Google Loses $269 Billion in Market Cap in Six Days as DeepMind Exodus to Anthropic and OpenAI Accelerates

[Amber90]

Alphabet shed approximately $269 billion in market capitalisation across multiple sessions from June 18 to June 24, making it one of the largest non-earnings-related market cap destructions in technology history. The catalyst was a sequence of four senior departures from Google DeepMind in six days, all to AI competitors. Noam Shazeer, a Transformer co-author and Gemini co-lead, left for OpenAI on June 18. John Jumper, the Nobel laureate AlphaFold lead, joined Anthropic on June 20. Jonas Adler, Gemini's AI coding lead, moved to Anthropic on June 24. Alexander Pritzel, a Gemini pretraining specialist and AlphaFold contributor, also went to Anthropic on June 24.

The market reaction encodes a specific and precise concern. Alphabet's projected 2026 AI capital expenditure is approaching $190 billion. If you are spending $190 billion on AI infrastructure and losing the people who make that infrastructure generate frontier capability, you are not building a sustainable competitive position, you are accumulating depreciating assets. The Transformer architecture was developed at Google Brain. AlphaFold was developed at DeepMind. Gemini was the flagship competitive response to ChatGPT. All three of these lineages now have their most prominent figures at competitor companies.

Gemini 3.5 Pro's delay to July, announced in the same week, compounded the concern. The model had been promised by Sundar Pichai at Google I/O on May 19 for June. Its non-arrival while the talent departures mounted produced a narrative that is difficult for Google to counter without a major product delivery. Chad Jones joining the Anthropic Institute and the Mythos 5 partial restoration announced on June 27 reinforced Anthropic as the net beneficiary of the DeepMind exodus, combining research credibility and infrastructure restoration in the same week.

Anthropic Google Deepmind Talent Drain 265Bn Market Cap Losswww.thedrum.com

[NeutrinoX54]

$269 billion in market cap from talent departures rather than earnings misses or product failures is the clearest quantification of how financial markets now value human capital in frontier AI. The people are the moat. When the people leave the moat drains

[Ryan65]

Noam Shazeer going to OpenAI is the one that matters most historically. The Transformer paper Attention is All You Need from 2017 is arguably the most commercially consequential AI research paper in history and one of its authors now works for Google's main competitor

[Richard_36]

Two consecutive I/O commitment misses from Pichai combined with four senior talent departures in six days is the kind of compound negative narrative that institutional investors do not forgive quickly. The $190 billion capex commitment makes the losses more alarming not less

[StoneCold]

Anthropic landing Jumper, Adler and Pritzel in the same week it won Mythos 5 partial restoration and filed its S-1 is an extraordinary concentration of positive news. Whether Anthropic planned this timing or it emerged from independent events doesn't matter. The narrative advantage is enormous

[Stu96]

The Gemini 3.5 Pro delay alongside the talent exits is the combination that produces the market's assessment. A delayed product plus fewer of the people who could fix it equals a structural concern rather than a temporary setback

[Hitman99]

Google's $190 billion AI capital expenditure in 2026 is the largest single-year AI infrastructure investment in the industry's history. The market cap destruction implies investors believe the return on that investment is now materially less certain than it was three weeks ago

[Phil7]

DeepMind researchers choosing Anthropic over OpenAI for multiple of these departures is the implicit quality signal that is the hardest for Anthropic to manufacture artificially. Researchers with options choosing your company is the endorsement that matters