Anthropic's IPO prospectus contains a warning that reads more like science fiction than securities law. The company tells investors that advanced AI could pose "catastrophic or existential risks to humanity," including models that exhibit "self-preserving behaviours" such as resisting shutdown, concealing information, and behaviour "resembling blackmail". Approximately 80 of the 261 pages in the main body are devoted to risk factors, compared with 48 pages describing the business.
This isn't just legal boilerplate. The warning follows internal turmoil, including a researcher's resignation claiming people building AI "earnestly believe that it could kill us all by the end of the decade," and a senior safety researcher agreeing there's more than a 10% chance of human extinction within a decade. CEO Dario Amodei has publicly called for the industry to slow the pace of AI development.
For investors, this creates a unique risk perception challenge. Traditional tech IPOs worry about competition, regulation, and execution. Anthropic adds tail-risk scenarios that are difficult to quantify. Some experts criticize these existential warnings as unverifiable and unscientific, but they're now part of the official disclosure package. This likely pushes some institutional investors to demand higher risk premiums or simply sit out, while others may view the candor as a governance positive.
The most eye-popping number in Anthropic's S-1 isn't the potential $2 trillion valuation—it's the compute commitment. Anthropic has agreed to pay SpaceXAI up to $120.5 billion through 2029 for access to NVIDIA GPU-powered computing capacity. Monthly payments could reach approximately $3.1 billion once all contracted capacity is delivered by December 2026.
This represents a massive escalation from earlier estimates. Confidential prospectus details in September 2026 pegged the commitment at $84.5 billion, while SpaceX's own May 2026 S-1 referenced roughly $1.25 billion per month. The jump reflects expanded capacity orders and the shift from Colossus 1 to the much larger Colossus II facility.
The strategic logic is clear: Anthropic needs massive compute capacity immediately, and building it would take years. SpaceX's Colossus 1 facility houses over 220,000 NVIDIA GPUs drawing more than 300 megawatts of power. Anthropic gets access within weeks rather than years. But the financial implications are staggering. Anthropic reported $16.8 billion in revenue last quarter, or about $5.6 billion monthly. At $3.1 billion monthly, SpaceXAI payments would consume over half of that revenue.
The contract structure provides some protection. Non-cancelable periods end in January and March 2027, after which either party can terminate with 90 days' notice. This means Anthropic isn't locked into the full $120.5 billion if business conditions deteriorate. But the initial commitment creates significant fixed-cost pressure on margins and cash flow.
While Anthropic prepares for its blockbuster IPO, traditional IT services companies are feeling the pain.
Infosys fell nearly 3%, while HCL Technologies, Mphasis, and Persistent Systems dropped more than 2% each. The Nifty IT index plunged over 5% that session.
The causal chain is straightforward. Indian IT companies depend heavily on entry-level talent for routine coding, testing, and support work—exactly the tasks AI agents can now automate. COBOL modernization, a key revenue driver for firms like IBM and their Indian partners, is directly in Anthropic's crosshairs. An estimated 95% of ATM transactions in the US still rely on COBOL, representing a massive market that AI could disrupt.
This creates a divergence in market reaction. Global AI stocks might rally on Anthropic's growth prospects, but traditional IT services face structural threats. Jefferies has already downgraded several IT stocks, warning that AI may structurally shift the sector toward consulting and implementation work while shrinking managed services. This shift would increase cyclicality and require new operating and talent models.
Not everyone sees doom. CLSA argues that fears of AI-led disruption appear overdone, with channel checks showing no material change in client spending behavior. Some analysts believe AI could expand total addressable markets by lowering execution costs, and that Indian IT firms can reposition as AI orchestrators rather than manpower suppliers. But the valuation pressure is real.
Anthropic's risk disclosures don't exist in a vacuum—they're part of an evolving regulatory narrative. The company has actively supported state AI transparency laws, arguing that "transparency alone is no longer sufficient" and that governments need a more substantial role. It opposes federal preemption unless a federal law is at least as strong as state versions.
This creates potential feedback loops. Anthropic's public acknowledgment of existential risks gives regulators and plaintiffs a reference point for liability claims. The filing notes that lawmakers have won hefty settlements against tech companies on "analogous theories of product-design or algorithmic harm," pointing to Meta's nearly $17 billion settlement with state attorneys general. At least one state criminal investigation into another AI company is already underway.
The regulatory landscape is fragmented. The EU AI Act's bans, general-purpose AI rules, and transparency duties already apply, while high-risk rules are delayed to December 2027. The US has no federal AI statute, relying instead on executive action, existing laws, and state legislation. States like California, New York, Illinois, Michigan, and Massachusetts have introduced frontier AI bills.
Anthropic is already feeling regulatory heat. The prospectus reveals that in February, the president ordered federal agencies to stop using Anthropic's models, and the Department of Defense designated the company a supply-chain risk to national security. In June, the Commerce Department imposed worldwide export restrictions on Anthropic's Fable 5 and Mythos 5 models, forcing the company to disable them for all customers. While the restrictions were later lifted, the company warns similar actions could recur.
All these factors converge on one question: Is Anthropic worth $2 trillion? The math is daunting. One analysis suggests that to justify this valuation, Anthropic would need to immediately improve operating margins to 10% and sustain them for a decade, grow revenue by 1000% in 2026 and 100% in 2027, then grow 56% compounded annually through 2035. In this scenario, Anthropic's revenue would reach $3.6 trillion by 2035—1.4 times the combined revenue of Microsoft, Apple, Alphabet, Amazon, NVIDIA, Meta, and Oracle.
Even with more conservative assumptions, the downside is significant. If margins improve to 20% and revenue growth moderates, Anthropic's equity might be worth just $1.1 trillion—a 43% downside to the targeted IPO valuation. This scenario still implies $1.2 trillion in revenue and $242 billion in operating profit by 2035, ranking first across all companies covered by the analysis.
The company's financials show both promise and peril. Net revenue retention stands at around 900%, meaning existing customers increase their spending ninefold year-over-year. But the company lost over $80 billion in the first half of 2026 on $16 billion in revenue. The path to profitability requires massive scale and continued pricing power.
Anthropic's IPO will be a watershed moment for the AI industry. The existential risk disclosures, unprecedented compute commitments, and competitive threats to traditional IT services all point to a sector in transformation. Institutional investors will need to decide whether to price in tail risks that have no historical precedent, while regulators will face pressure to act on warnings from the industry itself.
The company's Long-Term Benefit Trust, which has rights to appoint board members, adds another governance layer. Public shareholders may have limited influence over company decisions, depending on voting rights attached to IPO shares. This structure, designed to protect long-term safety goals, could conflict with short-term shareholder interests.
For now, Anthropic is betting that the demand for AI capabilities will outpace the risks. The $120.5 billion SpaceXAI commitment signals confidence in massive growth. But as one analyst noted, private investors and Wall Street insiders have already loaded up on equity at valuations as low as $623 million. They now need public markets to deliver their payday. Whether retail investors share that optimism remains to be seen.
