On a Wednesday afternoon stuffed with tech earnings news, the biggest surprise was Snowflake’s announcement that CEO Frank Slootman had decided to retire, effectively immediately, handing over the reins to Sridhar Ramaswany, who has been leading the company’s AI business.

Slootman has been at the helm of Snowflake for nearly five years, steering the company through its 2020 IPO, and overseeing a period of hypergrowth for the cloud-based data software provider.

Slootman has called Snowflake his third act. He took Data Domain public in 2007 (it was later acquired by EMC) and managed ServiceNow through a 2011 IPO. Now 65, Slootman has decided—to the surprise of Wall Street—that it was time to take a step back.

While Slootman will remain Snowflake’s chairman, Ramaswamy takes over as CEO immediately.

“When I joined Snowflake five years ago, I was already pretty far along in my career, obviously,” Slootman said in an interview with Barron’s, in which he was joined by Ramaswamy. “Quite honestly, I had no plans to ever take the helm again, but somehow I did.”

Slootman said his approach to finding a successor was always tied to finding the right person, rather than an arbitrary month or year. “Time was not the driver. We discussed lots and lots of people.”

Slootman said that in Ramaswamy’s relatively short tenure with Snowflake—he arrived as part of the company’s May 2023 acquisition of the AI-based search engine Neeva—he has become “our tip of the spear driving our AI strategy.” Slootman said he developed a “very, very high degree of confidence” that Ramaswamy was the right person to lead the company next.

Of course, the other big news from Snowflake was guidance for the January 2025 fiscal year that came in well short of expectations. Snowflake shares fell 20% in after-hours trading on Wednesday.

Ramaswamy told Barron’s that the company provides guidance based on historical consumption patterns. He notes that unlike companies that sign long-term contracts, Snowflake starts every quarter with zero revenue.

So his view is that the forecast is conservative.

“I mean, I’m chuckling here,” Slootman said. “We had an exceptional quarter. Our workload growth was 62%. The business is very, very strong. The thing you should be reacting to is the real results that we published rather than the guide. For consumption companies like ours, it’s hard to give guidance.”

“This is very different from a subscription model like a Salesforce, Workday or ServiceNow,” Slootman says. “On day one of the quarter, we literally have zero revenue. Those companies have 96% already under contract. People are coming to terms slowly and gradually to learn consumption models, because they’re very, very different. We think they are super equitable for customers, but it’s harder for inventors, there’s no doubt about it.”

Both executives said that the CEO change was unrelated to the poorly-received guidance.

“It’s an absurd notion,” Slootman said.