Salesforce CEO Marc Benioff: This isn’t our first SaaSpocalypse

Salesforce did everything it could to convince investors that the AI ​​revolution was far from over when it reported fourth-quarter earnings on Wednesday.

Salesforce reported strong quarterly revenue of $10.7 billion, up 13% year over year. For the year, it reported revenue of $41.5 billion, up 10% from the previous year, with both results boosted by its $8 billion acquisition of data management company Informatica last May.

Net income came in at $7.46 billion, and the company provided strong guidance for next year, forecasting revenue of $45.8 billion to $46.2 billion — an increase of 10% to 11%. It also said its “residual performance obligation,” or RPO, is more than $72 billion. This number shows revenue within the contact that has not yet been delivered or recognized as earned revenue.

However, numbers can only do so much. SaaS stocks, with Salesforce as its poster child, have been under pressure recently. Investors fear that the rise of AI agents will undermine these companies, making their per-seat business models obsolete. The situation has been called “Saaspocalypse”.

The concept hung so high in the air during the earnings call that CEO Marc Benioff mentioned the term at least six times.

“Have you heard about SaaSpocalypse? And it’s not our first. We’ve had a few of them,” he said, later adding, “If there’s a SaaSpocalypse, SaaSpocalypse might eat it because there’s a lot of companies using a lot of SaaS because they’ve gotten better with proxies.”

In an attempt to convince the world of its continued validity, Salesforce dropped everything and sunk the kitchen in this earnings report. The company increased its dividend nearly 6% to $0.44 per share. It launched a new $50 billion stock buyback program. This is always a favorite with shareholders because it creates a powerful buyer for the shares and reduces the number of shares outstanding (which can boost the stock price).

The company also renewed the earnings call itself. It was part podcast, part business, and part regular Q&A with some questions from Wall Street analysts.

Rather than run through the numbers, Benioff conducted on-camera interviews with three Salesforce customers to attest to their love of the new agent options: the CEO of home appliance company SharkNinja, the CEO of Wyndham Hotels & Resorts, and, just to be clear, the CEO of SaaStr, the conference and media software maker. We’ll trim the interviews down to the shortest summary: They all love Salesforce’s AI agent products.

Salesforce also introduced a new metric for its agent products: Agent Work Units (“AWU”). The idea here is that instead of just counting “tokens” — the standard unit of AI processing volume — AWU is trying to measure something more important: whether the agent actually completed a task, such as writing into a log, rather than just generating text. (Salesforce recorded 19 trillion tokens last quarter, a number that sounds like a lot but is really not in the world of AI.)

“You can ask him a question and he can write you a poem, but that’s not really valuable in the enterprise world,” Patrick Stokes, Salesforce president and chief marketing officer, said on the call. So AWU is meant to measure when an agent writes to a log or performs some other verifiable task.

Moreover, Salesforce also presented its own architectural vision for the coming world of agents. It shows that SaaS software like it owns most of the technology stack, with AI modellers sitting at the bottom as invisible, interchangeable, and commoditized business engines.

This was a direct reversal of one of the reasons for the SaaSpocalypse selloff earlier this month, after OpenAI launched its enterprise agent platform Frontier. The OpenAI architectural vision shows that OpenAI owns most of the stack, with SaaS providers of systems of record (databases and business software platforms where companies store their underlying data) at the bottom as invisible drivers.

And if all that wasn’t enough to sway investors: Benioff was wearing a black leather jacket, echoing the distinct look of the CEO who’s clearly crushing it in the AI ​​world: Nvidia’s Jensen Huang.