Google/Alphabet CEO, racing to catch up
Sundar Pichai
Profile
Sundar Pichai runs the most consequential AI company that almost missed the AI moment. As CEO of Google since 2015 and Alphabet since 2019, he presides over the lab that invented the Transformer, trained some of the first large language models, and built the hardware (TPUs) that made modern AI economically possible — and then watched OpenAI weaponize that research against Google Search with ChatGPT. The “code red” he reportedly declared in late 2022 wasn’t panic; it was an admission that Google had been sitting on the future while Sam Altman shipped it.
What followed is the most-watched turnaround in big-tech history. Pichai merged Google Brain and DeepMind under Demis Hassabis, killed the Bard brand, unified everything under Gemini, and pushed AI into Search, Workspace, Android, and the Pixel line faster than Google has ever shipped anything. Gemini 1.5 introduced the 1-million-token context window that made everyone else’s models feel cramped. Gemini 2.5 and 3 brought Google back into genuine parity — arguably ahead — on reasoning and multimodal benchmarks. The TPUs that Jeff Dean and team have been quietly iterating on for a decade are now a strategic moat nobody outside Google can replicate.
For developers, Pichai’s Google is finally behaving like a startup with infinite resources. The Gemini API is cheap, the free tier is generous, AI Studio is a legitimately good prototyping environment, and Gemini inside Google Cloud has become a serious Anthropic/OpenAI competitor rather than an also-ran. Watching Pichai turn a company of 180,000 people into something that ships weekly AI updates is a masterclass — not in vision, but in organizational physics. He didn’t invent Gemini; he made Google capable of building it.
The open question is whether he can protect the cash cow while the cow is on fire. Search ad revenue still pays for everything, and Gemini answers in Search are directly eroding the click economy that funded the last 25 years of Google. Add the DOJ antitrust judgment threatening Chrome and default-search deals, and Pichai is running the most expensive transition in tech history with regulators holding a scalpel to the wallet. Whether this ends with Google dominating the AI era or being remembered as the next IBM is the single biggest story in the industry.
Key Articles & Papers
AI at Google: our principles An important next step on our AI journey Google DeepMind: Bringing together two world-class AI teams Introducing Gemini: our largest and most capable AI model Our next-generation model: Gemini 1.5 A note from our CEO: The next chapter of our Gemini era Google's Response to ChatGPT: A 'Code Red' Sundar Pichai's Congressional testimony on AI and competition U.S. v. Google: The antitrust ruling that could reshape search Introducing Gemini 2.0: our new AI model for the agentic eraControversies
The Gemini image generation incident (Feb 2024) — Gemini’s image model refused to generate white historical figures and produced racially diverse Nazi soldiers, becoming a global story about AI safety tuning gone wrong. Pichai called the behavior “completely unacceptable” in an internal memo. The episode crystallized critiques that Google’s RLHF was ideologically heavy-handed and accelerated a broader industry pullback on aggressive safety tuning.
Project Nimbus and defense AI — Google’s $1.2B cloud contract with the Israeli government, combined with quiet revisions to the company’s AI principles that removed explicit prohibitions on weapons and surveillance applications, has drawn sustained internal protest and public criticism. Pichai has defended the work as standard enterprise cloud business.
The DOJ antitrust case — In 2024 a federal judge ruled Google an illegal monopolist in search. The proposed remedies — potentially including forced divestiture of Chrome and restrictions on default-search payments — represent the most serious existential threat to Google’s business model in its history, and they landed at the exact moment Pichai needs maximum flexibility to fund the AI transition.
The “code red” and strategic drift narrative — A persistent critique, articulated by former Google researchers and investors alike, is that Pichai’s consensus-driven management style let Google sit on transformer research for five years while startups ate its lunch. Noam Shazeer and other transformer authors left; the $2.7B deal to rehire Shazeer via Character.AI is often cited as evidence of how expensive that drift became.
Spotify Podcasts
