Snowflake's CFO Reveals How AI Products Drove 50% of Last Quarter's Beat While Headcount Growth Slowed to a Third of Last Year's Pace
Piper Sandler Growth Frontiers Conference, September 15, 2026
Snowflake CFO Brian Robins used a fireside chat at Piper Sandler's 5th Annual Growth Frontiers Conference to pull back the curtain on how the company's AI product suite is reshaping both its growth trajectory and its cost structure, offering investors granular detail behind a quarter that Robins called one of the strongest beats in Snowflake's history. The headline number from that print: full-year revenue guidance was raised from 31% year-over-year growth to 36%, a 500 basis point increase in a single quarter. Robins told the audience that roughly half of that beat came from AI products and half from the core data warehousing business, though he was careful to note the two are increasingly intertwined.
AI Adoption Is Lifting the Core Business, Not Just Adding a New Revenue Line
The most consequential disclosure was less about the AI products themselves and more about their spillover effect on Snowflake's legacy consumption business. Robins cited a LinkedIn post from CEO Sridhar Ramaswamy noting an 11% uplift to core business consumption when a customer adopts CoCo, the company's flagship AI agent product. That figure matters because it reframes the investment thesis: AI adoption isn't cannibalizing or merely supplementing the core, it's expanding it. Robins disclosed that roughly 9,100 accounts have now adopted CoCo and just over 5,000 have adopted Cowork, with the current sales motion focused on deepening penetration within those accounts rather than simply adding new logos. "We're trying to focus on now is get our sales team to go in and talk about the use cases to drive deeper penetration in those accounts, so we can get more users," Robins said.
Forecasting Discipline Explains the Beat-and-Raise Pattern
Robins offered an unusually detailed explanation of why Snowflake, as a 100% consumption-based business, has been beating guidance by wider margins under his tenure than investors are used to. The company runs machine-learning forecasting models nightly, with a four-hour pipeline that produces a detailed forecast email to Robins, Ramaswamy and others by 5 a.m. every morning, built up from customer and product-level data rolled up by region. For the mature core business, Robins said the standard deviation on forecasts is very tight given years of historical data. New products are a different story. "With just a short amount of data with sort of high numbers, I think it would somewhat be reckless to extract that out and just apply that to guidance," he said, explaining that the company applies a deliberate haircut to nascent AI product trajectories rather than extrapolating early hypergrowth. With CoCo now two quarters into its life rather than one, Robins suggested the confidence interval is narrowing, which could mean smaller beats going forward but also more credible guidance.
Path to GAAP Profitability Draws Skepticism, but Robins Insists Growth Isn't Being Sacrificed
Snowflake's commitment at its Analyst Day to reach GAAP profitability in the fourth quarter of next year proved more contentious among investors than the company anticipated, according to Robins, with some applauding the discipline and others arguing Snowflake should be reinvesting more aggressively while it has an AI-driven growth advantage over unnamed competitors running at free-cash-flow breakeven. Robins pushed back directly on the idea that the two goals are in tension. "Did I do anything that put constraints on the business to be able to make that comment? And the answer is absolutely not. We're spending as much as we want to spend in the business," he said, pointing to his track record at VeriSign, where he said he took operating margins from low single digits to a level that helped multiply the company's market cap during his tenure.
Headcount Efficiency Is the Clearest Evidence of the AI Thesis Playing Out Internally
The most concrete data point supporting Snowflake's profitability push was headcount growth. Robins disclosed that the company has added roughly 330 people year-to-date, of which 170 to 180 came from the Observe acquisition, meaning organic headcount growth was closer to 150 people. That compares to approximately 940 net new hires over the same period last year, even as revenue growth guidance moved up 500 basis points. Robins attributed part of this to an engineering hiring strategy shift: rather than hiring senior engineers requiring large equity packages, Snowflake is increasingly hiring junior engineers who use AI coding tools to be more productive at lower cost. Internally, Robins said Snowflake has canceled all of its BI and analytics software contracts, replacing them with CoCo-built dashboards, and has deployed a sales enablement agent called Raven that automates account-executive prep work, along with an internal tool he calls "Good Morning CFO" that pulls data from roughly 40 sources each morning and can be interrogated down to the SKU or customer level.
Migration Times Have Been Cut Nearly in Half, but a Legacy Migration Opportunity Remains Largely Untapped
Robins said Snowflake's AI tooling, including capabilities from its acquisition of Natoma for MCP connectivity, has already compressed customer migration timelines from 10 to 11 months down to roughly 6 months, though he cautioned that customer-side dependencies prevent this from going to zero. He framed the broader opportunity in terms of total addressable market, citing a figure from Investor Day of a market approaching $300 billion over the next five to six years, a number he said excludes adjacent analytics, business intelligence, and other AI-driven use cases layered on top of CoCo and Cowork. "We are just starting to scratch the surface," Robins said, arguing that legacy data warehousing workloads represent a large, still largely untapped migration pipeline as enterprises seek to consolidate data into AI-ready platforms.
Fourth-Quarter Renewal Concentration Is a Known Seasonal Pattern, Not a New Risk
Asked about the risk of a large concentration of customer renewals landing in the fourth quarter, which aligns with when many enterprise customers finalize annual budgets, Robins downplayed any incremental risk, noting that account executives are in constant contact with customers throughout the year on usage, ROI, and expansion opportunities well ahead of renewal dates. "It's not like a major event. It's just a normal event for us," he said, adding that reps are already working renewal conversations well in advance given Snowflake's consumption-based commercial model.