Here is the uncomfortable question Alphabet’s management memo did not answer: what happens when the people who designed the infrastructure leave, not just the people who trained the models that run on it?
Alphabet shares dropped after news that Google DeepMind CEO Demis Hassabis would take on a bigger role and veteran AI leader Jeff Dean is leaving. The stock fell over 4%.
Why This Stock Matters Now
None of this happened because Alphabet’s core business is struggling. The company reported revenue of $119.8 billion for the quarter ended June 30, up 24% year over year, with Google Cloud growing 82% to $24.8 billion and operating margin widening to 34%.
Cloud backlog grew to $514 billion, and the Gemini app reported 950 million monthly active users.
That contrast is the real story here: strong earnings did not stop the sell-off, because investors are not looking at last quarter’s results. They are focused on what comes next.
The Investment Thesis
Alphabet’s AI ambitions rest on two pillars: the models themselves and the infrastructure layer underneath them. Today’s departures struck both simultaneously. Jeff Dean did not just lead AI teams. He co-created MapReduce, built Bigtable and Spanner, and architected the distributed systems that Google’s entire AI operation runs on. Reports say Dean is leaving after roughly 27 years at Google to start a new AI company with other Google AI veterans, and Alphabet will invest in it.
The market is not selling because a company lost good people. The market is selling because it lost the people who are hardest to replace.
The Business Behind the Stock
Google shares are up sharply over the past year, and much of that momentum has been tied to the company’s AI growth, with Google Cloud sales climbing 82 percent year over year to $24.8 billion last quarter.
Below the revenue line, a more complicated picture has emerged. Alphabet reported approximately -$5.9 billion in free cash flow in Q2 2026, the first negative FCF quarter since its 2004 IPO. Capital expenditures surged to $44.9 billion, and management raised 2026 capex guidance to $195 billion to $205 billion, with further increases expected in 2027. The company that was once the gold standard of capital efficiency is now spending faster than it generates operating cash.
What’s Changing
Under the new structure, Hassabis will transition out of his day-to-day operational responsibilities at Google DeepMind to become Chair of the unit and Chief Scientist of Alphabet, while Koray Kavukcuoglu has been elevated to run day-to-day leadership as a Senior Vice President reporting directly to Sundar Pichai. He will be responsible in particular for overseeing the development of Gemini 4, Google’s next major AI model.
That reshuffling is the official version. The less official version is a sustained talent exodus that began well before today. Noam Shazeer, a key figure behind the transformer architecture underpinning nearly every chatbot in use today, left Google for OpenAI. John Jumper, who shared the 2024 Nobel Prize in Chemistry with Hassabis for his work predicting protein structures, has moved to Anthropic. Additionally, senior Gemini model developers Jonas Adler and Alexander Pritzel recently left Google for Anthropic.
OpenAI and Anthropic, both widely expected to pursue IPOs, can offer pre-IPO equity that Alphabet’s status as a publicly traded company cannot structurally match. That structural disadvantage is not solvable by writing a bigger paycheck. It is a feature of being a mature public company competing for talent against privately held startups sitting on asymmetric upside.
The restructuring comes amid mounting pressure on Google to deliver faster AI products after reporting highlighted delays affecting Gemini releases. Losing the architects of the system while simultaneously racing to ship faster is not an easy combination to manage.
The Risks
The bull case on Alphabet requires believing that Kavukcuoglu, a long-time DeepMind veteran with real credentials, can absorb responsibilities from multiple senior exits without missing a stride on Gemini 4. Kavukcuoglu has led major DeepMind and Google DeepMind efforts, and will now oversee Gemini model development. That is a credible bet, but it is still a bet.
The capex trajectory compounds the concern. Alphabet raised its 2026 capex guidance to a range of $195 billion to $205 billion, and CFO Anat Ashkenazi told investors to expect a further significant increase in 2027. If Google Cloud growth decelerates even modestly while that spending continues, free cash flow could remain pinned negative for multiple quarters. Markets are becoming less willing to reward spending simply because it is large. Investors increasingly want to see those billions translate into higher revenue, stronger margins and sustainable free cash flow.
There is also a competitive dimension that goes beyond talent. Alphabet stock has now reacted sharply more than once this summer to high-profile AI departures. The pattern is becoming a market event in its own right.
What Investors Should Watch Next
Three metrics define whether the thesis holds. First, Google Cloud growth relative to the prior 82% pace. If that growth cools sharply, the capex burden becomes much harder to defend.
Second, Gemini 4’s timeline. Kavukcuoglu will be responsible for overseeing the development of Gemini 4, and the new structure is intended to strengthen governance while preparing for the next development phase. Any further delays after the ones previously reported will register as confirmation that the talent losses are affecting velocity.
Third, whether the new Jeff Dean-led startup attracts follow-on talent from inside Google. If it quickly raises a marquee funding round and begins pulling more Google engineers into its orbit, that is a material negative for Alphabet that the market may not have fully discounted.
Bottom Line
The numbers from Q2 are not the problem. Alphabet revenues grew 24% year-over-year and Google Cloud revenues accelerated to 82% growth, driven by demand for AI infrastructure and AI solutions, with nearly 90% of the Fortune 100 using Gemini Enterprise. Those are strong facts.
The problem is that the people most responsible for building the next layer of that advantage are leaving to build something new, and the structural incentive pulling them out is one Alphabet cannot easily eliminate. Wednesday’s decline interrupts what had been a standout year for Alphabet investors. Whether it becomes a turning point or a short-lived correction depends entirely on whether Kavukcuoglu and the restructured DeepMind can match the velocity the departing architects took with them. That answer will not arrive in a press release. It will arrive in Gemini 4.
