Which AI company has the highest revenue multiple in June 2026?

Cognition is the clear outlier. At $26B valuation on roughly $492M annualized revenue, Cognition trades at ~53x ARR. Anthropic ($965B / ~$47B ARR) trades at ~20x. OpenAI ($852B tracking / ~$60B ARR) trades at ~14x. Cognition's premium reflects growth rate (1,230% YoY reported) more than absolute scale.

Is Cognition's $26B valuation justified vs Anthropic and OpenAI?

On growth, yes — Cognition's 1,230% YoY ARR growth is roughly 3-4x faster than Anthropic's and 5x faster than OpenAI's. On durability, less clear — Cognition is a single-product company (Devin) competing directly with Cursor, Claude Code, GitHub Copilot, and Windsurf/Devin Desktop, while Anthropic and OpenAI are platform companies. The multiple is a growth bet, not a moat bet.

Which of these would be the safest investment if all three IPO'd?

OpenAI looks safest on the math: largest absolute revenue, lowest multiple, broadest distribution (ChatGPT consumer + API + Microsoft channel). Anthropic is the highest quality on margins and enterprise mix but most expensive in absolute terms. Cognition is the highest-variance bet — could 10x if the autonomous-agent thesis plays out, or compress hard if Claude Code/Cursor cap its TAM.

How do these multiples compare to historical SaaS peaks?

At the 2021 SaaS peak, top growth names like Snowflake traded around 80-100x forward revenue at IPO; Datadog peaked near 50x. Cognition's ~53x ARR is in that high-growth-SaaS-peak zone. Anthropic's ~20x is closer to mid-cycle SaaS comps. OpenAI's ~14x is the cheapest of the three but still richer than most public software comps today.

Quick Answer

Cognition $26B vs Anthropic $965B vs OpenAI $852B: Multiples

Published: June 5, 2026

Cognition $26B vs Anthropic $965B vs OpenAI $852B: Multiples

June 2026 is the most expensive month in AI funding history. Cognition closed $1B+ at $26B (announced May 27). Anthropic’s Series H hit $965B post-money (closed May). OpenAI is tracking $852B on the secondary market ahead of its confidential IPO filing. Here’s how the revenue multiples actually compare.

Last verified: June 5, 2026

Side-by-side: valuation, revenue, and multiple

Metric	Cognition	Anthropic	OpenAI
Latest valuation	$26B (May 27, 2026)	$965B (May 2026 Series H)	~$852B (secondary tracking)
Reported ARR	~$492M	~$47B	~$60B
ARR multiple	~53x	~20x	~14x
YoY ARR growth	~1,230%	~10x	~6x
Gross margin (est.)	~50-60%	~50%	~45-55%
Lead investors	Lux, General Catalyst, 8VC	Multiple incl. ICONIQ	Multiple incl. Thrive
Product	Devin (autonomous SWE)	Claude (chat + API)	ChatGPT + API
Status	Private	Public IPO filed	Confidential IPO filed

Sources: Cognition press, Anthropic Series H disclosures, OpenAI secondary trading data, public reporting as of June 5, 2026.

What the multiples actually mean

Cognition at 53x: the growth bet

Cognition went from roughly $40M ARR a year ago to ~$492M ARR today — that’s the 1,230% YoY figure widely reported. If you believe that arc continues, $26B is cheap. If growth halves each year (typical for coding-tool S-curves), the implied 2-year-forward multiple is closer to 20x, similar to Anthropic today.

The risk: Cognition is one product (Devin) in a crowded field — Cursor 3, Claude Code, GitHub Copilot X, Windsurf/Devin Desktop, Codex CLI. Every major foundation lab now ships a competing autonomous coding agent. Cognition’s moat is its early enterprise traction and benchmark lead on SWE-bench, not a structural advantage.

Anthropic at 20x: the platform bet

$47B ARR at 20x is a real revenue multiple, not a “vibes” multiple. For context, Microsoft at peak AI hype traded around 13x; NVIDIA peaked around 25x. Anthropic is being valued like a top-tier hyperscaler with embedded software growth.

The premium over OpenAI reflects perceived enterprise quality (Claude Code dominance, financial services and PWC partnerships, Pentagon and Workday deals), margin profile, and the IPO-priced-on-its-own-terms factor. It’s also the IPO market saying “Anthropic’s enterprise mix is more durable than OpenAI’s consumer mix.”

OpenAI at 14x: the scale bet

$60B ARR is roughly 28% larger than Anthropic’s. The 14x multiple reflects two things: heavier consumer exposure (more cyclical), and Microsoft’s commercial entanglement (less clean monetization). But on absolute dollar growth, OpenAI added more revenue last year than Cognition’s entire ARR.

If OpenAI IPOs at 14-16x, that’s the bargain of the trio on conventional metrics — but only if you believe consumer AI subscriptions don’t churn meaningfully.

How these compare to historical IPO multiples

Snowflake (2020 IPO): ~80x forward revenue
Datadog (2019 IPO peak): ~50x revenue
Google (2004 IPO): ~7x revenue
Facebook (2012 IPO): ~28x revenue at debut
NVIDIA (peak 2024): ~25x revenue

Cognition’s 53x is in the Datadog-peak-zone. Anthropic’s 20x is below Facebook’s IPO multiple. OpenAI’s 14x is well below Facebook IPO — closer to a mature large-cap multiple.

What this tells you about the AI funding cycle

The market is bifurcating. Platform AI is being valued at growth-SaaS multiples. Single-product agents are being valued at peak-cycle SaaS multiples. There’s no premium for “AI generally” anymore.
Revenue quality matters more than revenue volume. Anthropic’s premium over OpenAI per dollar of ARR reflects enterprise mix and margin durability.
Cognition is the marginal-investor test. If the next round prices flat or down, the autonomous-agent thesis is in trouble. If it prices up at $40B+, the growth bet holds.
IPOs will compress these multiples. Public markets typically apply a 20-30% discount to private peak. Cognition at 53x in IPO would be more like 35-40x; Anthropic at 20x would price closer to 14-16x.

Bottom line

If you have to pick one based purely on revenue multiples: OpenAI at ~14x is the cheapest and largest of the three. Anthropic at ~20x is the premium platform bet. Cognition at ~53x is the highest-conviction growth bet but also the highest-variance — closer to a venture position than a public-market position, even at $26B.

The June 2026 funding wave isn’t an “AI bubble” reading on these multiples. It’s a sorting event: the market is pricing platforms differently from products differently from growth bets, and revenue multiples now tell you which is which.