Kevin McLaughlin at The Information had a piece this week we’ve been passing around the office. Snowflake is up 35% on the quarter. Salesforce is down 30% on the year. Kevin makes a strong case that the gap between those two numbers tells you where enterprise AI value will actually live.
His framing is the part that stuck with us. Snowflake’s advantage, he argues, is that it doesn’t have a legacy enterprise app business; it has to drag into the AI age. Sridhar Ramaswamy was pretty blunt about it on the earnings call. Customers are leaning into Snowflake’s coding agent. They’re asking natural-language questions about their sales metrics. They’re searching across Snowflake, Microsoft, Salesforce, and SAP, and the data layer is where it all converges. As Kevin put it, throwing AI on databases is a decent business.
Now look at Salesforce. $1.2B in Agentforce ARR, which is a real number, and the stock still got hammered. CRPO growth slowed. Free cash flow guidance was cut by 5 points due to a $25 billion debt raise to repurchase stock. When Goldman’s Gabriela Borges asked the obvious question about how Salesforce plans to charge for Headless 360, the answer was a promise to figure it out with customers and partners later. There’s a webinar on Friday.
That’s not a strategy. That’s a stall.
Here is where we think Kevin only got half the story.
Throwing AI on databases isn’t just a decent business. It’s the business. But raw AI pointed at raw data is how you end up with confidently wrong KPIs, hallucinated revenue numbers, and a dashboard that contradicts your board deck. The companies that win this cycle are the ones that feed AI governed, consistent, business-ready metrics. Same numbers across every agent, every query, every model, every time.
Which is why next week matters.
We can’t say much yet. But if you’ve been watching where AtScale and Snowflake have been heading, and you read Kevin’s piece and felt the same thing we did, you probably already know what’s coming.
Snowflake is the platform. We’re about to make it speak fluent business.
More soon.
SHARE
Guide: How to Choose a Semantic Layer