on superintelligence and frontier capability:

Ineffable Intelligence — $1.1B seed (April 27 2026), DeepMind alum founder, Sequoia + Lightspeed co-led
Mira Murati’s Thinking Machines — $2B+ raised since founding
Ilya Sutskever’s SSI — multi-billion at frontier valuations

Pattern: founder pedigree (DeepMind, OpenAI, Meta FAIR alumni), explicit superintelligence thesis, sovereign capital participation, no immediate product.

2. Vertical AI agents ($50M-$1B at growth stage)

Industry-specific autonomous agents replacing tier-1 work:

Decagon — $4.5B tender (March 2026), AI customer support, 100+ enterprise customers
Avoca AI — $1B (April 2026), Kleiner Perkins-backed, missed-call recovery for HVAC/plumbing/roofing trades
Sierra (Bret Taylor) — multi-billion, AI customer experience
Cresta, Forethought — active growth-stage rounds

Pattern: clear vertical, enterprise customers, replacing labor, autonomous (not chatbot) deflection.

3. AI coding tools ($100M-$10B at growth stage)

The coding agent category is having a banner 2026:

Cursor — multi-billion at frontier coding IDE valuations
Windsurf — solid growth at $1-5B range
Augment Code — enterprise-focused coding agent rounds

Pattern: developer love, enterprise expansion, frontier-model partnerships.

4. AI legal / regulated industries ($500M-$5.6B)

Legora’s $5.6B Nvidia-backed round (April 30 2026) is the headline. Pattern: high-stakes industries (legal, healthcare, finance), AI as augmentation not replacement, large ACVs ($100K-$2M).

5. AI infrastructure ($30M-$1B+)

ComfyUI ($30M led by Craft, April 2026) is a recent example. Pattern: developer tools + workflow + open source distribution + enterprise monetization layer.

What investors look for in May 2026

Stage	What they want
Pre-seed	Pedigree (Big Tech AI background) OR working prototype with usage
Seed	$100K-$500K MRR OR strong demo with frontier-grade output
Series A	$1-5M ARR, 100+ enterprise customers (Decagon bar), defensible product
Series B+	$10M+ ARR, clear path to $100M+, strong gross margins (60%+)
Growth	$50M+ ARR, category-leading, viable IPO or strategic exit path

Application-layer AI bar (most relevant for most founders)

Per Mean CEO blog’s analysis (May 2 2026), the typical Series A bar in May 2026:

100+ paying customers with retention >90%
$1-5M ARR growing 3-5x year-over-year
AI-native moat — not just an OpenAI API wrapper
Unit economics path — frontier model costs <25% of revenue at scale
Founder + team capable of executing through capability shifts (model upgrades, regulatory changes)

What’s NOT getting funded

A few patterns that stopped working in 2026:

Generic chatbot wrappers — too easy to replicate, no defensible moat
“AI for [industry]” without enterprise customers — investors want proof
Open-source AI without commercial layer — capital follows monetization
Late-stage AI without defensibility — capable models commoditize fast; investors want product moats beyond model quality

How to approach AI VCs

Practical advice for founders raising in May 2026:

Lead with revenue or pedigree, not vision. Investors have heard every superintelligence pitch. They’re looking for either real revenue or proven research talent.
Use Decagon, Avoca, Legora as comparables. Cite specific recent deals when explaining valuation expectations and exit paths.
Don’t hide your AI stack. Investors know you use Sonnet 4.7 / GPT-5.5 / Gemini 3.1 Pro. The interesting question is what proprietary layer you build on top.
Plan for tender offers, not IPOs. Like Decagon’s $4.5B private liquidity event, plan for $1-2B and $5-10B tender offers before considering an IPO. The IPO window is uncertain.
Talk to sovereign funds. UK, UAE, Saudi, France, Singapore — these are now first-tier early-stage participants. Don’t ignore them.

Notable themes for May 2026

Ex-Big-Tech founder labs dominate prestige funding. Meta, Google, OpenAI alums are getting $200M-$1.1B seed rounds.
Vertical agents are the application story. Decagon, Avoca, Sierra, Cresta — autonomous vertical agents replacing labor.
Sovereign AI funds are capital-significant. Treat them as first-tier investors, not strategic add-ons.
Private liquidity events substitute for IPOs. Plan for tender offers at $1B, $5B, $10B before public markets.
Compute access matters. NVentures’ GPU allocation is often worth more than the cash.

Bottom line

AI VC in May 2026 is hyper-bifurcated: $1B+ seeds for prestige labs, $1-5M seeds for everyone else with a hard application-layer Series A bar. The most active investors are Sequoia, Lightspeed, NVentures, Kleiner Perkins, and the sovereign AI funds. The pattern is clear: pedigree or revenue, real moat, plan for private liquidity events, treat sovereign capital as first-tier.

Sources: CNBC “Big Tech firms seeing top staff leaving to launch AI startups” (April 28 2026), CNBC “Former Google DeepMind researcher’s AI startup raises record $1.1 billion” (April 27 2026), Fortune “Avoca AI Kleiner Perkins” (April 27 2026), CNBC “Nvidia backs Legora” (April 30 2026), Dealroom AI startup funding data 2026, Mean CEO blog AI funding analysis May 2026, TechCrunch and TechFundingNews Decagon coverage March 2026, Qubit Capital “Top AI Investors & VC Firms 2026.”