/ai-search833 words

Alphabet Stock Fell 12% as AI Mode Cannibalizes Search Ads

Alphabet shares fell 12% after execs said AI Mode earns less per query than search ads, as the company plans $200B in AI infrastructure spending this year.

Can Google win the AI race without killing its cash cow?
Can Google win the AI race without killing its cash cow?AI-generated
  • Alphabet stock fell more than 12% in late October — its worst single-day drop since its 2004 IPO.
  • Alphabet expects to spend roughly $200 billion this year on AI infrastructure.
  • Google's U.S. search share grew from 80% in 2009 to nearly 90% in 2020; revenue grew from $200,000 in 1999 to over $280 billion by 2022.
  • Google invented the transformer in 2017 and built the unreleased Meena chatbot before OpenAI launched ChatGPT on Nov. 30, 2022.
  • Google says search queries grew at their fastest year-over-year rate in five years last quarter.

Alphabet's stock dropped more than 12% in a single day in late October — its worst one-day percentage decline since going public in 2004 — after executives admitted that AI Mode, its chatbot-style search feature, earns less per query than traditional search ads.

The disclosure came alongside another number: Alphabet expects to spend roughly $200 billion this year alone on infrastructure, mostly data centers, to fuel its AI push. The combination crystallized a question that has hung over the company since ChatGPT launched on Nov. 30, 2022: can Google win the AI race without killing the search business that made it rich?

That business is enormous. Google generated just over $200,000 in revenue in 1999, the year Marissa Mayer joined as employee No. 20. By 2022, under parent company Alphabet, revenue exceeded $280 billion — a more-than-millionfold increase. Its U.S. search market share grew from 80% in 2009 to nearly 90% in 2020, cemented partly by default-search deals a federal judge ruled illegal in 2024.

Did Google squander its AI head start?

By one reading, yes. A team of eight Google employees invented the transformer — the technology behind modern chatbots — in 2017. Google built an internal chatbot, Meena, years before ChatGPT, and never released it. The company reportedly declared a "code red" after ChatGPT's launch.

Google's chief scientist, Jeff Dean, explained the hesitation on Dwarkesh Patel's podcast last year: "From a search perspective, you want to get the right answer 100% of the time," he said. "These models were not near that bar." He acknowledged Google underestimated how useful chatbots would be for tasks people don't bring to search engines, like drafting notes or summarizing text.

That fits the innovator's dilemma, the theory Harvard Business School professor Clayton Christensen laid out in his 1997 book: companies making tons of money from an existing business keep disruptive technologies "bottled up in their labs," risking loss to rivals willing to ship imperfect products.

Mayer — now CEO of AI startup Dazzle and still a Google shareholder — pushes back. "The standards for an established company like Google... are fundamentally different than for an upstart or a newcomer," she said. "This is not a technology that's happening to Google. They helped invent this new technology."

Can AI pay Google's mounting bills?

The advertising question is the hard one. Mayer sees no dilemma in using AI to answer consumer questions — "They're just answering questions differently and arguably better" — but concedes trouble on the revenue side: "How do I take that advertising spend and move it to a new format that keeps my advertising customers as happy?" She still expects AI advertising to be more lucrative than what came before: "I don't see why it can't be as profitable."

Google, in a statement, said AI features are driving usage: "This past quarter, Search queries grew at their fastest year-over-year rate in five years." The company says it's taking a "deliberate approach": "building the best user experience first and then scaling monetization."

Analyst Benedict Evans, formerly of Andreessen Horowitz, frames the shift as a recurring "platform shift" — like PCs, the internet and smartphones — where "the whole table gets reset." His verdict on search share is blunt: "Google's share of search has really got only one way to go."

Does losing share mean losing money?

Maybe not. Evans points to Microsoft: it lost dominance when the internet era arrived, yet "the market became so much bigger" and "the shareholders got a lot more money." Caveat: Microsoft's stock didn't pass its dotcom-era peak for nearly 17 years.

Mayer expects multiple winners, not one. "There will be multiple big players," she said. "I don't think there will be just one." The massive compute and energy costs of frontier models (the most capable AI systems) create barriers to entry, but she argues the pie grows as AI attracts more users and ad dollars.

Evans flags a deeper risk: AI models could become so similar that developers can't charge much for them — like cellphone carriers who built the mobile-internet infrastructure while smartphone and app makers captured the rewards. "Just because it's amazing and cool and exciting and it's going to change the world doesn't mean that you have value capture or pricing power," he said.

Google hedges across scenarios. It sells access to its own models, supplies compute to Anthropic (in which it holds a stake that could appreciate at IPO), and runs YouTube, cloud and Waymo. CEO Sundar Pichai dismissed the dilemma framing on the All-In podcast: "The dilemma only exists if you treat it as a dilemma. You lean into it as hard as you can. It's the only way to do it."

What to watch next: per-query monetization of AI Mode, whether query growth offsets lower ad rates, and whether frontier models develop winner-take-all dynamics or keep converging in capability.

via feeds.npr.org (Source)

More from James Calloway

James Calloway

Show full bio

Market editor covering money, funding and market shifts around Search rankings and algorithm updates, generative engine optimization at SERP Journal.

231 articles