Salesforce's AI Agent Business Is Growing Over 200%, But Will It Show Up in Its Order Book?

Salesforce’s Agentforce AI agents hit $1.2B in annual recurring revenue (+205% YoY) and completed 3.8B tasks, but signed customer contracts have not accelerated versus management’s growth projections. Bookings for premium seat upgrades grew nearly 60% YoY in Q1 and 7 of the top 10 deals added seats last quarter, while remaining performance obligation (cRPO) is the key read on whether usage converts to revenue. Shares are up ~40% from the June low and trade at ~15x forward earnings into the quarterly report, keeping investor expectations highly sensitive to near-term contract conversion.

Analysis

CRM is becoming a test case for whether agentic AI creates incremental software spend or simply reshuffles it across pricing buckets. The market should care less about task volume and more about whether contract value starts to outrun usage, because that is what protects margin and the multiple. If bookings stay steady while management keeps leaning on mixed pricing, the signal is that AI is being bundled to defend share rather than to expand wallet share.

Second-order winners are the service layers around enterprise deployment, but the larger structural loser could be the whole per-seat SaaS cohort if customers conclude one agent can replace several human users at renewal. That would hit net retention before it hits revenue, which is why the next 1-3 quarters matter more than the current quarter. NVDA is only a partial beneficiary unless agent workloads meaningfully raise inference intensity; otherwise this is a software monetization debate, not a compute demand shock.

Contrarian view: the market may be overpricing near-term seat compression. Large enterprises move slowly, and many agent programs will remain embedded in flat-fee contracts long enough to keep reported revenue intact, even if usage is rising underneath. The key falsifier for the bearish thesis is an upward inflection in organic cRPO or evidence that premium-seat conversions are broadening beyond a few large deals over the next 1-2 quarters.

Key Decisions for Investors

• Into the print, express a cautious bearish view with a short CRM / long IGV hedge for 2-6 weeks; thesis is multiple compression if organic cRPO remains flat and the market re-prices AI as defensive bundling rather than growth. Cover if management implies sustained billings acceleration or >5% organic cRPO growth.
• Use CRM put spreads 1-2 months out only if the stock rallies on headline AI adoption but the order-book metrics do not inflect; target a retrace toward the pre-rally range with defined premium risk.
• Do not chase NVDA on this event alone; treat it as a second-order beneficiary only if management later quantifies higher inference usage. Reassess NVDA exposure only if enterprise AI rollouts show materially higher compute intensity in subsequent quarters.