The entire AI ecosystem is now credit exposure to two private labs.
With hyperscaler AI revenue concentrated in OpenAI and Anthropic, the ecosystem's $250B purchase promises, vendor equity stakes, and IPO pricing all function as credit exposure to whether two barely-audited labs can generate the cash they committed to spend.
The same conclusion keeps arriving from across the workspace's research — 2 topics independently instantiate this theme. Filter the evidence by where it came from:
Anthropic's guaranteed lease obligations now span both silicon and the data centers housing it, converting chip-layer sales into credit exposure to whether the lab can generate the cash it promised to spend.
The chipmaker's own receivables book concentrates the ecosystem's credit exposure in a handful of lab customers, making NVIDIA's balance sheet a levered claim on a few labs' solvency.
Eisman's 70%-concentration math restates the ecosystem's dependence on OpenAI and Anthropic as a systemic credit position.
The ecosystem functions as credit exposure to two labs, and the first hard print shows one lab's cash generation doubling into its $518B commitment stack.
The leaked S-1 showing Anthropic spending $2.75 to make a dollar hardens the ecosystem's credit exposure to whether two unprofitable labs can generate the cash they committed to spend.
The first audited look at one of the two load-bearing labs shows a widening $8.06B operating loss, two-customer revenue concentration, and 90-day cancellation clauses carrying a half-trillion-dollar commitment stack.