The recent news of Apple and Microsoft stepping away from their ‘observer seats’ on OpenAI’s Board serves as recognition that Big Tech’s slow takeover of generative AI startups has landed them in hot water with regulators around the globe. But the move was a superficial fig leaf to ward off necessary antitrust scrutiny. Regulators including the Federal Trade Commission (FTC), the United Kingdom’s Competition and Markets Authority (CMA), and the European Commission’s Competition Authority (EU), should not back down.

This past spring, we submitted a letter with Demand Progress, Open Markets Institute, and Revolving Door Project applauding the FTC for its investigation and urging the Commission to closely monitor the actions of Microsoft and other tech giants to ensure that market competition is preserved, consumer rights are protected, and ethical standards are upheld in the development of AI technologies. NextGen Competition also supports the joint official statement between the CMA, EU, FTC, and Department of Justice committing to protecting competition within the generative AI space.

So how did Big Tech covertly take over generative AI firms? Microsoft’s slow and covert takeover of OpenAI exposes Big Tech’s playbook.

A HISTORY OF REGULATORY EVASION

The pseudo-merger between Microsoft and OpenAI has played out over years in what appears to be a calculated attempt to evade government antitrust scrutiny.

The following outlines the relevant history:

• In 2019, Microsoft invested $1 billion in the company through OpenAI’s for-profit arm, OpenAI Global, LLC. As part of that investment, Microsoft became OpenAI's exclusive cloud provider, making OpenAI’s models dependent on Microsoft software and hardware to run.
• In 2020, Microsoft became the exclusive licensor of OpenAI's GPT-3 model.
• In 2021, Microsoft made a further investment in an undisclosed amount.
• In January 2023, Microsoft invested an additional $10 billion, acquiring a 49% stake in OpenAI.
• In November 2023, OpenAI announced that Microsoft would obtain a board observer seat on OpenAI’s board of directors after saving Sam Altman from being ousted from the company and reinstating him as OpenAI’s CEO.
• In June 2024, multiple media sources reported that Microsoft’s head of AI efforts Mustafa Suleyman took an unprecedented and unthinkable step of examining OpenAI source code and proprietary algorithms as he leads Microsoft’s development of a ChatGPT killer.

U.S. antitrust authorities examine investments for anti-competitive concerns when the investment exceeds approximately $110 million—if it is in a corporation. However, acquisitions of interests in entities other than corporations, such as limited liability companies (LLC), trigger antitrust scrutiny in the US only where the investment exceeds approximately $110 million and the investor acquires 50 percent or more of the entity.

If OpenAI were a corporation, then Microsoft's $10 billion investment would have easily triggered a government antitrust review—it would have exceeded the Hart-Scott-Rodino Antitrust Improvements Act of 1976 filing threshold by almost 100 times. But because Microsoft invested in OpenAI through an LLC, Microsoft’s 49 percent investment—despite its $10 billion price tag—fell just below the reporting threshold for investments in non-corporate entities. Microsoft was able to close its investment and continue its integration with OpenAI without a regulatory review or approval within the United States.

Microsoft’s former observer seat on OpenAI’s board is equally problematic. While Microsoft downplayed the seat as “non-voting,” it gave Microsoft access to confidential information about OpenAI. The former board position raised similar questions of evasion. Under a provision of an antitrust statute called the Clayton Act, it is illegal for a person to serve as an officer or director of two competing companies. The basic purpose of the law is to prevent competitors from sharing sensitive business information or otherwise coordinating their competitive activities. The law technically applies only to investor representatives formally appointed as board directors and not board observers, even though the same risks from information-sharing and coordinated conduct could be present.

Although Microsoft has now relinquished its board seat, the damage has already been done. With 49% ownership of OpenAI, Microsoft has woven its technology into some of its products. Microsoft's rights to OpenAI's intellectual property—including the ChatGPT and Dall E 3 models—gives Microsoft unparalleled access to OpenAI’s most valuable assets without any federal oversight or regulatory approval. ChatGPT chatbot is powered by its large language models, which run on Microsoft’s Azure cloud technology. Now, Microsoft’s engineers are using their access to OpenAI algorithms to develop its own “competing” product.

Perhaps most alarming is that Microsoft appears to have successfully pushed Open AI to abandon its original non-profit ethos—to build AI safely for the benefit of humanity—and instead pursue profits at all costs. Indeed, Sam Altman’s hero to villain arc has lawmakers and other observers alarmed.

INVESTIGATIONS ARE WARRANTED

It is imperative that the FTC continues its investigation of Microsoft’s relationship with OpenAI. An investigation is crucial for ensuring a diverse and competitive AI industry that prioritizes public interest and ethical standards over corporate gain. Other startups are not immune. Mistral, Inflection AI, Anthropic are facing similar pressures from Microsoft, Google, Amazon and others.

The supposed success of companies like OpenAI and Anthropic is often touted as evidence of a competitive market. However, the reality is that these emerging stars are significantly reliant on, and beholden to, the infrastructure and resources of these tech giants. And research shows a pattern of Big Tech offering resources and “guidance” to startups as a way of steering them away from any products that might disrupt their monopolies.

These concerns are further heightened by the broader context in which this relationship exists, as the world is already suffering the consequences of pre-AI tech monopolization—from dangerous cyber vulnerabilities and privacy violations to price increases and bowing to authoritarians. The dominance of a few key players—namely Google, Microsoft, and Amazon—in the realm of cloud computing casts a long shadow over AI and risks seeing history repeat itself.

We can’t let this happen.