Morgan Stanley reiterates Alphabet stock Overweight on AI agent rollout

Alphabet unveiled Gemini Spark and Gemini 3.5 Flash, with the new Flash API priced at $1.50 per million input tokens and $9.00 per million output tokens, about 70% below leading competitors. Morgan Stanley reiterated an Overweight rating and a $375 price target, while BofA also reiterated Buy with a $430 target, citing Alphabet’s expanding leadership in consumer AI. The new tools will roll out across Gemini, Search, Gmail, Chrome and third-party apps, supporting continued AI adoption and monetization.

Analysis

Alphabet is using product breadth, not just model quality, to widen its moat: the real leverage is distribution. Embedding a lower-cost agent across Search, Gmail, Calendar, Drive, and Chrome creates a default workflow layer that competitors have to dislodge one user action at a time, which is far harder than winning a standalone chatbot battle. The pricing cut on the new model tier also matters because it can compress inference costs for Google’s own products faster than it compresses industry margins, letting Alphabet subsidize adoption while preserving operating leverage.

The first-order winners are the consumer and travel/commerce platforms that become reachable through agentic actions, but the second-order winner is Google’s ad and commerce funnel. If the agent becomes a planning layer inside Chrome, the valuable event shifts from search query to transaction completion, which should increase monetizable intent density over 6-18 months. For the named partners, the upside is distribution and conversion, but the risk is disintermediation: once users habitually ask an agent to book or buy, branded apps may lose checkout control and pricing power.

The main bearish setup is that expectations are already high, so the stock can underperform even if execution is good. The market is likely discounting product launches but not yet fully discounting the margin tradeoff from heavier AI usage, support costs, and potential partner cannibalization inside Google’s own surfaces. A catalyst that would reverse the move is any sign that agent adoption is stuck at novelty usage rather than becoming a repeat habit, especially if usage growth fails to translate into higher ARPU or better retention over the next 2-3 quarters.

Contrarianly, the most interesting trade may be not long GOOGL outright, but long Google’s ecosystem share of time spent while short the weakest standalone beneficiaries of agentic discovery. The challenge is that many investors will treat this as a pure AI win; the better framing is that it is a browser-and-workflow land grab that shifts bargaining power toward whoever owns the default interface. If Google can keep inference cheap and latency low, it can expand engagement without a proportional cost spike, which is the key variable the street may be underestimating.

Key Decisions for Investors

• Long GOOGL vs. long-only AI software basket over 3-6 months: buy GOOGL on any 2-3% pullback and fund it by trimming high-multiple AI beneficiaries that lack distribution, targeting a relative outperformance trade as workflow ownership becomes the scarce asset.
• Add to GOOGL via call spreads for a 2-4 month catalyst window: use upside exposure rather than stock if the market is already crowded, because the risk/reward is better if monetization headlines continue but multiple expansion stalls.
• For travel/commerce adjacency, prefer BKNG over EXPE on a 6-12 month basis: BKNG has stronger pricing power and should capture higher-quality intent if agentic booking grows, while EXPE is more exposed if Google keeps more of the discovery layer in-house.
• Avoid chasing DASH on this headline; if anything, watch for a later entry after agent adoption is proven. The upside from incremental traffic is real, but the margin risk from Google-owning the interface is a longer-duration negative that could cap multiple expansion.