Down More Than 30% This Year, Could Salesforce Be an Underrated Artificial Intelligence Stock to Buy Right Now?

The market is treating CRM as a referendum on whether AI compresses software moats faster than it expands TAM. The key second-order issue is that a strong quarter with weak guidance usually means customers are still committing to seat-based and workflow spend, but management is not yet seeing enough net-new budget conversion from AI to justify a re-rate. That creates a narrow window where the stock can stay depressed even if operating execution is solid, because investors want proof of incremental monetization, not just retention.

The bigger hidden variable is competitive substitution inside the enterprise stack. If Agentforce adoption is real but not yet material enough to lift guidance, it may actually be a sign that buyers are testing AI copilots from multiple vendors before standardizing; that delays wallet share gains and could cap CRM’s multiple for several quarters. On the flip side, this same experimentation phase can be positive for infrastructure names: as enterprises pilot agentic workflows, demand intensity for model training, inference, and integration layers tends to migrate upstream before app-layer winners are obvious.

The valuation signal is tempting, but cheap software often stays cheap when the market believes growth is decelerating structurally rather than cyclically. The risk is not a near-term earnings miss; it is a 6-12 month narrative failure where AI features become table stakes and CRM is forced to defend pricing or bundle more aggressively, pressuring margin quality. A rerating likely needs a cleaner proof point: either sustained ARR acceleration from Agentforce or evidence that AI is expanding deal sizes instead of merely preserving renewals.

The contrarian view is that consensus may be overfocusing on whether AI helps CRM directly, when the first durable monetization may be indirect: partner ecosystem, implementation services, and adjacent compute/network demand. If CRM remains range-bound, that is not necessarily bearish for the broader AI complex; it may just mean value is shifting away from application-layer incumbents toward picks-and-shovels beneficiaries. In that sense, the stock’s weakness could be an input signal for rotating capital into faster-cycle AI infrastructure rather than a standalone buy-the-dip setup.