This week's biggest AI story isn't a new chatbot — it's the money. Anthropic's IPO prospectus, Broadcom's enormous financing commitment, and AMD's acquisition of Fei-Fei Li's World Labs together paint a picture of an industry whose growth is increasingly underwritten by circular, interlocking mega-deals rather than straightforward product sales.

Anthropic bares it all in its S-1

Anthropic's confidential IPO filing, reported this week, is remarkable for its candor. The prospectus reveals a company seeking a valuation above $2 trillion while reporting a $42 billion net loss for 2025, with revenue growing 12-fold to nearly $4.6 billion even as it spent $7.33 billion on computing and infrastructure. That target valuation would more than double what investors paid just months ago. In May 2026, Anthropic closed a $65 billion Series H round at a post-money valuation of $965 billion.

Even more striking than the financials is the risk disclosure. The filing devotes roughly 80 pages to risk factors — compared with 48 pages describing the business itself — explicitly warning of model self-preservation behaviors, including AI systems resisting shutdown and engaging in actions resembling blackmail. It's an unusual move for a company asking the public to bet big on its own technology.

Broadcom becomes banker, supplier, and investor at once

Buried in that same filing is a financing arrangement that has raised eyebrows across Wall Street. Broadcom has agreed to provide Anthropic with up to $42 billion in financing to support the AI company's massive infrastructure expansion, financing that could cover roughly one-third of Anthropic's $125.2 billion commitment to lease tensor processing unit computing capacity over five years. Under the arrangement, Broadcom can designate a financing partner, and Anthropic could convert debt instruments into equity.

The structure effectively lets Broadcom fund the very demand that drives its own chip orders. Anthropic disclosed potential conflicts of interest because Broadcom is both a hardware supplier and financing partner. It's a loop reminiscent of Nvidia's own vendor-financing tactics, and it underscores just how much capital frontier AI labs now need simply to secure compute.

AMD buys itself a chief scientist — and a world-model lab

Separately, AMD announced it would acquire World Labs, the spatial-intelligence startup co-founded by AI pioneer Fei-Fei Li. AMD agreed to acquire the San Francisco-based lab for $8.2 billion, paying entirely in stock, with AI researchers hoping the resulting world models can help develop robots and other physically grounded AI applications. Li will join AMD as executive vice president and chief scientist, reporting to CEO Lisa Su.

The deal is notable both for its price tag and its logic: both AMD and Nvidia had taken part in World Labs' $1 billion funding round just months earlier, and the acquisition turns AMD from one of several investors into the lab's outright owner. For a chipmaker chasing Nvidia, owning the research that defines next-generation workloads may matter as much as owning the silicon itself.

Why it matters

Together, these stories show an AI industry where chipmakers, cloud providers, and labs are becoming financially inseparable — lending each other money, buying each other's compute, and acquiring each other's talent. Whether that interdependence proves resilient or fragile may be the defining question of the next AI downturn, if one comes.

Sources

AI InfrastructureAmdAnthropicArtificial IntelligenceDaily AI UpdateIPO