Dreamforce verdict: last year's AI models are already closing sales, and Salesforce sets $63B FY2030 target

Salesforce told a 50,000-person Dreamforce floor this week that it’s aiming for $63 billion in fiscal-2030 revenue, roughly $3.8 billion above the LSEG analyst consensus of $59.2 billion and above the $61.4 billion Bloomberg-compiled average. The number is the frame. The subtext, delivered by the enterprise buyers walking the show, is more consequential: the models already shipping are good enough to close deals.

“The majority of agentic outcomes aren’t driven by frontier capabilities. They’re driven by last year’s AI,” G2 chief innovation officer Tim Sanders told CNBC. Nice technology chief Kevin Lee said models “one generation behind” are “highly performant and effective at doing the things that our customers need. It’s almost like everything beyond this point is icing on the cake.” Nagarro’s Ram Reddy described waiting roughly three months before plugging in the latest models. SummitX’s Jaya Rohit Vuyyuru echoed the pattern.

The financial spine backs the mood. Agentforce ARR hit $800 million in Q4 FY26, up 169% year-over-year, with more than 29,000 deals closed since launch and production accounts up nearly 50% quarter-over-quarter. Combined Agentforce and Data 360 ARR now exceeds $2.9 billion, up over 200% year-over-year, having processed 19 trillion tokens (5x YoY) and 2.4 billion agentic work units. FY26 revenue landed at $41.5 billion, up 10%, with remaining performance obligation of $72.4 billion, up 14%.

Marc Benioff and COO Robin Washington paired the target with a fresh $50 billion buyback on top of $60 billion already repurchased. Applications president Patrick Stokes unveiled Claudeforce, an Anthropic-powered beta with up to 1,000 enrolled clients, alongside a new reasoning model called Koa. Salesforce’s Informatica acquisition closed last November. Shares closed Wednesday at $250.54, up 67% from the June 22 trough but still 5.4% below where they started the year.

For a small-team owner watching from outside the Moscone bubble, the operative signal isn’t the GPT-6 Astra or Claude Fable 5.1 roadmap. It’s that pipeline software built on last year’s models is already producing the outcomes buyers are paying for. Waiting is the expensive posture now.

Sources

• https://www.cnbc.com/2026/09/18/at-dreamforce-business-leaders-say-older-ai-models-are-enough.html
• https://www.cnbc.com/2026/09/18/the-saaspocalypse-is-over-how-salesforce-now-plans-to-thrive-in-an-ai-world.html
• https://www.cnbc.com/2026/09/16/salesforce-issues-revenue-target-of-63-billion-for-fiscal-2030.html
• https://www.sec.gov/Archives/edgar/data/1108524/000110852426000056/crm-q4fy26xexhibit991.htm
• https://finance.yahoo.com/markets/stocks/articles/salesforce-sets-63-billion-fiscal-111846987.html