SpaceX has signed a binding commitment for about $40 billion in AI chip leases through an Apollo Global Management-managed entity. The leases start in the first half of 2027, but the master agreement and interest rate are still to be finalised.
SpaceX will lease GPUs bought by an entity managed by Apollo Global Management, according to a regulatory filing. The master finance lease is expected to close before the end of 2026, so the binding commitment is not yet a completed lease agreement.
SpaceX absorbed xAI in February and is financing graphics processors used to train and run AI models. The arrangement works like a loan: quarterly payments over roughly 5 years would cover about $40 billion in principal, with fixed-rate interest added.
The chips will arrive in batches, making this a future capacity commitment rather than computing power available today. For companies building with AI, the unset interest rate also leaves the financing cost unresolved despite the scale of the commitment.
A binding commitment of this scale backs AI infrastructure expansion with financing rather than just a spending ambition.
The 2027 lease start means builders should not treat the deal as an immediate increase in available computing capacity.
The economics remain unproven because the interest rate is unset and the master lease has yet to close.
Written automatically by the Made in 10 Newsroom from the source above.
OpenAI has won most of Replit’s AI spending from Anthropic with discounts, The Information reports. The reported contract is worth over $100 million a year, following an OpenAI discount of more than 30%.
OpenAI has detailed how its dots AI assistant grew out of its coding tools to work proactively in the background. It is available in ChatGPT desktop and mobile apps for Pro subscribers and above, with plans starting at $100 a month.
adsilico has raised ten million pounds to expand AI simulations that test medical devices on virtual patients. Its Virtual Chimera platform combines generative AI with physics-based simulations to assess devices across diverse patient populations.