These days, you can hear snippets of the same conversation if you stroll through the lobby of practically any sizable tech campus in Menlo Park or Mountain View. who is departing. who is being pursued.
Who received the call over the weekend? There used to be rumors about new products and stock updates. They are now about engineers, and the numbers that are associated with them seem almost fictitious.
| Category | Detail |
|---|---|
| Largest AI talent markets globally | United States and India, each with nearly a million specialists |
| Europe’s biggest hub by headcount | United Kingdom — roughly 145,000 AI professionals |
| Germany’s standing | 17,000 AI engineers, fourth-largest pool worldwide |
| Most competitive country per capita | Ireland, 4.19 AI professionals per 1,000 inhabitants |
| Meta’s Scale AI deal | 49% stake for $14.3 billion |
| Google’s Windsurf move | $2.4 billion deal pulling in Varun Mohan |
| Notable founder-researcher launch | Mira Murati’s Thinking Machines Lab, valued at $10 billion |
| Reported signing bonus offers | Up to $100 million for top OpenAI talent |
| Source dataset | 1.6 million AI professionals tracked by Revelio Labs |
It’s difficult to ignore how rapidly the atmosphere has changed. AI talent was costly a year ago. It’s more akin to mythology now. What was once quiet poaching has become more akin to open warfare due to Mark Zuckerberg’s recent hiring binge, which included the $14.3 billion deal that brought Alexandr Wang of Scale AI into Meta’s new Superintelligence unit. Never one to keep things to himself, Sam Altman claimed that Meta has been offering OpenAI researchers $100 million in signing bonuses. It almost doesn’t matter if every detail is perfect. The key is the signal.
Even though the numbers seem ridiculous, there is a logic to it. In the words of Alexandru Voica of Synthesia, “if you’re spending a billion dollars to train a single model, ten million for the engineer who can actually build it starts to look like a bargain.” Not much has changed in terms of supply. Demand is now vertical. Even when compared to the software hiring frenzies of fifteen years ago, the result is wage inflation of a type the industry has rarely witnessed.
The map is also evolving. With nearly a million AI experts apiece, the US and India continue to hold a dominant position. However, Interface, a think tank with headquarters in Germany, recently examined data from Revelio Labs on 1.6 million AI workers and discovered something intriguing: Europe is subtly establishing a third pole. China’s pipeline seems to be closing, and tighter US visa regulations have made America less appealing to foreign talent. Just behind Singapore, Ireland currently has the second-highest AI density per capita in the world. The UK has about 145,000 AI engineers overall, while Germany has 17,000, the fourth-highest number in the world.
Then there’s the emergence of the founder-researcher, which seems like a truly novel development. Before the company had actually shipped anything, Mira Murati, a former employee of OpenAI, raised $2 billion for Thinking Machines Lab at a $10 billion valuation. John Schulman, a co-founder, and senior Character engineers joined.Mistral and AI. More and more, capital is following people rather than plans. That is a significant change. Investors are willing to pay because they seem to think that a small team of talented individuals can create something that the giants cannot.
Simultaneously, the definition of a “elite” researcher is changing. Companies are paying more attention to individuals who can combine statistics with cognitive science, neuroscience, or even ethics, but deep machine learning expertise is still the floor. Though not as much as the engineer who can actually push a frontier model into production and worry about safety along the way, the pure-paper publishing academic is still important.
As I watch all of this, I get the impression that the industry has entered a stage where standard economic reasoning isn’t quite applicable. This feeling is somewhere between excitement and uneasiness. Meta is hiring like a war-torn nation. Google has invested $2.4 billion to attract the CEO of Windsurf. Two dozen DeepMind employees have been discreetly absorbed by Microsoft. It doesn’t point to any impending cooling. There are still very few people available to construct these systems. The amount of money pursuing them continues to increase. Who wins and whether the wagers are profitable are still up in the air. However, talent currently has the upper hand. They are also unlikely to abandon that stance with ease.
