Salesforce did not just beat earnings on Wednesday. It delivered a quarter that fundamentally altered the conversation around enterprise AI, a conversation that had turned deeply skeptical after weeks of trillion-dollar software selloffs and existential hand-wringing about whether artificial intelligence would destroy the very industry that built it.
The numbers deserve to be stated plainly. Fourth-quarter revenue came in at $11.18 billion, up 11.7% year over year and landing at the top end of the company's prior guidance. Non-GAAP diluted earnings per share hit $3.04, comfortably ahead of the $2.93 consensus. Full-year revenue for fiscal 2026 crossed $40.9 billion, a milestone that would have seemed aspirational just two years ago when the company was navigating activist investor pressure and a workforce reduction of 10% of its employees.
Agentforce Is No Longer an Experiment
The headline number that will define this quarter, and possibly this year, is 22,000. That is the number of paid Agentforce deals Salesforce closed in Q4 alone, representing nearly 50% growth quarter over quarter. When Marc Benioff launched Agentforce at Dreamforce in September 2025, the product was met with cautious optimism from analysts who had watched too many enterprise AI products fail to move beyond pilot programs. Fourteen months later, Agentforce is not a pilot. It is a production platform processing work at a scale that rivals some of the largest AI deployments in the world.
The combined annual recurring revenue for Agentforce and Data Cloud reached approximately $1.8 billion, up from $1.4 billion just three months earlier. That $400 million sequential increase in ARR is roughly equivalent to adding a mid-size SaaS company's entire annual revenue in a single quarter.
11.14 Trillion Tokens Tell a Bigger Story
Perhaps the most significant data point buried in the earnings release is one that received relatively little attention in the initial analyst reactions: Salesforce's platform served 11.14 trillion tokens during the quarter. To put that in context, a token is roughly three-quarters of a word, meaning Salesforce's enterprise customers collectively processed the equivalent of approximately 8.4 trillion words through its AI infrastructure in 90 days.
This is not experimentation. Companies do not process 11 trillion tokens as a test. They process that volume when AI agents are embedded in daily workflows, handling customer service inquiries, analyzing sales pipelines, generating reports, and executing tasks that previously required human intervention. The token count is proof that the enterprise AI adoption curve has shifted from "are companies buying this?" to "companies are already running their businesses on this."
The Software Selloff Context
Salesforce's results arrive at a moment of maximum uncertainty for the software sector. Over the past month, enterprise software stocks shed more than $1 trillion in collective market value as investors panicked over the possibility that AI would replace, rather than enhance, traditional software platforms. Anthropic's demonstration that Claude could autonomously operate legal software sent shockwaves through the sector, and even companies with strong fundamentals saw their stock prices crater.
Against that backdrop, Salesforce's quarter is a data point that cuts against the doom narrative. The company is not being replaced by AI. It is becoming the delivery mechanism for AI in the enterprise. Every one of those 22,000 Agentforce deals represents a customer who chose to deploy AI through Salesforce rather than around it.
"We are witnessing the most significant platform shift in the history of enterprise software," Benioff said during the earnings call. "Agentforce is not a feature. It is a new category of enterprise computing, and our customers are deploying it at a pace we have never seen with any product in our history."
Marc Benioff, Salesforce CEO
What the Guidance Reveals
The company guided fiscal 2027 revenue to a range of $44.7 billion to $45.2 billion, implying roughly 10% growth at the midpoint. More importantly, Salesforce signaled that Agentforce and Data Cloud will represent an increasingly larger share of new bookings, with management projecting that AI-related products will account for more than 20% of new annual contract value by the end of the fiscal year.
Operating margins also held firm. The company reported a non-GAAP operating margin of 33.1% for the quarter, up from 31.5% a year ago, demonstrating that the investment in AI infrastructure is not coming at the expense of profitability. This is a critical distinction in a market that has grown weary of companies promising AI revenue in the future while burning cash in the present.
The Competitive Landscape
Salesforce's results will inevitably be measured against the broader enterprise software earnings season. ServiceNow, which reported earlier this month, showed similarly strong AI adoption metrics, while Microsoft's Copilot has been gaining traction in the productivity suite. The emerging picture is not one of AI replacing enterprise software but of AI becoming the most important growth driver within it.
The 11.14 trillion token figure also positions Salesforce as one of the largest AI compute consumers in the world, trailing only the hyperscalers themselves. That scale gives the company negotiating leverage with GPU suppliers and cloud infrastructure providers that smaller competitors simply cannot match.
What Investors Should Watch
The stock rose more than 4% in after-hours trading following the release, a sharp contrast to the 15% decline it suffered in the weeks leading up to the report. Whether that rally holds will depend on several factors. First, investors will scrutinize the Agentforce deal pipeline for any signs of deceleration in Q1. Second, the margin trajectory will need to hold as AI-related infrastructure costs scale. Third, the competitive threat from Anthropic, OpenAI, and other AI-native platforms remains real, even if this quarter's results suggest it is manageable.
But the most important takeaway from Salesforce's quarter may be the simplest one. In a market consumed by fear that AI would make enterprise software obsolete, the largest pure-play enterprise software company just posted its strongest AI adoption quarter in history. The narrative is shifting, and for the first time in months, the data supports the bulls.
