Salesforce Crushes Earnings as AI Bets Pay Off With 12% Stock Surge

Salesforce investors got exactly what they wanted on Wednesday: a beat, solid guidance, and proof that the AI revolution isn’t killing the software business after all. The cloud giant’s shares jumped 12% in extended trading after delivering results that surpassed Wall Street expectations across the board.

The numbers tell the story. Revenue grew 11% year-over-year in the fiscal second quarter ending July 31, while net income more than doubled to $3.53 billion, or $4.29 per share, compared to $1.89 billion the year prior. That’s an 87% jump that instantly silenced the doomsayers.

Where AI’s Real Money Is

Here’s the kicker: much of that income gain came from a $2.6 billion windfall on strategic investments, specifically Salesforce’s stake in AI darling Anthropic. The startup’s valuation hit $965 billion in May when it closed a funding round. Alphabet and Microsoft have also recently booked gains on their Anthropic bets, signaling that big tech’s AI bets are finally paying dividends in the short term.

But beneath the surface, there’s real operational momentum too. Free cash flow spiked 81% to $1.10 billion, blowing past the StreetAccount consensus of $643.2 million. That’s the kind of efficiency gain that doesn’t require accounting wizardry or one-time gains to explain.

What’s particularly impressive is how Salesforce’s AI products are gaining traction with customers. Annualized revenue from Agentforce AI products topped $1.5 billion, up 240% year-over-year. Yes, the growth rate slowed from over 200% the quarter before, but momentum remains undeniably strong across artificial intelligence products.

The Guidance Nobody Wants to Mess With

For the third quarter, Salesforce sees adjusted earnings between $3.42 and $3.44 per share on revenue of $11.42 billion to $11.50 billion. Wall Street was expecting $3.38 and $11.41 billion. Another beat, another raise.

Full-year revenue guidance came in at $46.1 billion to $46.4 billion, implying 11% growth at the midpoint. The prior guidance called for $45.9 billion to $46.2 billion. This is measured confidence, not hype. The company isn’t throwing down wild promises or hoping generative AI solves all problems by Friday.

Marc Benioff, Salesforce’s co-founder and CEO, struck a defiant tone on the earnings call. “This is not the SaaSpocalypse,” he said. “We’ve been hearing about this for the last two quarters, these dire predictions about the end of software and how the models eat everything, but none of them have come true for us.” It’s a fair point. Salesforce isn’t getting disrupted into irrelevance. Instead, it’s absorbing AI capabilities and turning them into real revenue.

The company also announced a plugin for Anthropic’s Claude that can compose emails for salespeople, arm them with information, and update records through chat. Practical AI, not theoretical. During the quarter, Salesforce also landed a $1.6 billion contract from the U.S. Department of Veterans Affairs and revealed plans to acquire customer service startup Fin for $3.6 billion.

Not Without Bumps

Still, not everything is smooth sailing. Robin Washington, Salesforce’s chief operating and financial officer, acknowledged “headwinds and volatility” in selling licenses for integration and analytics software. There are cracks in the foundation worth monitoring.

The company’s current remaining performance obligation came in at $33.5 billion, slightly above the $33.22 billion StreetAccount expected. That’s the business that’s locked in and waiting to be recognized, a comforting buffer against near-term uncertainty.

Yet context matters. Salesforce shares remain down 22% year-to-date, while the S&P 500 has gained 12%. Investors have been whipsawed by debate over whether generative AI threatens legacy software makers or enhances them. Wednesday’s earnings suggest the answer isn’t binary: it’s both threat and opportunity, depending on execution.

Salesforce’s biggest risk now isn’t disruption. It’s whether investors will reward operational excellence and AI traction consistently, or whether the next quarter brings fresh panic about models “eating” software as a category.

Source: CNBC