Published in Svenska Dagbladet, 2024-10-09. Translated from Swedish.
Two Nobel Prizes linked to AI went to Google this week. But it is ordinary search ads that have paid for the development behind them — and now those ads are under threat.
Most companies would be thrilled to have had even one Nobel Prize recipient among their staff at any point. Google has just received two — in one week. Geoffrey Hinton, joint recipient of the physics prize, has admittedly left Google since last year, but worked there for ten years before that. Demis Hassabis, one of three sharing the chemistry prize, heads Google’s AI unit, DeepMind.
Both appointments are in different ways related to the development of artificial intelligence — an area that has been a core focus since Google’s current CEO Sundar Pichai took over in 2015. Less than ten years later, the world is in the midst of an accelerating AI race, with entirely new competitors like OpenAI’s ChatGPT challenging Google at the very heart of its business: delivering the right information, quickly, to those who search for it.
The Nobel announcements come at a particularly interesting — and somewhat ironic — moment for Google. AI development has so far been primarily a cost for the company. The hope is that it is an investment in the future. But what has paid for it to date is something considerably more mundane and undramatic: search advertising.
The way those ads are sold has begun to be questioned — including in the courts. In September a legal process began in the United States to examine whether Google’s ad sales constitute a form of monopoly. The outcome will be known in a few months. In a separate ruling from August, it was established that Google holds a monopoly in the search market. On Tuesday the US Department of Justice submitted its views on what remedies should be imposed to address that monopoly.
It was not pleasant reading for Google. “Radical and sweeping” was how a Google legal executive described the proposals. The DOJ wants Google to be broken up. If the American Justice Department gets its way, products like the Chrome browser and the Android mobile operating system could be forced to separate from the rest of the business — both to prevent Google’s dominant position in search from being further entrenched, and to make the search giant less competitive in artificial intelligence.
The significance of such a break-up would be hard to overstate. Revenue from search advertising accounts for around 57 percent of Google’s total revenue. Add roughly 10 percent from YouTube and around 9 percent from ads shown on third-party sites. Google may want to position itself as an AI company — but in all material respects it is still an advertising company.
The strong profitability from those ads is what has financed major acquisitions and AI investments. Both Geoffrey Hinton and Demis Hassabis came to Google through companies it acquired. Being able to pick up promising research projects early and let them develop inside Google has been a strategy that worked well for many years. Few other companies would have the means — or even the ambition — to run that kind of project without being able to directly attribute revenue to it.
DeepMind — the division where Demis Hassabis works — was acquired ten years ago for around four billion kronor at today’s exchange rate. Add ten years of salaries and expensive infrastructure on top. Even the T in competitor ChatGPT — transformer — is a technology originally developed by Google.
Something as simple as a text ad in a search result has thus paid for one of the most advanced areas of technological development we have today. What happens if the US Justice Department gets its way and Google is split into smaller pieces — and the company’s strong position in the advertising market weakens? The tap for some future technologies could be turned off. That may sound like a good outcome for the free market. But it is today’s ads that are paying for tomorrow’s Nobel Prizes.
