"[The Microsoft] partnership had 'been foundational to our success' but had 'also limited our ability to meet enterprises where they are,'" OpenAI revenue chief Denise Dresser wrote in an internal memo reported by CNBC. That admission frames a major realignment in cloud distribution for one of the most visible artificial-intelligence companies.
Microsoft’s license becomes non‑exclusive, but key ties remain
OpenAI and Microsoft resolved a dispute that began when OpenAI signed a $50 billion deal with Amazon, creating competing exclusivity claims. Microsoft — which previously was the sole distributor of OpenAI products under a contractual clause that guaranteed exclusivity "until AI achieves consciousness on par with a human" — has converted its license to OpenAI intellectual property into a non‑exclusive agreement that runs through 2032.
Under the renegotiated terms, Microsoft no longer pays a revenue share to OpenAI. OpenAI will continue paying 20% to Microsoft through 2030, and that amount is now subject to a cap. Microsoft retains its designation as OpenAI's "primary cloud partner," a status the source says means OpenAI products are expected to ship first on Azure unless Microsoft cannot or chooses not to support them. The company also still holds roughly 27% ownership of OpenAI's for‑profit entity, and therefore continues to benefit financially from OpenAI's growth regardless of which cloud serves it.
OpenAI launches paid access on Amazon Web Services
As part of the new arrangement, OpenAI on Tuesday launched paid access to its models, the Codex coding tool and a new agent deployment service on Amazon Web Services. Codex, which OpenAI says has more than four million weekly users, can be configured to run through AWS Bedrock and count against existing AWS cloud commitments.
AWS product announcements cited in the reporting include GPT‑5.4 being available on Bedrock now, with GPT‑5.5 arriving "within the next couple of weeks," WS CEO Matt Garman reportedly said at a San Francisco event. Bedrock Managed Agents handles the underlying infrastructure for AI agents, allowing enterprises to deploy automated workflows without assembling the supporting architecture themselves.
Money, chips and commitments: Amazon’s incentives and OpenAI’s obligations
The commercial mechanics are consequential. To claim all $35 billion from Amazon under its pact, OpenAI will need to spin up two gigawatts of Amazon's Trainium accelerators — Amazon's custom chips that compete with other hardware for AI workloads. That capacity requirement ties potential near‑term revenue to substantial infrastructure deployment on AWS.
The broader market also shows large parallel commitments: Google announced plans to invest up to $40 billion in Anthropic, including five gigawatts of Google Cloud compute capacity over five years, while Amazon has made a separate commitment of up to $25 billion to Anthropic. Those numbers frame the scale of cloud and lab bets now in play.
Market share and competitive pressure from Anthropic
OpenAI is entering a market where Anthropic has built a significant lead in the enterprise. A Menlo Ventures report cited here shows Anthropic holds 32% of enterprise AI model usage while OpenAI holds 25% — a reversal from 2023, when OpenAI commanded half of the enterprise market and Anthropic held 12%. Analyst Ben Thompson is quoted saying Azure's prior exclusivity had constrained OpenAI's enterprise reach, forcing enterprises to adopt Azure even when their workloads ran elsewhere and creating friction that benefited Anthropic.
What this means for enterprises, Microsoft, and Amazon/Anthropic
- Enterprises and procurement leaders: They can now access OpenAI models and Codex through AWS and have the option to count usage against existing AWS commitments, removing a previous requirement to route those workloads through Azure.
- Microsoft: The company loses exclusive distribution but preserves several strategic advantages — a capped revenue stream from OpenAI through 2030, "primary cloud partner" status that favors Azure-first launches, and roughly 27% ownership of OpenAI's for‑profit business.
- Amazon and Anthropic: Amazon strengthens its enterprise offering by adding OpenAI models to Bedrock and tying commercial incentives to Trainium deployment; Anthropic enters competitive dynamics backed by large cloud commitments from Google and Amazon and holds a leading share of current enterprise usage.
The settlement rearranges the technical and commercial plumbing of enterprise AI: OpenAI gains the freedom to ship across clouds, Microsoft safeguards financial and launch advantages, and Amazon secures a high‑profile lab onto its infrastructure with a multibillion‑dollar, capacity‑dependent payout on the line. The immediate questions the facts leave open are operational — principally whether OpenAI can spin up the Trainium capacity required to access the full Amazon payments and how quickly enterprises will migrate or expand workloads to take advantage of the new multi‑cloud options.
Original story: https://www.govinfosecurity.com/openai-trades-azure-exclusivity-for-enterprise-reach-a-31537
