Snowflake Leaps 20.9% as AI Drives a Third Acceleration

Snowflake shares rose 20.86% to 369.64 on Sept. 3, 2026 after the company reported a third consecutive quarter of accelerating product revenue growth and raised its full-year outlook, with CEO Sridhar Ramaswamy attributing about half of the outperformance to AI.

Snowflake Inc. (SNOW) delivered the rarest thing in enterprise software: growth that is speeding up rather than settling down. The data cloud company reported a third consecutive quarter of accelerating product revenue growth and lifted its full-year outlook, and the market repriced the shares in a single session.

By the last trade at 17:40 GMT on Sept. 3, 2026, Snowflake was changing hands at 369.64, up 20.86% from the previous close of 305.84 — a gain of 63.80 points. The stock traded between 364.00 and 384.56 on the day, meaning even the low of the session sat far above where it closed the day before. That is a repricing, not a drift.

Chief Executive Sridhar Ramaswamy, speaking to Ed Ludlow on Bloomberg Tech, put a number on the driver: artificial intelligence accounted for roughly half of the company's outperformance. He also pointed to rapid adoption of Coco, Snowflake's AI coding agent, and to rising usage among existing customers. The interview aired on Bloomberg Technology.

Why a third straight acceleration matters more than a single beat

Consumption-based software companies bill for what customers actually use. That makes their revenue line an unusually honest read on demand: if workloads slow, the invoice shrinks the same quarter. It also means acceleration is hard to manufacture. One quarter of faster growth can come from a large contract landing at a convenient moment. Three in a row implies the underlying volume of queries, model calls and stored data is compounding.

That is the mechanism behind Ramaswamy's claim that AI supplied about half the outperformance. Enterprise AI projects are data-hungry by construction — retrieval, fine-tuning, evaluation and agent workflows all pull against a governed store of company information. When those projects move from pilot to production, they show up as consumption, and consumption shows up as revenue.

The raised full-year outlook is the second half of the story and arguably the harder one. Management teams that beat a quarter but leave guidance untouched are telling investors the quarter was a timing artifact. Raising the year says the company expects the new run rate to hold.

What the 20.86% move implies about positioning

A one-day gain of that size in a large-capitalization software name is not a reaction to a small revenue beat. It is a reaction to a beat that arrived against low expectations. The comparison with the broad market on the same day makes the point: the S&P 500, via SPY, was at $773.05 and up 1.03%, the Nasdaq 100 through QQQ was at $717.43 and up 1.15%, and the Dow 30 via DIA was at $537.26 and up 1.25%. Snowflake outran the Nasdaq 100 by roughly 19.7 percentage points on the session — an illustrative gap, but one that shows the move was company-specific rather than a rising tide.

The intraday range is worth reading too. Snowflake touched 384.56 before settling back to 369.64 by the time of the last trade, giving up part of the initial surge. That pattern — a spike, then a partial fade with the price still enormously higher — usually reflects fast money taking the first move and longer-horizon buyers deciding what multiple the new growth rate deserves.

Coco, agents and the case for stickier consumption

Coco, the AI coding agent Ramaswamy singled out, is the kind of product that changes the arithmetic of a consumption business if it lands. Coding agents lower the effort required to build something on top of a data platform. Lower effort means more projects, and more projects mean more workloads billed. It is also a retention tool: an agent that understands a customer's own schema and pipelines is harder to walk away from than a storage layer.

Coco, the AI coding agent Ramaswamy singled out, is the kind of product that changes the arithmetic of a consumption business if it lands.

The strategic risk is the mirror image. Agents that write queries efficiently can, in principle, reduce the compute a given task consumes. Snowflake's bet — and the read implied by three quarters of acceleration — is that the volume of new work created outweighs any efficiency drag on the old work.

The Databricks contest is about where AI data lands

Ramaswamy also addressed the competition with Databricks, the privately held data and AI platform, over which system enterprises make the default home for AI-ready data. That contest is less about feature checklists than about gravity: whichever platform holds the governed, cleaned, permissioned version of a company's data tends to capture the AI workloads built on top of it, because moving data is expensive and moving governance is worse.

Snowflake's accelerating revenue is evidence in that argument, not a verdict. Databricks does not report public quarterly results, so investors cannot line the two growth rates up side by side. What they can do is watch whether Snowflake's acceleration extends into a fourth quarter and whether the raised outlook proves conservative.

The next markers to watch

Whether the acceleration extends. A fourth consecutive quarter of faster product revenue growth would make the AI-consumption thesis very difficult to argue against.
How much of the raise was AI. Ramaswamy put AI at about half the outperformance; the composition of the next guidance revision will show whether that share is climbing.
Coco's usage disclosures. Adoption described as rapid is a qualitative claim until the company attaches metrics to it.
Where the stock settles. The session high of 384.56 and the prior close of 305.84 now bracket the market's disagreement about what this growth rate is worth.

For now the tape has answered the immediate question. A software company whose customers are buying more each quarter, in a market that has spent two years asking who actually monetizes AI, got paid for it.

Key facts

SNOW price: 369.64, +20.86%, as of 17:40 GMT Sept. 3, 2026
Previous close: 305.84; day range 364.00–384.56
Growth streak: Third straight quarter of accelerating product revenue growth
AI contribution: CEO Sridhar Ramaswamy: about half of the outperformance

Frequently asked questions

How much did Snowflake stock move after the results?

Snowflake shares were quoted at 369.64 as of the last trade at 17:40 GMT on Sept. 3, 2026, up 20.86% from the previous close of 305.84. The stock traded in a range of 364.00 to 384.56 during the session, meaning the day's low still sat well above the prior closing level.

What did Snowflake actually report?

Snowflake posted a third consecutive quarter of accelerating product revenue growth and raised its full-year outlook. Chief Executive Sridhar Ramaswamy, speaking on Bloomberg Tech, said artificial intelligence accounted for roughly half of the company's outperformance in the quarter, alongside growing usage from existing customers.

What is Coco?

Coco is Snowflake's AI coding agent. Ramaswamy described its adoption as rapid in his Bloomberg Tech interview. For a consumption-billed platform, a coding agent matters because it lowers the effort required to build new projects on the data platform, which in turn can increase the workloads that generate revenue.

Why does accelerating growth matter more than beating one quarter?

Snowflake bills for consumption, so revenue tracks actual customer usage in near real time. A single strong quarter can reflect contract timing. Three consecutive quarters of faster growth suggests the volume of underlying workloads is compounding, and the raised full-year outlook signals management expects the higher run rate to persist.

How does Snowflake compare with Databricks financially?

A direct comparison is not possible for public investors. Databricks is privately held and does not publish quarterly results, so its growth rate cannot be lined up against Snowflake's. The competition between them centers on which platform becomes the default home for governed, AI-ready enterprise data.

How did the broader market perform on the same day?

Major benchmarks were higher but far behind Snowflake. As of 17:40 GMT on Sept. 3, 2026, SPY tracking the S&P 500 was at $773.05, up 1.03%; QQQ tracking the Nasdaq 100 was at $717.43, up 1.15%; and DIA tracking the Dow 30 was at $537.26, up 1.25%. Snowflake's move was company-specific.

Sources

AI Fuels Snowflake’s Accelerating Revenue Growth — Bloomberg Technology

Photo: Mutante · BY-SA 3.0 — source