Anthropic's Two-Trillion-Dollar Ambition Faces Its Sternest Test Yet

Anthropic's potential $2T IPO hits political and financial hurdles, with revenue soaring but losses mounting and OpenAI's retreat adding uncertainty.

The road to what could become the largest initial public offering in modern market history is getting bumpier by the day. Anthropic, the artificial intelligence lab backed by Amazon and a constellation of blue-chip investors, is preparing to list on the Nasdaq with aspirations of a valuation north of $2 trillion — a figure that would dwarf the benchmark set by SpaceX, which raised $86.2 billion in June at a $1.78 trillion valuation. Yet the company finds itself navigating a thicket of political friction, eye-watering losses, and a rival's sudden retreat from the public markets.

OpenAI's decision to pause its own IPO plans has injected fresh uncertainty into Anthropic's timetable, according to market observers. The company filed its paperwork confidentially with the SEC back in June, and a public filing could land as early as late August. But a firm date remains elusive, with reports suggesting an October window that sources describe as little more than an unconfirmed rumor.

The Numbers Tell Two Stories

The financial picture fueling this saga is genuinely split-screen. On one hand, the growth trajectory is almost without precedent. Annualized revenue has rocketed from roughly $9 billion at the end of 2025 to $47 billion in May, and then to $65 billion by the end of July. Second-quarter 2026 revenue topped $11.5 billion — more than fourteen times the $787 million booked in the same period a year earlier, and comfortably ahead of OpenAI's $6.7 billion. Reuters projects 2028 revenue could land between $190 billion and $200 billion.

The company also notched a milestone that changes the conversation: its first positive adjusted operating result in the second quarter, a stark contrast to OpenAI's $12.3 billion operating loss over the same stretch.

But the other side of the ledger is sobering. Anthropic posted a net loss of nearly $42 billion for 2025, up from $8.3 billion the year before. NYU's Aswath Damodaran, the valuation scholar known as the "Dean of Valuation," has crunched the numbers and arrived at a daunting conclusion: to justify a $2 trillion price tag, Anthropic would need annual revenue of roughly $1.2 trillion within a decade — 18 times its current run-rate and more than one and a half times Amazon's entire 2025 turnover of $716.9 billion.

Secondary Markets Already Voting

Investors haven't waited for the official listing to price the company. On secondary markets, Anthropic shares changed hands last week at a valuation of $1.5 trillion — a 25 percent jump in a single month and a substantial premium to the $965 billion valuation set during the Series H round in May, when the company raised $65 billion from Altimeter, Dragoneer, Greenoaks, and Sequoia, with Amazon chipping in $5 billion.

That momentum got an additional jolt on Monday when the preliminary quarterly figures leaked into the market. The news was enough to lift semiconductor stocks by one percent, as traders bet on surging demand for AI infrastructure.

Infrastructure Bets of Historic Proportions

The capital requirements behind this growth are staggering in their own right. Anthropic has committed to spending more than $100 billion with Amazon Web Services over the next decade, while Amazon is investing up to $33 billion in the company. Broadcom is in talks to take on over $60 billion in debt — with Blackstone and Apollo involved — to finance chip capacity aimed at delivering more than 20 gigawatts of computing power by 2028. There's also a monthly tab of roughly $1.25 billion to SpaceX for compute, an arrangement running through May 2029.

The company is also shoring up its balance sheet ahead of the listing, with its credit line reportedly set to expand beyond $10 billion to fund the insatiable demand for computing capacity.

Politics and Product Launches

Friday brought a White House meeting between CEO Dario Amodei, Treasury Secretary Scott Bessent, and chief of staff Susie Wiles — described by observers as "productive" — following the release of the Claude Mythos model preview. The engagement comes against a backdrop of friction: the Trump administration has directed federal agencies to steer clear of Anthropic products.

The company is also keeping its most capable internal model, referred to as "Model 2," out of public release, using it exclusively for coding tasks and synthetic data generation, according to insiders.

Preparing the Corporate Structure

Anthropic is leaving nothing to chance on the governance front. Reports from Bloomberg and The Information indicate the company plans a multi-class share structure granting founders enhanced voting rights, designed to preserve control over its AI safety protocols after the listing. Morgan Stanley, Goldman Sachs, and JPMorgan are the banks expected to lead the underwriting.

Operationally, the company is expanding its enterprise footprint. On Thursday, it launched "Ode," a standalone unit developed with Blackstone, Hellman & Friedman, Goldman Sachs, GIC, and Sequoia to help businesses implement Claude-driven transformations. Ode has already absorbed the AI services firm Casper Studios. Claude Code has been upgraded into an always-on, event-driven agent system capable of managing long-running software development tasks across devices. And to ease enterprise concerns, Anthropic is revising its data retention policies — a June rule allowing marked prompts to be stored for up to two years had drawn criticism.

Infrastructure partnerships continue to multiply: a new platform called "Theseus Infrastructure" with Macquarie Asset Management and GIC will focus on large-scale data center capacity, while a Wednesday deal with British startup Fractile covers roughly $250 million in AI inference chips, with room for expansion.

For prospective investors, the calculus is anything but simple. A record-shattering valuation, historic losses, political headwinds, and a rival's stalled IPO all point in different directions. The one certainty is that the timeline remains very much in flux.