The number was delivered to a room of investors at Dreamforce, and it was higher than anyone in the room had projected. On September 16, Salesforce gave guidance for its fiscal 2030 revenue: more than $63 billion. The figure from analysts, as compiled by LSEG, stood at $59.2 billion. The company’s own target, by a comfortable margin, said the skeptics are wrong.
Chief Executive Marc Benioff framed the moment in his usual register. The industry, he said, has reached a time when everything is changing. The remark was characteristically broad, but the audience understood the subtext: Salesforce has spent the past year fighting a narrative that the software business it built is the one the AI era will leave behind.
That narrative has a name. The shorthand for it, the SaaS doom thesis, runs like this: as frontier models get better, the seat-based software business becomes the thing being replaced. If an AI can draft the sales email, reconcile the records, and answer the customer, why does a company need to pay per seat for a platform of forms and dashboards? Salesforce, as the largest pure software-as-a-service company, has been the thesis’s favorite exhibit.
The guidance is an attempt to answer that question with arithmetic. A target of more than $63 billion implies growth that continues for the rest of the decade, and it assumes that Salesforce finds a way to sell AI, in the form of agents and autonomous tools, at a scale that offsets whatever the models do to its traditional seat revenue. The company has been pushing its Agentforce line of AI agents precisely to make that case.
The timing of the announcement mattered as much as the number. Software stocks have been beaten down in part because the market could not decide whether AI was a tailwind for the industry or a threat to it. Salesforce’s guidance, and the confidence behind it, gave the sector something to rally around, and software stocks bounced in premarket trading on September 17.
The rebound had a wider context. The previous day, the Federal Reserve had raised rates, and growth stocks sold off as long-term yields climbed. A day later, the tone shifted, and Salesforce’s long-term target was one of the catalysts that helped the software group recover some of the lost ground.
Salesforce’s position in the AI debate is distinctive because it sits at the point where the technology meets the enterprise. The company’s data, the customer records, the sales pipelines, the service histories, is exactly the data that AI agents need to be useful. The optimistic case for Salesforce is that it owns the system of record that every enterprise agent will have to consult.
The pessimistic case is that the models do not need the system of record at all, at least not in its current form. If an agent can hold the customer context in its own memory, the argument goes, the database is a commodity and the seat is an anachronism. The next several quarters of Salesforce’s results will be read, fairly or not, as evidence for one of these two views.
Benioff has spent the year on the offensive against the doom thesis, arguing that the productivity gains from AI will expand demand for software rather than shrink it. The 2030 target is that argument reduced to a single number, and its effect is to raise the cost of doubting him, since a target delivered publicly to investors is a commitment the company will be measured against for years.
The gap between the company’s figure and the analysts’ figure is the interesting part. Analysts are, by training, cautious, and a gap of nearly four billion dollars in a long-range forecast is a meaningful difference of opinion about the direction of an entire industry. One side of that gap is going to be embarrassed, and Salesforce has put itself on the side that says the industry grows.
For the broader software sector, the guidance functioned as a signal. If the largest pure SaaS company is willing to commit to growth through 2030, the argument that AI kills software is harder to sustain, and the sector’s share prices responded accordingly. The question now is whether the execution matches the confidence, and whether the agents Salesforce is selling actually do the work the number implies.
The Agentforce bet is the mechanism behind the number. Salesforce has reorganized its product story around autonomous agents that act on customer data, and it has argued that these agents expand the addressable market rather than cannibalize it. The 2030 target presumes that argument is right, and that agents become a revenue engine comparable to the seat licenses they are meant to augment.
The software sector’s reaction on September 17 suggested investors were willing to give the thesis a hearing. After months in which software multiples compressed on AI anxiety, a long-range commitment from the category’s leader was enough to lift the group. Whether that confidence survives the next earnings season is a different matter, and the company has given itself eight years to be proven right.
