Snowflake just turned two years of AI promises into a number Wall Street could not argue with. Shares jumped roughly 36% in extended trading on May 27 after the data-cloud company beat on the quarter, lifted its full-year forecast, and signed the largest infrastructure commitment in its history: a five-year, $6 billion pact with Amazon Web Services.
That is a lot of good news landing in a single evening, and the market treated it accordingly. The more interesting question for anyone holding the stock, or competing with it, is which piece actually moved the needle. The earnings were strong. The guidance raise was the tell. The AWS deal is the strategy.
The Quarter That Bought Back the Growth Story
The headline numbers were clean. Revenue came in at $1.39 billion, up 33% from a year earlier, with product revenue of $1.33 billion, up 34% and ahead of the $1.32 billion analysts had penciled in. Net revenue retention held at 126%, a figure that says existing customers keep spending more, and the company counted 779 customers generating more than $1 million in trailing product revenue, up 29% year over year.
Then management did the thing investors actually pay for: it raised the full-year product-revenue outlook and guided second-quarter product revenue to a range of $1.415 billion to $1.42 billion, comfortably above the roughly $1.37 billion Wall Street expected. For a stock that spent much of the past year fighting the perception that its consumption model was decelerating, a confident guide is worth more than any single quarter’s beat. The 36% move is the market repricing the growth curve, not just the print.
A $6 Billion Bet on Amazon, and on Itself
The structural story is the AWS deal. Snowflake committed $6 billion over five years to Amazon Web Services, its largest infrastructure commitment to date, in a strategic collaboration aimed squarely at enterprise agentic AI. According to the joint announcement posted to Amazon’s press center, the agreement leans on AWS Graviton processors for price-performance gains and GPU-accelerated EC2 instances for AI training and inference, with deeper product integrations and expanded distribution through AWS Marketplace.
Follow the money and the logic gets sharper. A $6 billion compute commitment is a cost, not revenue, and a company does not lock in that kind of spend unless it expects demand to justify it. Snowflake is buying two things at once: better unit economics on the compute it resells to customers, and a tighter go-to-market alliance with the cloud provider that already hosts much of its workload. Reuters reported that the company struck the deal as enterprise AI adoption accelerates, and the framing matters. This is a vendor betting that the agentic wave is real enough to pre-fund.
There is a circular quality worth naming. Snowflake commits billions to AWS, AWS pushes Snowflake harder through its marketplace, and both sides get to point at the partnership as proof of AI demand. It is the same flywheel logic running through the chip and cloud economy right now, where the biggest players keep writing checks to one another and calling the sum a market. The deals are real. Whether the end-customer revenue grows fast enough to make them pay off is the open question.
The Agentic Pivot Is the Real Tell
Strip away the dollar figures and the message is a repositioning. Snowflake wants to be the control plane for what it calls the Agentic Enterprise, with Cortex AI letting customers build applications for text-to-SQL, summarization, sentiment analysis, and entity extraction directly on governed data without shipping it outside the security perimeter. New surfaces like Cortex Code and Snowflake Intelligence are the company’s answer to a worry that has dogged the entire software sector.
That worry is not hypothetical. Earlier this year, the prospect of AI agents eating into seat-based software revenue rattled enterprise names hard enough to wipe out billions in value, a dynamic that surfaced when fears about autonomous agents threatening the SaaS model hit incumbents across the category. Snowflake’s pitch is that data gravity protects it: the agents have to run somewhere, on data that has to be trusted and governed, and that somewhere can be Snowflake. The AWS compute commitment is what makes the pitch credible at scale. It is also a useful contrast to the broader hyperscaler spending debate, where investors have spent months pressing Microsoft, Amazon, Alphabet, and Meta on whether their enormous AI capital outlays and cloud growth will ever convert into proportional profit.
What the 36% Pop Is Really Pricing
A one-day move of this size is rarely about a single quarter. The market is pricing a narrative shift: Snowflake as a structural AI-infrastructure winner rather than a decelerating data warehouse. The guidance raise gives that story near-term support, and the AWS alliance gives it a multiyear spine.
The skeptic’s note still applies. Consumption revenue is lumpy, a $6 billion spend commitment raises the bar on the returns the company has to generate, and the agentic-enterprise category is still more roadmap than revenue line for most of the industry. Snowflake has bought itself runway and a stronger Amazon relationship. The next few quarters will show whether the demand it is pre-funding actually shows up, or whether the market got ahead of the business again. For now, Snowflake has the one thing that was missing from the story for a year: momentum the numbers back up.