Salesforce reports first-quarter results after Wednesday’s close, and Marc Benioff walks in with a problem that money alone has not fixed. The stock is down roughly 30% this year even as the company keeps repeating that its Agentforce AI platform has reached $800 million in annual recurring revenue. Wednesday is the quarter where Wall Street stops taking the agent story on faith and starts pricing it on proof.
The setup is unusually binary. Either Agentforce is the early read on a real enterprise AI business, or it is a well-marketed pilot program that has not yet shown up as durable, billable revenue. The print will start to settle the argument.
The Number Benioff Keeps Repeating
Salesforce has built its AI narrative on three figures, all from its fourth-quarter results in late February: $800 million in Agentforce annual recurring revenue, up 169% year over year; more than 29,000 deals signed since launch; and 112 trillion records ingested into its Data 360 platform. Stack Agentforce together with Data 360 and the company puts combined ARR above $2.9 billion, a total that includes its Informatica acquisition.
These are company-claimed numbers, not independently audited, and that is precisely the tension. The skeptic’s question for Wednesday is how much of that $800 million is recurring, billable revenue versus credit and pilot commitments that have not yet converted into spend a CFO would defend in a downturn.
A Reporting Change That Cuts Both Ways
Starting this year, Salesforce is changing how it reports. Rather than break out a clean standalone AI revenue line, the company is collapsing its disclosures into two segments, “Agentforce Apps” and “Data 360, Platform and Other,” with Sales, Slack, MuleSoft, and Tableau folded into the mix. Investing.com reported the reorganization into two reporting segments, with dual reporting through fiscal 2027 and a full switch in fiscal 2028.
You can read that two ways, and both are fair. The generous read is that Salesforce is reorganizing around its agent business because that is where it is steering the company. The cynical read is that bundling makes it harder for outsiders to see exactly how fast, or how slowly, the pure agent revenue is growing. A clean standalone number would have settled the debate. This structure keeps it open.
The Stock Market’s Verdict So Far
Investors have already rendered a preliminary verdict. CRM trades near $180, down about 30% year to date, and UBS recently cut its price target to $185. Against that backdrop, Salesforce launched what Tikr described as its largest-ever buyback, a $25 billion accelerated share repurchase begun in March and funded with debt, representing half of a $50 billion program the board authorized in February.
Buying back stock at a 30% discount is a credible bet on yourself. It is also exactly what a mature software company does when growth alone will no longer lift the multiple. The buyback supports the share count. It does not answer the Agentforce question.
The Analyst Split
The sell side is divided in a way that frames the stakes. Bank of America’s Tal Liani sits at Underperform with a $160 target, flagging slowing growth, weaker upsell, and limited near-term Agentforce monetization. TD Cowen’s Derrick Wood is more constructive on in-line results. Consensus lands near $3.13 in adjusted EPS on about $11.05 billion in revenue, and the options market is pricing a post-earnings move of roughly 8.7%, more than double the stock’s recent average. The market expects fireworks in one direction or the other.
The Bigger Test
This print is bigger than Salesforce. It is the cleanest proxy for whether “AI agents are the new SaaS” survives contact with real enterprise budgets. The same agent wave threatening Salesforce’s seat-based model is the one Anthropic is riding into the enterprise, which is why a soft Agentforce number would not stay contained. The read-through would hit every legacy software vendor trying to bolt an agent layer onto a subscription business, from the CRM incumbents to the back-office suites. There is a labor story underneath the revenue story as well. Agentforce is sold partly on the premise that software agents can absorb work that headcount used to do, and every enterprise weighing that trade is watching whether Salesforce’s own results justify the pitch before they restructure their own teams around it.
What Wednesday Decides
The agent thesis gets validated or quietly buried on Wednesday afternoon, and the new reporting structure means this is the baseline every future quarter gets measured against. Benioff has spent a year telling Wall Street the agents are working. Now he has to show the receipts. A miss does more than dent Salesforce. It tells every CIO who has been piloting agents that the vendor selling the story loudest could not yet prove the model pays, and that is the kind of signal that freezes budgets across an entire category.