The rise of agentic AI in early 2026 kicked off what’s come to be known as the “SaaSpocalypse,” a massive selloff in enterprise software grounded in the premise that LLMs had put an expiration date on the SaaS business model. Salesforce (CRM), one of the companies most closely associated with SaaS in the first place, became a favorite target, falling as much as 43% in the first half of 2026 alone.
The basic bear case is straightforward: If AI agents can do more work with fewer employees, companies need fewer software seats. If employees increasingly work through ChatGPT, Claude or another AI interface, they may spend less time inside Salesforce. And if frontier models can generate bespoke applications on demand, then perhaps companies won’t need to spend on enterprise software at all.
I expressed my skepticism in an episode of the Street Check podcast (and in the pages of the Cabot Wealth Daily) in February on the heels of the Citrini Research report. Not because I believed that software stocks possessed some insurmountable moat, but simply because most of those fears overstated the capabilities of large language models.
Salesforce’s fiscal second-quarter results, released Aug. 26, showed a clear path forward for companies (and investors) wrestling with fears of the “SaaSpocalypse.”
[text_ad]
“SaaSpocalypse” Was a Whimper, Not a Bang
Salesforce shares surged on the heels of the report, rising 23% the next day, after the company reported fiscal Q2 revenue of $11.35 billion, up roughly 11% year over year and slightly ahead of Wall Street expectations. Subscription and support revenue rose 12% to $10.8 billion, though the company’s acquisition of Informatica contributed $440 million to that number.
The headline figures (non-GAAP diluted EPS of $5.90, up 103% year over year) were buoyed by investment gains, without which adjusted EPS was closer to $3.37, above the $3.28 analysts had been expecting.
What moved shares wasn’t the 3% earnings beat; it was how the company positioned its prospects moving forward.
Current remaining performance obligations, or cRPO, rose 14% to $33.5 billion, and CFO and COO Robin Washington said net-new annual order value growth was the strongest Salesforce has produced in four years, while customer attrition remained near record lows.
In other words, the metrics that investors would expect to deteriorate first if AI were rapidly hollowing out Salesforce’s business model showed resilience.
More importantly, Salesforce laid out a plausible path to adapting to widespread AI adoption.
A quote that has stuck with me persistently over the last few years as AI has emerged is, “AI won’t take your job. Workers who use AI will.”
That comports with my broader view of AI as a productivity multiplier that is broadly applicable across many industries, even if it falls short of the promises of the most dyed-in-the-wool AI boosters.
It appears, based on the market’s reaction to Salesforce’s earnings call, that the quote is just as applicable for companies as it is for workers: AI won’t kill your company; companies that use AI will.
To put a more positive spin on it, the market is rewarding companies that show they can integrate AI tools into their own product offerings—that they are “keeping up” and not being “left behind.”
On that note, Salesforce’s offerings really shone brightly.
Annual recurring revenue from Agentforce and Data 360 reached nearly $3.9 billion, up more than 210% from a year earlier. Agentforce ARR alone surpassed $1.5 billion, growing more than 240%. Customers consumed 3.2 billion “Agentic Work Units” during the quarter, up 97% sequentially.
Salesforce also said bookings for Agentforce One Edition and Agentforce for Apps more than doubled quarter over quarter, while Slack posted its fastest quarterly net-new annual order value growth since Salesforce acquired it. Slackbot users rose more than 150% from the previous quarter.
It is worth noting, however, that ARR is not a strong metric in general, as companies can annualize a strong quarter (or month, or even week) to project a much more impressive growth rate than may be justified. Also, Salesforce has expanded what it includes in Agentforce ARR, and usage metrics such as Agentic Work Units are still new enough that investors don’t have a baseline against which they can compare them.
Claudeforce May Be the Most Important Part
That brings us to the most important announcement of the quarter: Claudeforce.
Salesforce and Anthropic expanded their partnership with a new integration designed to combine Claude’s reasoning capabilities with Salesforce’s customer data, workflows, permissions and business logic.
The first product, Salesforce in Claude, is a Claude plugin with 37 prebuilt sales skills. A salesperson working inside Claude could, for example, reason over live Salesforce data, update a pipeline or take an approved action without having to leave the AI interface. More integrations across Claude, Salesforce and Slack are planned.
That is (at least based on the market reaction), exactly the kind of “keeping up” that investors are looking for these days.
The existential fear surrounding SaaS assumes that the AI model becomes the application. Claudeforce represents a different future: The AI model becomes the interface, while Salesforce remains the system containing the data, permissions, workflows and business rules that make the model useful inside an enterprise.
In that scenario, instead of becoming a threat to Salesforce, AI potentially increases the value of the data and infrastructure Salesforce already controls.
Keep Up or Get Left Behind
Since ChatGPT kicked off the generative-AI boom, “keep up or get left behind” has become something of a marching order for knowledge workers.
Learn to use the tools. Adapt your workflow. Figure out which parts of your job AI can do better, and make yourself valuable doing the parts it can’t.
Salesforce is discovering that the same rule applies to companies.
Had Salesforce spent the past three years insisting that AI wouldn’t change CRM, investors would probably be right to worry about its future. Instead, the company has spent aggressively on Agentforce, Data 360 and AI infrastructure, acquired businesses to fill gaps in its platform and, now, embraced a frontier-model company that could theoretically have become one of its biggest disruptors.
For investors, the most important takeaway is that, right now, the market is rewarding the companies showing that they’re adapting to AI, regardless of what they (or I) think about the long-term prospects of LLMs as a technology.
[author_ad]
