Nadella says Microsoft overcorrected on OpenAI reliance

Microsoft CEO Nadella Outlines Dual-Track AI Strategy, Acknowledges Early OpenAI Overreliance

In a wide-ranging interview published this week alongside Microsoft Build 2026, CEO Satya Nadella offered a candid assessment of the company’s AI trajectory, acknowledging that its early dependence on OpenAI left it playing catch-up and detailing how the tech giant is now charting a more self-reliant course through proprietary models, disciplined infrastructure allocation, and a new pricing paradigm for the agent era.

Building Independence From OpenAI

Speaking with Stratechery’s Ben Thompson on June 3, Nadella was asked whether Microsoft had been “lulled to sleep” by offloading too much of its AI strategy to OpenAI. He did not deny it, instead framing the situation as one that has since produced a stronger competitive posture. “Here we are competing with a bunch of new people, a bunch of old people, and we have our own game,” he said.stratechery

The company unveiled seven in-house MAI models at Build, including MAI-Thinking-1, a reasoning model, and MAI-Code-1-Flash, its first purpose-built coding model. Nadella described a “dual frontier” approach: Microsoft maintains its own MAI model lineage with “clean data lineage” while continuing to use OpenAI’s intellectual property through 2032, comparing outputs across both to improve quality. Mustafa Suleyman, who leads Microsoft’s rebranded Superintelligence team, told Fortune that the restructured OpenAI deal now allows Microsoft to train larger frontier models, though he estimated it would take two to three years to match the top labs.cnbc

GPU Discipline and Enterprise Priorities

Nadella was unusually direct about infrastructure trade-offs. He revealed that Microsoft is deliberately withholding GPU capacity from what he called “Neolabs” — smaller AI companies seeking raw compute on Azure — to prioritize enterprise customers and internal model development. “We’re not selling raw GPUs to a bunch of Neolabs,” he said. “We are being very disciplined on some business that we turn away”. He characterized selling compute to AI startups as “easy money” that Microsoft is forgoing in favor of long-term positioning across three investment buckets: the hyperscale cloud business, its own applications, and research compute for MAI models.stratechery

The End of Pure SaaS Pricing

Nadella also predicted that traditional per-seat software pricing will give way universally to a hybrid model combining subscriptions with usage-based billing. “100%,” he said when asked if hybrid pricing would define the future. On Microsoft’s April earnings call, he had noted that nearly 60 percent of customer service clients already purchase usage-based credits, and that GitHub Copilot shifted to consumption-based pricing on June 1. CFO Amy Hood framed the transition simply: “You’ll just bill for usage”.crn

The strategic logic, Nadella explained, flows from the rise of autonomous agents that consume tokens continuously — a cost structure incompatible with flat per-user fees. He identified coding, security, and knowledge work as the three domains where Microsoft will concentrate its agent businesses, calling them “massive domains where tokens make sense”. The company says nearly 90 percent of Fortune 500 firms now have active agents built with Microsoft tools, and its AI business has surpassed a $37 billion annual revenue run rate.stratechery