Anthropic's $47B Revenue Runway and the $1.25B Monthly Bet on Rented Compute

"It's a really big upfront cost to train the models and to serve inference on them," Amodei said of the capital demands driving the company toward public markets . The implication was clear: enterprise customers are not merely experimenting with Claude — they are committing real budgets to its capabilities.

The IPO Filing and the $965 Billion Valuation

Anthropic formally initiated its journey to the public markets on June 1, 2026, submitting a confidential draft S-1 registration statement to the U.S. Securities and Exchange Commission . Under the SEC's confidential review process, the filing does not yet reveal the number of shares to be offered or the target price range, and full financial disclosures remain private until regulators complete their review .

The IPO move came less than a week after Anthropic closed a $65 billion Series H funding round that valued the company at approximately $965 billion post-money . The round was reportedly greatly oversubscribed, with multiple investors telling TechCrunch that demand far exceeded the amount raised . That private-market enthusiasm now sets the stage for a public debut that could come as early as Fall 2026, should market conditions permit .

The valuation leapfrogged OpenAI's $852 billion figure, making Anthropic the most valuable AI startup in the world — and setting up what would likely be the largest public listing ever for an AI-native company .

Renting Compute Instead of Building Data Centers

One of the most strategically revealing portions of Amodei's appearance concerned infrastructure. While competitors have poured tens of billions into constructing their own data centers, Anthropic has deliberately chosen to rent compute capacity instead .

The reasoning, Amodei explained, comes down to capital efficiency. Building data centers locks billions into physical assets at a moment when the technology stack is evolving rapidly. By leasing capacity, Anthropic retains more financial flexibility — avoiding what she portrayed as an unnecessary bet on specific infrastructure at a time when model architectures, chip generations, and deployment patterns are still shifting .

The $1.25 Billion Monthly xAI Deal

The most extraordinary example of this rent-don't-build philosophy is a deal whose financial terms became public not through Anthropic, but through SpaceX's own IPO filing. Anthropic is paying $1.25 billion per month through May 2029 to access 300 megawatts of compute capacity at xAI's Colossus data center in Memphis, Tennessee .

The agreement gives Anthropic access to the supercomputing facility originally built to train Elon Musk's Grok model — a setup housing roughly 220,000 Nvidia GPUs, including H100, H200, and GB200 processors . At full rate, the total contract value could exceed $45 billion over its full term . A discounted rate applies during May and June 2026 while xAI completes its capacity ramp-up, and either party can terminate the agreement with 90 days' notice .

Musk later pushed back on the framing of a long-term commitment, posting on X that SpaceX has "not committed to leasing Colossus for years" and characterizing the arrangement as essentially a 180-day lease with cancellation rights thereafter . Regardless of the semantic disagreement, the scale of the monthly payments — revealed in SpaceX's S-1 filing on May 20 — makes this the single largest compute lease in AI industry history .

The deal also illuminates a broader market shift. Frontier AI firms are beginning to treat compute infrastructure as a standalone commercial business, with competitors buying capacity from each other rather than exclusively building their own . For xAI, the arrangement provides a way to monetize excess compute capacity amid declining Grok user numbers .

The Enterprise Spending Slowdown Risk

Despite Amodei's confident public posture, the risk of an enterprise AI spending pullback is real and well-documented. Forrester research indicates that many enterprises are postponing approximately 25% of planned AI spending as they struggle to demonstrate clear return on investment . The McKinsey Global AI Survey 2026 found that 73% of enterprise AI deployments fail to achieve their projected ROI — a stubbornly persistent failure rate despite improvements in model capabilities and tooling .

Amodei acknowledged this trend but framed it as a healthy market correction rather than an existential threat to Anthropic's model . Her argument rests on a distinction between the broader enterprise AI market — where pilot programs frequently fail to become production habits — and Anthropic's specific revenue base, which she characterized as sticky enterprise commitments from customers who have already validated Claude's utility .

The company's rapid revenue acceleration from $9 billion to $47 billion annualized in roughly five months lends some weight to that claim, at least in the short term .

A Tale of Two Amodeis: Cautious CEO, Confident President

A notable undercurrent in Anthropic's public messaging is the contrast between Daniela Amodei's bullish tone and her brother Dario's more measured warnings. The CEO has spent much of 2026 sounding alarms about the financial fragility beneath the AI boom. In February, he warned that even a slight miscalculation in growth forecasts could prove "ruinous" and potentially sink the company . He has described the timing of AI revenue as deeply uncertain: "It could be one year. It could be two years. I could even stretch it to five years, although I'm skeptical of that" .

Daniela, by contrast, presented an unambiguously confident face to the Bloomberg audience — brushing off skeptics, pointing to the revenue numbers, and projecting the image of a company whose fundamentals already justify its nearly trillion-dollar valuation. Together, the siblings appear to be running a deliberate two-track communications strategy: Dario warns the industry about excess while Daniela assures the markets about Anthropic's specific trajectory .