Google and Microsoft to Lead AI Market as Computing Expenses Rise

Sam Altman’s aim to secure approximately $7 trillion for the development of artificial intelligence chips reveals more than just his ambitious goals. The rising costs of building AI infrastructure and the concentration of value in the hands of a few tech giants highlight the challenges and potential risks of an increasingly monopolistic industry.

Despite all the competition fueled by the launch of ChatGPT in late 2022 and the influx of new startups jumping into the hyped generative AI market, most of these new players are likely to exit or join the ranks of incumbents in the near future. a year or so. The cost of doing business is too high for them to survive on their own.

Take Sasha Haco, CEO of Unitary, which scans videos on social media for content that violates the rules. That would cost his company 100 times what it charges customers to order OpenAI’s video scanning AI tools. So Unitary makes their own models, which is high management balancing in itself. His startup must lease access to these rare AI chips from cloud vendors such as Microsoft Corp. and Amazon.com Inc.’s Amazon Web Services. These chips have doubled in price since 2020, Haco says, and are hard to come by. “We’ve had times where we can’t get our hands on what we need, so we have to pay 10 times the price,” he told me.

Unitary makes it work, but Haco admits that no generative AI startup has figured out how to profitably run a large business, at least not in the same way that big tech companies do. Another AI founder in San Francisco tells me that some of his peers, who have to rent AI chips and cloud computing, find that the only way to make money “is if people don’t use the product.”

“The best analogy is electricity,” says Ronald Ashri, CEO of Startup Dialogue.ai, which creates custom chatbots for regulated industries. “You are connected to the basic model and it is your electricity, and you use it constantly. Consumption is the highest single cost in the solution we deliver to our customers.”

Generative AI startups can build their technology in two different ways. They can develop their own version of OpenAI’s GPT-4 or, for example, Google’s Gemini, the so-called basic model that requires investments of hundreds of millions of dollars. Or they can build on top of an existing model that only requires an investment of tens of millions, which is what most AI startups are doing today.

In both cases, the main beneficiaries are cloud computing giants Microsoft, Amazon and Alphabet Inc.’s Google, as well as artificial intelligence chip maker Nvidia Corp. “Right now, all these startups are taking money from venture capitalists and giving it to cloud companies and Nvidia,” says Rodolfo Rosini, CEO of chip company Vaire Computing. That’s why Nvidia shares have more than doubled in the past year, close to $2 trillion.

You’d think big tech companies would be eyeing the AI startup landscape and licking their chops in this dynamic, hungry for new talent and ideas. But it’s not that simple. Most new generative AI startups don’t have many hard AI researchers, which makes them an attractive way to buy talent, because they depend on large, third-party models. These startups are often staffed by regular software engineers.

On top of that, big tech buyers like Meta Platforms Inc. are already investing heavily in their own internal AI efforts, says Nathan Benaich, founder of Air Street Capital, a London-based AI venture capital firm, and many of those companies are cutting back. significant costs only last year.

An even bigger stumbling block is regulation. Big tech companies are rightly wary of antitrust backlash in any big AI deal, thanks to the recent wave of tougher antitrust. Hence the shift to investing. Big tech investments in AI startups will reach more than $24.6 billion in 2023, up from $4.4 billion in 2022. The purpose of the change is to avoid regulatory oversight, says Brendan Burke, senior analyst at market research company Pitchbook, who also provided the figures.

Now that the Federal Trade Commission is investigating some of those investments — including Microsoft’s multibillion-dollar bet on OpenAI and Amazon’s investment in Anthropic — the pendulum may be swinging back toward traditional buyouts, Burke says.

Venture capitalists and startup companies have conflicting views on how many acquisitions will take place next year. What seems most likely: Regulatory pressure is preventing takeovers of leading AI startups valued at over $1 billion, such as Perplexity, Cohere, Character.ai, and Inflection. Instead, they attract investment – at least for the time being – as some of the long tail of smaller players are scraped off, while others under upward pressure give up.

The result is a playing field that looks very much like it does today, where the biggest players keep growing. That’s a win for big tech, and arguably for consumers, who will continue to get cheap access to AI. But it is also a loss for competition and society. When a small handful of companies dominate general purpose AI that is woven into our lives, it gives those companies enormous power and influence. We better avoid that outcome.