Why Microsoft’s surprise deal with $4 billion startup Inflection is the most important non-acquisition in AI

‍Microsoft stunned the tech world on Tuesday when it announced that it has hired Mustafa Suleyman, the cofounder of $4 billion AI startup Inflection, to run Microsoft’s AI operations. Karén Simonyan, another Inflection cofounder, is also joining Microsoft, along with an unspecified number of staffers.

Essentially, the cloud giant, worth $3.12 trillion, has nabbed one of the most coveted teams of AI experts at a pivotal time in the evolution of the buzzy technology.

But the deal—if that’s what this arrangement can be called—is highly unusual. Inflection, having recently secured a staggering $1.3 billion in funding just last year, has ranked among the most high-profile (or hyped, depending on your perspective) startups in the new crop of AI companies. As part of the deal, the company said it was shifting away from the consumer version of its Pi chatbot, launched less than one year ago. And, according to Forbes, Microsoft is not taking an equity stake in Inflection AI, and no intellectual property is changing hands.

To call this turn of events a head scratcher is an understatement.

There’s a lot to unpack here, from potential conflicts with Microsoft’s stake in OpenAI to Inflection cofounder Reid Hoffman sitting on Microsoft’s board of directors. Could this be the pin that pops the AI bubble, or is the technology bound to be run by a select, well-funded few that can afford the compute—that is, until the U.S. Justice Department and the Federal Trade Commission get involved?

Here are some of the key questions raised and items to watch following Microsoft and Inflection’s most unusual announcement.

If this isn’t acquisition, what exactly is this deal?

The proverbial elephant in the room here is that Big Tech companies like Microsoft are under intense antitrust scrutiny right now. In January, the FTC said it was launching an inquiry into Microsoft’s $13 billion partnership with OpenAI, the company behind ChatGPT (the FTC also said it was looking at investments between Amazon and Anthropic and Google and Anthropic).

Given the government’s concerns about the concentration of power in the AI market, it’s impossible to imagine Microsoft acquiring Inflection without facing a regulatory challenge. So what exactly is Tuesday’s arrangement between the two companies? An acquihire? Microsoft’s press releases describes it as an “organizational update” and says only that “several members” of the Inflection team in addition to the two cofounders are moving to Microsoft.

Microsoft has run this play before, although not to completion. When OpenAI’s Sam Altman was fired from the startup in November, Microsoft CEO Satya Nadella extended a welcoming hand to all OpenAI staff, including Altman and other executives, offering them a continuation of their work under Microsoft’s umbrella. The offer, which Altman accepted, ultimately became moot when he was restored to his throne at OpenAI. As many observers at the time noted, if Microsoft had managed to onboard the entire OpenAI team, it would have effectively achieved an acquisition of OpenAI without the usual antitrust scrutiny.

Will such a move pass muster with regulators, or will come across as a challenge that regulators feel compelled to respond to? Stay tuned.

What does this say about the Microsoft-OpenAI relationship?

It’s fair to wonder why Microsoft even wants, or needs, Suleyman’s AI expertise, since it already has a pricey partnership with Altman and the OpenAI team. Speculation immediately swirled on Tuesday that the move could hint at strains in the Microsoft-OpenAI partnership. According to Bloomberg, Nadella only informed OpenAI CEO Sam Altman on Monday about Suleyman and his team’s integration into Microsoft even though the conversations with Suleyman had apparently been underway for months.

While we don’t really know what the current state of the relationship between Nadella and Altman looks like, this move signals that at the very least Microsoft is determined not to keep all of its AI eggs in Altman’s basket. There have been other signs: In February, Microsoft made a $16 million investment in Mistral AI, a French competitor of OpenAI. And Microsoft has also taken steps to develop its own, lower-cost generative AI product that doesn’t rely entirely on OpenAI technology, The Information reported in January.

Is this the beginning of the end for AI chatbots?

With Tuesday’s news, Inflection said that it was moving away from developing its current Pi chatbot and would instead focus on building custom chatbots for business customers. Inflection said that the soon-to-be-extinct Pi is used by millions of people a week. But that apparently is not enough in a crowded and expensive chatbot field, many of which are based on expensive LLM models (training these models requires “eye-watering” levels of compute, OpenAI CEO Sam Altman once said.)

Suleyman’s original vision for Inflection looked very different from the “empathetic” AI chatbot called Pi that it went to market with in May. The initial plan was to develop an AI “chief of staff” that would be capable of seamlessly managing both personal and professional aspects of life—a generative AI agent poised to handle a multitude of tasks. Suleyman repeatedly insisted that it remained committed to creating this AI chief of staff, but those promises were tempered with caveats that it would require considerable time and effort.

Even since OpenAI launched ChatGPT in November 2022, the tech world has been experiencing a collective mania for AI chatbots, pouring billions of dollars into all manner of bots with friendly names (there’s Claude, Rufus, Poe, and Grok — there’s even a chatbot name generator). In January, OpenAI launched a GPT store that’s chock full of bots. But how much differentiation and value can these bots really provide? The general concept of chatbots and copilots is probably not going away, but the demise of Pi may signal that reality is crashing into the exuberant enthusiasm that gave birth to a countless chatbots.

Inflection cofounder Reid Hoffman is on Microsoft board

There’s another curious aspect to the Microsoft-Inflection news: Reid Hoffman, a co-founder of Inflection, is a director on Microsoft’s board. As Fortune reported in December, Hoffman is one of several venture capitalists, including Marc Andreessen, who sit on Big Tech boards and have potential conflicts because of the AI startups that their firms back.

Hoffman also previously served on the board of OpenAI but resigned to avoid potential conflicts of interest, citing a desire in invest in companies leveraging OpenAI’s software. However, Semafor reported he was “privately unhappy” about being asked to leave OpenAI’s board.

If Microsoft had acquired Inflection, Hoffman would no doubt have had to recuse himself from any involvement in the deal. But this isn’t an acquisition. Still, the move, which effectively eliminates a Microsoft rival in consumer AI chatbots, raises questions about how the deal came about and the details of its terms.

After publication, Hoffman posted on LinkedIn that “this agreement with Microsoft means that all of Inflection’s investors will have a good outcome today, and I anticipate good future upside.” It’s unclear whether the “good outcome” refers to a financial component in the deal for Inflection investors. And while the deal calls for Microsoft to license some of Inflection’s technology, the future value of that technology would seem to be lower now that Inflection has lost key members of its team.

Sourced from Fortune