Anthropic’s $2 Trillion IPO Bid: Can It Top SpaceX’s Record?

Anthropic filed a confidential draft S-1 with the SEC on June 1, 2026 and is preparing an IPO whose size could surpass the record set by SpaceX just three months earlier. At the center of investor models: a valuation north of $2 trillion that, if confirmed, would make the creator of Claude one of the most valuable companies ever to list on day one.

SpaceX set the bar everyone is now racing to clear

To gauge what Anthropic is aiming for, the benchmark SpaceX set on June 11, 2026 must be calibrated first. Elon Musk’s group priced 555.6 million Class A shares at $135, raising $75 billion at an implied market capitalization of $1.77 trillion. With the overallotment option exercised, total proceeds reached $86.2 billion. SPCX closed its first session at $161 — up 19% — briefly minting Musk as the first US-dollar trillionaire in history.

The feat looks all the more striking given SpaceX posted a $4.28 billion net loss in Q1 2026 alone, with an accumulated deficit of $41.3 billion. The prospectus leaned on a $28.5 trillion total addressable market — described as the largest in human history — of which roughly 90% was attributed to AI through the recently acquired xAI subsidiary. Investor demand peaked above $250 billion, or 3.5 to 4 times the amount sought.

The epilogue, however, cooled the enthusiasm. By late July 2026, SPCX had shed roughly 50% from its post-IPO peak, slipping back below $135 on a routine share unlock. Musk’s trillionaire status evaporated inside two months. The cautionary tale was inscribed before Anthropic even set its price.

A confidential S-1 and a $965B funding round

Anthropic confirmed the filing of its draft S-1 on June 1, 2026, days after closing a $65 billion Series H at a $965 billion post-money valuation — already the highest ever reached by a private AI company, ahead of OpenAI’s $852 billion mark from March 2026. The round was led by Altimeter Capital, Dragoneer, Greenoaks and Sequoia, with participation from Amazon, Google-related entities and infrastructure partners.

Since roughly August 13, CFO Krishna Rao has been leading early investor meetings, according to sources cited by CNBC. Notably, Rao has explicitly avoided discussing specific valuation targets, choosing to present Claude’s product roadmap, Claude Code and the management team instead. A rare discipline in a market where leaked figures usually pre-empt pricing. The implicit message: let the market discover the price.

« This gives us the option to go public after the SEC completes its review. The proposed initial public offering will depend on market conditions and other factors. »

Anthropic, official statement, June 1, 2026

Bloomberg reported in August 2026 that Anthropic expected its IPO to match or exceed SpaceX’s record, with a public filing possible as early as late August. The timeline is indicative — confidential filings routinely shift by weeks. But the directional intent is unmistakable.

A revenue trajectory that resets the benchmarks

Beneath the valuation chatter, the underlying financial dynamic is, by any measure, without recent equivalent. Anthropic’s annualized revenue run rate hit $65 billion in July 2026, per company disclosures. The path over thirty months is vertiginous: $87 million in January 2024, $1 billion by December 2024, $9 billion by end-2025, $14 billion in February 2026, $19 billion in March, $30 billion in April, $47 billion in May, and $65 billion in July. A 750x increase in roughly two and a half years.

Enterprise and API usage drive the bulk of that growth. As of October 2025, Anthropic had more than 300,000 business customers accounting for roughly 80% of revenue. More than 100,000 of those run Claude on Amazon Bedrock (April 2026 data). The number of customers spending over $100,000 annually has grown sevenfold in the past year. More than 1,000 customers now spend over $1 million per year, doubling from 500+ in under two months by April 2026. Eight of the Fortune 10 are Anthropic customers.

Claude Code: the product that didn’t exist a year ago

Launched publicly in May 2025, the Claude Code coding assistant already generated $2.5 billion in annualized run-rate revenue by February 2026 — more than doubling from January. Business subscriptions quadrupled over the same period. According to recent data, Claude Code now authors roughly 4% of all public GitHub commits, with projections above 20% by year-end. In under a year, the product pulls in more revenue than most public SaaS companies ever will.

Gross margin is approaching breakeven

The most underappreciated evolution in the file concerns unit economics. Anthropic’s gross margin has moved from roughly -94% in 2024 to 44-60% in 2026, depending on the quarter and whether amortization of training costs is netted against compute spend. The improvement rests on a single ratio: compute cost per revenue dollar, which fell from $0.71 in Q1 2026 to $0.56 projected in Q2 2026, according to Yahoo Finance.

« Every dollar of Claude usage is becoming cheaper to serve as inference costs fall and model utilization improves. »

SaaStr analyst, 2026

Anthropic reported positive adjusted operating income for Q2 2026. The picture must be qualified, however: the GAAP net loss for 2025 reached $42 billion, five times the $8.3 billion loss in 2024. The gap between adjusted operating income and GAAP net loss is explained primarily by massive non-cash charges and stock-based compensation — common in hyper-growth private companies but certain to face intense scrutiny once full GAAP disclosures become mandatory.

The structural gap with OpenAI

The OpenAI comparison is less balanced than it appears. Internal documents at the rival project a $14 billion loss for 2026, with over $1 trillion committed to infrastructure and positive free cash flow not expected until 2029. Anthropic, with a projected 2026 annual loss of roughly $2 billion, burns a fraction of that. Anthropic’s peak training cost of about $30 billion runs roughly four times lower than OpenAI’s, per SaaStr. That is a structural cost advantage, not a mere operational differential.

Is the $2 trillion justifiable?

At $65 billion in annualized revenue, a $2 trillion valuation implies a multiple of roughly 31x revenue. At $100-120 billion projected for full-year 2026, that multiple compresses to between 17x and 20x — elevated by historical standards, but not absurd for a company growing at this pace. Nvidia historically traded at 30-40x earnings during AI booms; Microsoft at peak multiples, and — in a stretched comparison — early Amazon and Google, serve as reference points. None of them, however, served as the AI inference backbone for a significant portion of global enterprise software.

The bull case is coherent: if AI genuinely becomes the operating system of the global economy, and if Claude is the productivity layer enterprises build on, then a company multiplying revenues 7x per year while improving unit economics simultaneously deserves a premium multiple. Enterprise network effects — every new Fortune 500 customer adopting Claude makes the product more indispensable, generates more training signal, and deepens switching costs — create the kind of moat public market investors pay up for.

The bear case is equally coherent: $65 billion in annualized revenue run rate is not the same as $65 billion in cash collected. Enterprise contracts remain cancellable. Usage-based pricing can be lumpy. The $42 billion net loss in 2025 — even partially explained by non-cash items — describes a company still fundamentally burning cash at scale. And compute costs that fall today may not fall fast enough to offset the next generation of model training, which is exponentially more expensive than the last.

The DoD regulatory overhang

Anthropic has already clashed with the Department of Defense over the use of its models, leading to a Pentagon blacklist the company is contesting in court. A DoD ban on a major AI provider is not a routine regulatory headache — it is a fundamental question about where Anthropic’s models can be deployed at the largest institutions in the world. President Trump has said a deal is possible, but the litigation is ongoing and the outcome uncertain.

2026 is already the biggest IPO year since 2000

The macroeconomic context amplifies the shockwave. Companies have raised roughly $160.6 billion in IPO proceeds through August 19, 2026 — within reach of the $175 billion full-year record set in 2021, and that is before Anthropic’s own listing is even counted. If Anthropic clears $86 billion, the year moves into genuinely uncharted territory.

California startups have raised approximately $366 billion so far in 2026, according to PitchBook data cited by the Wall Street Journal — more than three times the total for the other 49 US states combined. OpenAI and Anthropic together account for roughly half of those dollars. Behind Anthropic, the pipeline remains substantial: Databricks wrapped a $5 billion round at a $190 billion valuation in August 2026; Broadcom is reportedly raising a debt package of $60 billion to nearly $100 billion to fund AI infrastructure buildout.

The SpaceX preview: what to expect post-IPO

SpaceX offers a useful — and somewhat cautionary — preview of what Anthropic’s public shareholders might expect. The roughly 50% drop from the post-IPO peak within weeks of listing is a reminder that matching a record-setting raise on day one says nothing about where a stock trades ninety days later. The 366-day lock-up for Musk and insiders, and 180-day staggered lock-ups for other pre-IPO investors, mean the float will expand dramatically at expiry — and the supply shock that creates is not theoretical. Block sizes and their clearing prices will test whether the demand that drove the IPO pricing is durable or speculative.

Conclusion: a verdict that will be rendered on the roadshow

Anthropic’s IPO, if it materializes at the size and valuation currently being modeled by some investors, will be a defining moment for the AI investment thesis. It will answer questions the private markets have been unable to resolve: can an AI company multiply revenues 7x per year while improving unit economics? Can that growth hold as competition from OpenAI, Google, Meta and Amazon intensifies? And can it absorb the regulatory fire — from the DoD blacklist to broader AI governance debates — without a meaningful impact on enterprise adoption?

The revenue numbers are real and extraordinary. The path to profitability is narrowing. The competitive moat, measured by enterprise adoption and developer share, is deepening. Whether that justifies a $2 trillion day-one valuation is a question only the roadshow — and the first few quarters of public trading — can answer. One thing is already certain: when Anthropic files publicly, the world will be watching.

Sources

• CNBC — Anthropic confidentially files IPO prospectus with SEC (June 1, 2026)
• Value Add Pulse — Anthropic Eyes IPO Big Enough to Beat SpaceX (August 2026)
• Wikipedia — Initial public offering of SpaceX (2026)
• Binance Square — Anthropic IPO Could Break SpaceX’s Record (2026)
• Sacra — Anthropic revenue, valuation & funding (2026)
• SaaStr — Anthropic Just Passed OpenAI in Revenue (2026)
• Value Add VC — Is Anthropic Profitable? 2026 Losses, Burn Rate (2026)
• Yahoo Finance / CNBC — Anthropic compute cost per revenue dollar reporting (Q1-Q2 2026)

This article is published for informational and educational purposes only. It does not constitute investment advice. Do your own research (DYOR) before making any decision.