Google unveils broad new push to put AI everywhere

Google unveiled a broad AI rollout across Search, YouTube, Android, Workspace and its Gemini app, including Gemini 3.5 Flash, Gemini Omni, and new agentic features like Spark. The company said it is processing 3.2 quadrillion tokens per month, up from 480 trillion a year ago and 9.7 trillion two years ago, underscoring rapid AI adoption. The announcements are positive for Google's competitive positioning, though most features are still in testing or rolling out gradually.

Analysis

This is less a product keynote than a distribution war update: Google is trying to turn AI from a standalone product into a default interaction layer across search, video, workspace, and devices. The first-order read is bullish for GOOGL because it expands monetizable touchpoints, but the second-order effect is more important: every new AI surface increases switching costs for users while raising the bar for AI-native competitors to win by product quality alone. The company’s real advantage is not model novelty; it is usage frequency plus proprietary intent data, which should compound ad targeting and agent reliability over the next 6-18 months.

The biggest underappreciated lever is pricing architecture. By pushing lower-cost tiers and making faster models broadly available, Google is likely defending usage share rather than maximizing near-term AI margin, which should pressure standalone inference economics across the sector. That is negative for pure-play model vendors and some SaaS copilots, but positive for Google’s ability to subsidize adoption until the market normalizes on an AI bundle benchmark. If the usage funnel holds, the long-run winner is whoever owns the interface where intent starts, not the model with the best benchmark.

WRBY is a real option on hardware distribution, but timing matters: glasses remain a 12-24 month commercialization story, and the risk is that Google’s ecosystem keeps the category in “developer/enthusiast” mode longer than bulls expect. The contrarian view is that the market may be overestimating near-term hardware revenue and underestimating how much this helps Android and search engagement without requiring a breakthrough wearable cycle. If agentic search works, it could reduce query volume but increase value per query; the near-term ad risk is manageable, but the long-term mix shifts toward fewer, higher-intent interactions.

Key Decisions for Investors

• Long GOOGL vs. short an AI-native search substitute basket for 3-6 months: Google’s control of intent and distribution should outweigh narrative risk, while smaller pure-plays face multiple compression as AI features become commoditized.
• Buy GOOGL call spreads 6-12 months out, targeting a post-I/O re-rating as AI usage metrics convert into investor confidence; use spreads to cap premium given event-driven volatility and headline risk.
• Initiate a tactical long WRBY only as a paired expression against consumer electronics hardware optionality, not as a standalone growth bet; best entry is on weakness after the event fade, with a 3-6 month catalyst window tied to display-glasses testing.
• Short a basket of listed AI application names with exposed usage monetization models over the next 1-2 quarters; Google’s bundling and distribution advantage raises the odds of price pressure and feature parity erosion.
• Trim near-term hedges in GOOGL if search engagement data remains stable over the next 30-60 days; the key reversal risk is not model quality but evidence of query cannibalization from AI mode exceeding ad replacement gains.