Opinion | Big Tech’s A.I. Takeover and the Slow Death of Silicon Valley Innovation
Silicon Valley prides itself on disruption: Start-ups develop new technologies, upend existing markets and overtake incumbents. This cycle of creative destruction brought us the personal computer, the internet and the smartphone. But in recent years, a handful of incumbent tech companies have sustained their dominance. Why? We believe they have learned how to co-opt potentially disruptive start-ups before they can become competitive threats.
Just look at what’s happening to the leading companies in generative artificial intelligence.
DeepMind, one of the first prominent A.I. start-ups, was acquired by Google. OpenAI, founded as a nonprofit and counterweight to Google’s dominance, has raised $13 billion from Microsoft. Anthropic, a start-up founded by OpenAI engineers who grew wary of Microsoft’s influence, has raised $4 billion from Amazon and $2 billion from Google.
Last week, the news broke that the Federal Trade Commission was investigating Microsoft’s dealings with Inflection AI, a start-up founded by DeepMind engineers who used to work for Google. The government seems to be interested in whether Microsoft’s agreement to pay Inflection $650 million in a licensing deal — at the same time it was gutting the start-up by hiring away most of its engineering team — was an end run around antitrust laws.
Microsoft has defended its partnership with Inflection. But is the government right to be worried about these deals? We think so. In the short run, partnerships between A.I. start-ups and Big Tech give the start-ups the enormous sums of cash and hard-to-source chips they want. But in the long run, it is competition — not consolidation — that delivers technological progress.
Today’s tech giants were once small start-ups themselves. They built businesses by figuring out how to commercialize new technologies — Apple’s personal computer, Microsoft’s operating system, Amazon’s online marketplace, Google’s search engine and Facebook’s social network. These new technologies didn’t so much compete with incumbents as route around them, offering new ways of doing things that upended the expectations of the market.
But that pattern of start-ups innovating, growing and leapfrogging incumbents seems to have stopped. The tech giants are old. Each was founded more than 20 years ago — Apple and Microsoft in the 1970s, Amazon and Google in the 1990s, and Facebook in 2004. Why has no new competitor emerged to disrupt the market?
The answer isn’t that today’s tech giants are just better at innovating. The best available evidence — patent data — suggests that innovations are more likely to come from start-ups than established companies. And that’s also what economic theory would predict.
An incumbent with a large market share has less incentive to innovate because the new sales that an innovation would generate might cannibalize sales of its existing products. Talented engineers are less enthusiastic about stock in a large company that isn’t tied to the value of the project they are working on than stock in a start-up that might grow exponentially. And incumbent managers are rewarded for developing incremental improvements that satisfy their existing customers rather than disruptive innovations that might devalue the skills and relationships that give them power.
The tech giants have learned to stop the cycle of disruption. They invest in start-ups developing disruptive technologies, which gives them intelligence about competitive threats and the ability to influence the start-ups’ direction. Microsoft’s partnership with OpenAI illustrates the problem. In November, Satya Nadella, Microsoft’s chief executive, said that even if OpenAI disappeared suddenly, his customers would have no cause to worry, because “we have the people, we have the compute, we have the data, we have everything.”
Of course, incumbents have always stood to gain from choking off competition. Earlier tech companies like Intel and Cisco understood the value of acquiring start-ups with complementary products. What’s different today is that tech executives have learned that even start-ups outside their core markets can become dangerous competitive threats. And the sheer size of today’s tech giants gives them the cash to co-opt those threats. When Microsoft was on trial for antitrust violations in the late 1990s, it was valued in the tens of billions. Now it’s over 3 trillion.
In addition to their money, the tech giants can leverage access to their data and networks, rewarding start-ups that cooperate and punishing those that compete. Indeed, this is one of the government’s arguments in its new antitrust lawsuit against Apple. (Apple denied those claims and has asked for the case to be dismissed.) They can also use their connections in politics to encourage regulation that serves as a competitive moat.
Remember those Facebook ads advocating greater internet regulation? Facebook wasn’t buying them for charity. Facebook’s proposals “consist largely of implementing requirements for content moderation systems that Facebook has previously put in place,” concludes tech-investigations site The Markup. That would give it a first-mover advantage over the competition.
When these tactics fail to steer a start-up away from competing, the tech giants can simply buy it. Mark Zuckerberg made this clear in an email to a colleague before Facebook bought Instagram. If start-ups like Instagram “grow to a large scale,” he wrote, “they could be very disruptive to us.”
The tech giants also cultivate repeat-player relationships with venture capitalists. Start-ups are risky investments, so for a venture fund to succeed, at least one of its portfolio companies must generate exponential returns. As initial public offerings have declined, venture capitalists have increasingly turned to acquisitions to deliver those returns. And the venture capitalists know that only a small number of companies can acquire a start-up at that kind of price, so they stay friendly with Big Tech in hopes of steering their start-ups to deals with incumbents. That’s why some prominent venture capitalists oppose stronger antitrust enforcement: It’s bad for business.
Co-option may seem harmless in the short run. Some partnerships between incumbents and start-ups are productive. And acquisitions give venture capitalists the returns they need to persuade their investors to commit more capital to the next wave of start-ups.
But co-option undermines technological progress. When one of the tech giants buys a start-up, it might shut down the start-up’s technology. Or it might divert the start-up’s people and assets to its own innovation needs. And even if it does neither, the structural obstacles that inhibit innovation at large incumbents could sap the creativity of the acquired start-up’s employees. A.I. looks like a classic disruptive technology. But as the disruptive start-ups that pioneered it get tied up with Big Tech one by one, it may become nothing more than a way of automating search engines.
The Biden administration can step in to begin to solve this problem.
Earlier this year, the F.T.C. announced it was investigating Big Tech’s deals with A.I. companies. That’s a promising start. But we need to change the rules that make co-option possible.
First, Congress should expand the law of “interlocking directorates” — which prohibits a company’s directors or officers from serving as directors or officers for its competitors — to prevent the tech giants from putting their employees on start-up boards. Second, the courts should penalize dominant companies that discriminate in access to their data or networks on the basis of whether the company is a potential competitor. Third, as Congress moves to regulate A.I., it should take care to write rules that don’t entrench incumbents.
Finally, the government should identify a list of potentially disruptive technologies — we’d start with A.I. and virtual reality — and announce that it will presumptively challenge any mergers between the tech giants and start-ups developing those technologies. That policy might make life difficult for venture capitalists who like to give talks about disruption and then get drinks with their friends in corporate development at Microsoft. But it would be good news for founders who want to sell products to customers, not start-ups to monopolies. And it would be good for consumers, who depend on competition but have spent too long without it.
Mark Lemley is a professor at Stanford Law School and co-founder of the legal analytics start-up Lex Machina. Matt Wansley is an associate professor at Cardozo School of Law and was general counsel of the automated driving start-up nuTonomy.
The Times is committed to publishing a diversity of letters to the editor. We’d like to hear what you think about this or any of our articles. Here are some tips. And here’s our email: letters@nytimes.com.
Follow the New York Times Opinion section on Facebook, Instagram, TikTok, WhatsApp, X and Threads.
