Alphabet lifts 2026 spending plan as Google trains Gemini 4

Alphabet raised its 2026 investment forecast to between $195 billion and $205 billion as Google works on Gemini 4, a frontier model effort CEO Sundar Pichai said depends on training “much larger base models.” The new range is up from $180 billion to $190 billion, with Alphabet saying demand is still running ahead of investment.

The spending increase came alongside a strong second quarter. Alphabet reported $119.8 billion in revenue for Q2 2026, up 24% from a year earlier and ahead of analyst expectations of $116.9 billion. Google Cloud revenue rose 82% to $24.8 billion, while Google Services increased 15% to $94.5 billion.

Google also put larger usage numbers behind Gemini. The company said the Gemini app reached 950 million monthly active users, up from 750 million in February. AI Mode in Google Search has passed one billion monthly active users since its global launch in October, according to Google.

Gemini 4 is now the frontier focus

Pichai told investors that Google is training Gemini 4 and described it as the company’s most ambitious pre-training run so far. He said Google is seeing progress “at the frontier,” but did not disclose model size, training cost, benchmark results, launch timing or whether Gemini 4 will close performance gaps with rival systems.

Google is still planning to release new models on a monthly cadence, according to the company. Gemini 3.5 Pro, however, appears to sit outside that schedule despite being in testing. Pichai acknowledged areas where Google needs to improve, naming coding and agentic coding as examples.

That admission matters because Google is trying to serve two different AI markets at once: high-cost frontier systems that set capability benchmarks, and lower-cost models that can be deployed across consumer search, cloud products and developer workflows at large scale. Pichai said demand is strong for the Gemini Flash series because it balances performance and cost.

Search and ads claims still lack comparison data

Google said AI Mode, like AI Overviews, is leading to more search queries overall. The company also said it sends billions of clicks to websites each week through AI features in Search. Alphabet did not provide specific comparisons with traditional search volume or traffic patterns, which limits what can be concluded about the effect on publishers and other sites.

On the advertising side, Google said AI Max has exited beta and has 500,000 advertisers using it. Alphabet said AI is opening billions of new search queries that were previously difficult to monetize. It did not give revenue contribution, pricing, conversion rates or advertiser retention figures for AI Max.

Google also said the cost per AI Mode response has fallen to its lowest point since launch, even while the company uses stronger models. If sustained, that efficiency could help margins as AI search usage scales. For now, Alphabet is pairing lower serving costs with a higher investment plan, which signals that infrastructure demand remains the gating factor for Google’s AI strategy.

This story draws on original reporting from The Decoder.