Microsoft’s CTO testifies via email at the heart of Elon Musk’s allegations against the tech giant

Microsoft CTO Kevin Scott made the decision on Wednesday and, for the first time, spoke publicly about an internal email that Elon Musk’s lawyers cited to support allegations that Microsoft knew OpenAI was abandoning its nonprofit project before investing billions in the company.

That email, sent by Scott on March 7, 2018, read in part, “I wonder if OpenAI’s major contributors are aware of these initiatives? In theory, I don’t think they’ve funded an open-source effort to focus on ML.” [machine learning] talent so that they can go and build something that is closed, that makes a profit on its back.”

Musk alleges that Sam Altman and OpenAI received his donations to found a non-profit AI lab, and with the help of Microsoft, turned it into a for-profit business that enriched its leaders.

In Wednesday’s deposition, Scott said he was questioning whether OpenAI was standing up to the commercial plans it was putting into Microsoft, not asking big questions about its work. He explained that both companies are behind Google in AI, that OpenAI recently left Azure for Google, and that he was worried the talks would be “a big distraction.”

Scott said the OpenAI donor he had in mind was not Musk but rather his friend Reid Hoffman, the founder of LinkedIn, who sits on Microsoft’s board.

But later that year, Scott testified, at a dinner with Altman and retired Microsoft executive Craig Mundie at the Flea Street Cafe in Menlo Park, he learned important information: Hoffman, the donor he had been wondering about, was actually investing in a new for-profit OpenAI company and joining the nonprofit’s board.

Also at the dinner, Scott said he learned that OpenAI was raising $500 million, that Altman was leaving Y Combinator to lead the company full time, and that OpenAI had created a new “limited profit” business structure as part of a new funding round. Scott called the building “amazing and fascinating” – something he said he had never seen before.

The way to the party: But Microsoft was far from committed. Scott testified that the company had “a tremendous amount of due diligence to do,” including technical, financial, legal, and administrative.

By June 2019, the numbers were becoming clearer. In a confidential memo from the time, which was filed as an exhibit in the lawsuit, Scott and Microsoft CFO Amy Hood formally asked Microsoft’s board to approve a $1 billion investment in OpenAI. Scott warned that Google used its AI training infrastructure to advance, and that Microsoft was “trying to replicate” the results.

Without OpenAI, Scott wrote in an appendix to the memo, Microsoft faced “knowledge and talent gaps” that would make building its system “time-consuming and risky.”

A key part of the strategic case is that Microsoft needed what Scott called an “AI workload” on Azure — a customer pushing the platform at a scale that would dictate what infrastructure needed to be built. Google had that advantage; Microsoft did not.

The board approved the investment. Microsoft announced the deal in July 2019, the first investment in a multi-year partnership that will see the company make a total of $13 billion in OpenAI.

Within six months of that initial agreement, the companies had built their first AI computer together, and OpenAI used the computing power to train what would become known as GPT-3.

On Wednesday, Scott called the partnership a success. “I’m very proud of our infrastructure capabilities,” he said, adding that he’s proud of what Microsoft has allowed OpenAI to do.

Answer from Musk: One of Musk’s attorneys challenged aspects of Scott’s account with brief but pointed questions.

For example, Scott testified that he had no idea when he wrote a March 2018 email about whether OpenAI was releasing its technology as open source. Musk’s attorney showed Scott an email he received earlier, in which Microsoft chief scientist Eric Horvitz wrote OpenAI was “openly sharing their work, according to their core doctrine.” Scott confirmed that he had received it.

Musk’s attorney also pressed Scott on whether Microsoft had done due diligence to comply with nonprofit law. Scott said he did not know, adding that the legal work was handled by others on the Microsoft team.

New financial information: And at Wednesday’s event, Microsoft’s business development leader Michael Wetter talked about the extent of Microsoft’s commitment to OpenAI. He testified that Microsoft’s total spending related to OpenAI — including its $13 billion investment commitments, Azure infrastructure, and hosting costs — was “more than $100 billion” as of the end of the fiscal year in June.

Wetter testified that Microsoft has generated an estimated $9.5 billion in direct revenue from the partnership by March 2025. Separately, Information reported this week that total OpenAI-related revenue (including Azure server rentals, Copilot sales, and revenue sharing payments) will exceed $30 billion between 2023 and 2025.

Under their agreement announced last fall, Microsoft received a share of approximately 27% in OpenAI, with the commitment of OpenAI to spend $ 250 billion on Microsoft’s Azure cloud services.

When questioned by Musk’s attorney, Wetter admitted that Microsoft, having contributed 98% of the capital to the for-profit OpenAI business at one time, had active approval rights over major business transactions. This is the level of influence Musk’s lawyers argued he controlled.

Wetter said Microsoft has never turned down an authorization request.

Under the latest renegotiation of their agreement, announced as the trial began, OpenAI gained the ability to run its products on any cloud platform, ending its exclusive commitment to Azure. Amazon Web Services has moved quickly to offer OpenAI models on its platform.

Microsoft’s license to OpenAI technology was extended to 2032 but not exclusive, and the companies removed a clause that would have cut Microsoft off from future models if OpenAI announced it had achieved general artificial intelligence.

Musk’s lawsuit: Lawyers for SpaceX and the Tesla founder argued that Microsoft’s licensing rights gave it effective control over OpenAI’s transition from nonprofit to for-profit, and that the company continued despite its CTO warning of a potential problem in 2018.

Microsoft emphasized that it relied on OpenAI’s contractual guarantees that the partnership would not infringe on any third-party rights. Wetter testified that Microsoft did not find “circumstances related to Elon Musk” in its standard due diligence process.

Microsoft is named as a defendant in the lawsuit for allegedly aiding and abetting what Musk says is a breach of trust by Altman and OpenAI in turning a profit.

Next in suit: Testimony in the case ended at around 13:00 today at the federal court in Oakland. Closing arguments are scheduled for Thursday, and jury deliberations are expected to begin Monday.

A jury will decide whether OpenAI violated its charitable trust and whether Altman and others were unjustly enriched. If the jury finds for Musk, the jury will determine the amount of financial damages.

Musk is seeking up to $134 billion from all defendants, although US District Judge Yvonne Gonzalez Rogers questioned the methodology of those financial figures. Musk, the world’s richest man, said he would donate the proceeds to charity.

GeekWire reported on today’s proceedings via a live stream of court audio.