币圈稳健哥

A genius escapes Google, while Buffett is buying the dip? The signals are conflicting, but the answer might be straightforward.

Recently, the hottest topic in the AI space isn't about which company released a new model, but rather Google losing two key AI players in just 72 hours—Shah's departure to OpenAI, one of the eight original Transformer creators, and Nobel laureate Geoffrey Hinton joining Anthropic.

With this, all original authors of Transformers have left Google.

The market's reaction has been brutally honest: a single-day loss of 1.4 trillion RMB in market cap, the worst of the year.

But interestingly, while the market is in a panic sell-off, Buffett is continuing to accumulate Google shares. This is quite intriguing—on one side, top talent is voting with their feet, and on the other, the Oracle of Omaha is casting his vote with cold hard cash. Who's right? Who's wrong?

My take might make some uncomfortable: both are correct.

Short-term, the talent exodus is indeed a tangible bearish signal. The departure of key figures from Gemini and AlphaFold will inevitably impact the R&D pace, and the market has ample reason for its panic.

However, Buffett sees a different angle: Google's moat isn’t solely dependent on a few scientists.

Search, YouTube, Android—these three giants generate hundreds of billions in free cash flow annually, capable of funding AI R&D for a decade without profit. More critically, Google plays the role of "shovel seller" in the AI race: self-developed TPU chips, Google Cloud, massive user data—no matter which model vendor comes out on top, compute power and cloud services have to be provided by someone. This ecological position is far more stable than merely competing against large models.

To put it plainly, the market is fixated on where Shah is going, while Buffett is focused on where the money is flowing.

Talent will flow, technology will iterate, but the answer to the question, "who controls the next-gen compute distribution entry point" won’t change in the short term.

So these two perspectives are not contradictory—those in a panic sell are looking at quarterly rhythms, while those buying in batches are betting on a ten-year landscape. As for regular investors, which side you lean towards depends on what time frame you’re willing to commit to. History has repeatedly proven one thing: what truly determines long-term returns is not missing out on a genius, but losing a moat.