Cloud split: stateful vs stateless compute
A major cloud paradigm shift is underway — Amazon and OpenAI struck a multi‑billion relationship focused on “stateful” agent compute while Microsoft remains OpenAI’s exclusive host for stateless ChatGPT workloads, a change markets coverage reported. The split signals different infra patterns for persistent agent state vs ephemeral API scale — and interviewers are now probing consensus, sharding, and state management tradeoffs. This forces candidates to reason about consistency, latency, and multi‑cloud failover in design interviews.
Amazon pledged a $50 billion strategic investment in OpenAI, beginning with an initial $15 billion tranche, [announced]openai.com. AWS committed to provide roughly 2 gigawatts of Trainium accelerator capacity and will act as the exclusive third‑party distributor for OpenAI’s Frontier platform and its “Stateful Runtime Environment” on Amazon Bedrock, [reported]aboutamazon.com. Microsoft and OpenAI issued a joint statement reaffirming that Azure remains the exclusive cloud provider for stateless OpenAI APIs and that any stateless API calls created via third‑party collaborations would still be hosted on Azure, [stated]blogs.microsoft.com. SEC filings and reporting detail that the $50 billion package is structured in tranches with performance triggers (analysts flagged an initial close target and contingent payments), and independent filings analysis lays out contingencies, rights‑of‑first‑refusal and payment mechanics behind the headline, [analyzed]geekwire.com.
