Anthropic Files for IPO, Eyes $1 Trillion Debut

Anthropic went from a $61.5 billion private company to a near-$1 trillion IPO filer in 15 months, and it's now asking Wall Street whether the AI safety lab turned revenue machine deserves a public market at all.

On June 1, 2026, Anthropic confirmed it had confidentially submitted a draft registration statement on Form S-1 to the U.S. Securities and Exchange Commission for a proposed initial public offering of its common stock. No share count or price has been set. A potential October 2026 debut is the current target, according to reporting from CNBC, NPR, and TechCrunch.

TL;DR

Filed confidential S-1 with SEC on June 1, targeting an October 2026 debut
Valuation at filing: $965B - ahead of OpenAI's $852B pre-IPO mark
Revenue run rate reached $47B in May 2026, up from $10B at end of 2025
Q2 2026 showed first operating profit: $559M at a 5% margin
Three mega-listings now in motion: Anthropic (Oct), OpenAI (Sep), SpaceX (Jun)

The Trajectory in Numbers

The speed of Anthropic's ascent has no clean analogue in tech history. In March 2025 the company closed a $3.5 billion Series E at $61.5 billion. Fifteen months later it's filing for an IPO at roughly 16 times that valuation.

Date	Round	Valuation	ARR
Mar 2025	Series E	$61.5B	~$1B
Sep 2025	Series F	$183B	~$4B
Feb 2026	Series G	$380B	~$15B
May 2026	Series H	$965B	$47B
Jun 2026	S-1 filed	-	$47B (run rate)

That table tells an unusual story: valuation and revenue scaled together. Most AI companies have ballooned on hype with revenue lagging far behind. Anthropic's revenue run rate grew roughly 4.7x in six months - from about $10 billion at the end of 2025 to $47 billion by May 2026. Its valuation grew 2.5x over the same window. The multiple actually compressed as revenue outpaced paper value.

What drove each inflection

The Series F to Series G jump - $183B to $380B in five months - traces back almost entirely to Claude Code. Anthropic's developer product crossed $2.5 billion in ARR by February 2026, making it one of the fastest-growing developer tools in history. Enterprise deployments through Amazon Bedrock and Google Cloud compounded that: customers were booking compute before Anthropic could provision it.

The Series H closed at $965 billion in late May, co-led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia. It was the largest private funding round ever. Within two weeks of closing, the company was filing for an IPO. The message to investors: the private market has done its job, and the public market is the logical next stop.

The Golden Gate Bridge with San Francisco skyline in the background Anthropic is headquartered in San Francisco and has signed leases passing 1 million square feet in the city's "AI Alley" corridor along Howard Street. Source: unsplash.com

The Race to the Bell

Anthropic's filing doesn't exist in isolation. It's the second move in a three-company sprint to the public markets.

OpenAI filed its own confidential S-1 on May 22, 2026 - ten days earlier. Goldman Sachs and Morgan Stanley are leading that deal. OpenAI is targeting a September listing at a valuation above $1 trillion. Its revenue run rate sits at roughly $25 billion annualized, and it's losing an estimated $1.22 for every dollar of revenue it brings in - with no path to profitability projected before 2030.

SpaceX is moving even faster. The company is expected to launch its IPO roadshow on June 4, targeting roughly $75 billion in proceeds at a valuation around $1.75 trillion. If it hits that number, it would surpass Saudi Aramco's 2019 listing as the world's largest IPO.

Combined, the three companies represent roughly $3.6 trillion in claimed market value heading toward public markets in a single calendar year. Goldman Sachs projected earlier in 2026 that total U.S. IPO proceeds could reach a record $160 billion if the marquee names actually listed. That number now looks conservative.

"The company is experiencing strong demand for its AI solutions and services from enterprises and consumers, at levels that are exceeding the company's available supply."

That quote is from Alphabet, which announced on June 2 that it's raising $80 billion in equity - its first stock issuance since 2005 - partly to keep up with AI infrastructure demand. Berkshire Hathaway is putting in $10 billion. The same supply-constraint language could apply equally to Anthropic: demand is running ahead of capacity everywhere in AI infrastructure right now.

A financial trading screen showing market data and candlestick charts Public markets are now pricing AI companies for the first time, with three mega-listings expected before the end of 2026. Source: pexels.com

The Valuation Gap

At $47 billion in annual run-rate revenue, Anthropic's $965 billion pre-IPO valuation implies a price-to-sales multiple of roughly 20.5x. That sounds large, but it's lower than OpenAI's 34x multiple on $25 billion in revenue.

Company	Pre-IPO Valuation	ARR	P/S Multiple	Profitable?
Anthropic	$965B	$47B	~20.5x	Q2 2026 (5% margin)
OpenAI	$852B	$25B	~34x	No (2030 target)
SpaceX	~$1.75T	~$10B*	~175x	Yes (aerospace)

*SpaceX revenue excludes Starlink (estimated separately).

The relative cheapness of Anthropic on a revenue basis reflects one thing more than anything else: investors believe its revenue growth is real and defensible. A $47 billion revenue run rate in May 2026 started the year at $10 billion. The company isn't making up usage - Claude Code, Claude API enterprise contracts, and consumer subscriptions are all driving measured recurring revenue.

The loss picture is more complicated. Anthropic burned through an estimated $14 billion in 2026 against those revenues. Training runs for Claude Opus 4.8 and the restricted-access Mythos model don't come cheap. The company's commitment to supply Amazon, Google, and Microsoft with compute through 2029 creates long-term obligations regardless of revenue trajectory. Cumulative losses since founding are approaching $25 billion.

But Q2 2026 showed something new: a $559 million operating profit at a 5% margin - the first in the company's history, as we reported in May. That's a thin margin against a $14 billion annual burn projection, but the direction is the story. OpenAI doesn't have a Q2 like that to show investors.

Priced to Perfection - or Priced for Pain

The base case for Anthropic at $965 billion is straightforward: 5x revenue growth in six months, first operating profit, revenue above OpenAI, and a clear path to positive free cash flow by 2027 - not 2030. Claude remains the dominant enterprise AI product for reasoning-intensive workflows. Claude Opus 4.8 scores 69.2% on SWE-Bench Pro, and enterprise customers are paying $5 to $25 per million tokens for that. The Series H closed at $965B against hard revenue numbers, not projections.

The case against is also clear. A 20.5x revenue multiple at $47 billion ARR means Anthropic's implied market cap at IPO would be roughly $1 trillion. It needs to grow into that. The training and inference compute bill is truly enormous - $14 billion in losses this year, with $80 billion in infrastructure commitments to cloud partners stretching to 2029. And the competitive environment isn't standing still: DeepSeek's distillation success, OpenAI's model releases, and Google's Gemini ecosystem all pull at Anthropic's enterprise share.

The question public markets will actually price is simpler: does $47 billion in ARR, growing at 5x per year, justify $1 trillion? At Anthropic's current arc, it'd cross $200 billion in annual revenue sometime in 2027. That's the bet. Whether the 2027 revenue actually materializes at that scale - or whether compute costs eat the margin and growth normalizes - is what the S-1 will have to answer when it goes public later this year.

Sources: