A One-Way Door and the IBM Fear
When Satya Nadella approved Microsoft’s multibillion-dollar expansion of its OpenAI partnership in April 2022, he framed it as a “one-way door.” The company, he wrote internally, could not afford to build parallel supercomputers for both Microsoft and OpenAI. Instead, it would divert scarce computing resources to a single, high-stakes bet. In Nadella’s view, that meant “outsourcing” core AI research and creating a massive dependency on OpenAI’s intellectual property. His now-public email shows what he really feared: Microsoft becoming “the next IBM” while OpenAI evolved into a new Microsoft. The historical analogy underscored a strategic anxiety about repeating IBM’s early PC-era mistake—ceding long-term power to a supposedly junior partner. Nadella’s testimony in the Musk v. Altman trial reveals a leader determined to capture OpenAI’s breakthroughs without allowing the startup to become the sole gatekeeper of generative AI.
The November Shock: OpenAI’s Governance Crisis
That strategic anxiety erupted into the open during the OpenAI governance crisis in November 2023, when Sam Altman was abruptly ousted and then quickly returned as CEO. As OpenAI scrambled to redesign its board, new court filings show Microsoft’s leadership pulled in close, even though it had no formal control. Nadella, CTO Kevin Scott, and president Brad Smith joined Altman in a flurry of texts about who should oversee the nonprofit foundation that ultimately governs OpenAI’s for‑profit arm. The episode exposed the core asymmetry of the relationship: Microsoft had sunk extraordinary resources into OpenAI and integrated its technology across products, yet its influence over OpenAI’s board remained indirect and contingent. The crisis made tangible Nadella’s earlier worries that Microsoft’s destiny might rest on a governance structure it did not fully shape—and could, in theory, be locked out of.
Who Microsoft Wanted on the OpenAI Board—and Who It Didn’t
Unredacted text messages entered in the Musk v. Altman trial offer a rare, granular look at the Microsoft OpenAI board wish list. Kevin Scott and Satya Nadella weighed a roster of high‑profile candidates, from former Airbnb COO Belinda Johnson, whom Scott called “great,” to ex-Gates Foundation CEO Sue Desmond-Hellmann, Nadella’s own proposal who later joined the board. But the sharpest signals came in the vetoes. Nadella objected to former Google Cloud CEO Diane Greene and longtime Amazon-linked executive Bing Gordon, citing their deep ties to direct AI competitors. Scott even floated himself as a temporary board member, a suggestion Nadella rejected. By evening, Altman was pushing a three‑person board of Bret Taylor, Larry Summers, and Adam D’Angelo, with Altman as CEO but off the board. The messages reveal Microsoft’s informal but potent ability to shape who would oversee the very lab powering its AI future.
Influence Without Control: The Legal Battle over OpenAI’s Mission
In the Musk v. Altman trial, Elon Musk’s attorneys argue that Microsoft’s behind-the-scenes board lobbying shows it put commercial interests ahead of OpenAI’s original nonprofit mission. Musk claims Microsoft helped divert a charitable project—funded partly by his estimated USD 38 million to USD 44 million (approx. RM175 million to RM203 million)—into a powerful, profit-driven partnership, breaching the trust underlying OpenAI’s founding. Nadella counters that Microsoft assumed enormous risk to support a lab no one else would fund, helping create “one of the largest nonprofits in the world” and enabling products like ChatGPT and Copilot. Under cross‑examination, however, he acknowledged knowing of no full-time staff, grants, or open-sourced work at the OpenAI nonprofit before March 2026. The clash highlights a new corporate playbook: tech giants seeking deep access to cutting-edge AI while keeping formal governance at arm’s length, in ways that courts are only now starting to scrutinize.