Snowflake's latest quarter puts a hard number on a problem many software companies are still trying to explain: AI usage does not fit neatly into seat-based pricing.
Snowflake did not just report a strong quarter. It handed the software market a useful test case for how enterprise technology might be sold when AI starts doing more of the work that people used to do inside applications.
The company reported first-quarter fiscal 2027 revenue of $1.39 billion, up 33% year over year, while product revenue reached $1.33 billion, up 34%. That is not a small acceleration for a company already operating at scale. It also raised full-year product revenue guidance to $5.84 billion from $5.66 billion, a signal that management sees the AI demand curve as more than a one-quarter burst.
The more interesting part is not just the beat. It is what Snowflake Chief Executive Sridhar Ramaswamy says the beat proves. Snowflake has long charged customers based on consumption, meaning revenue is recognized when customers actually use its platform. In an AI-heavy software market, that distinction matters because usage can rise in ways that have little to do with how many employees have a login.
As Fortune recently noted after speaking with Ramaswamy, Snowflake's strong Q1 results helped validate the company's consumption-based pricing model at a moment when traditional software vendors are under pressure to defend per-seat contracts. That is the real story here. AI is very good at creating more work for machines. It is surprisingly bad at preserving the old logic of charging by human headcount.
For years, enterprise software pricing had a simple center of gravity. More employees using a product usually meant more seats, and more seats meant more predictable subscription revenue. That worked well when software mainly helped people click, search, approve, manage and communicate.
AI changes that relationship. A single analyst with a coding assistant can produce more work than a small team. A sales manager using an agent can query records, draft follow-ups and update systems without moving through each application by hand. A finance team can run more analysis without adding more finance employees. In that world, headcount becomes a weaker proxy for value.
This is why the pricing debate matters for founders and investors. If a software company charges by seat while its customers use AI to reduce the number of people needed for a workflow, revenue can come under pressure even if the product remains important. If the company charges by usage, the vendor can participate as workloads expand. That sounds obvious, but it is difficult to execute because usage is less smooth than a fixed annual contract.
Snowflake's quarter gives the consumption side a stronger argument. The company said its net revenue retention rate was 126%, remaining performance obligations rose 38% to $9.21 billion and 779 customers now generate more than $1 million in trailing 12-month product revenue. Those numbers suggest customers are not simply testing AI tools at the edges. They are putting more data and more workloads into the platform.
The company also said more than 13,600 accounts are using Snowflake AI capabilities, Cortex Code is in use across more than 7,100 accounts and accounts using Snowflake Intelligence more than doubled quarter over quarter. These are adoption metrics, not full proof of durable revenue, but they help explain why the market reacted so strongly. Investors have been looking for evidence that AI can expand software revenue rather than hollow it out.
The Investor Trade-Off
Consumption pricing is not a perfect answer. It can make revenue harder to forecast because customers can optimize usage, delay projects or shift workloads depending on budget pressure. That was one of the recurring concerns around cloud and data companies during the post-pandemic software slowdown. When customers tighten spending, usage-based models can feel exposed.
But AI also cuts the other way. Once companies begin building workflows around models, agents and data pipelines, usage can grow quickly. Snowflake's new five-year, $6 billion AWS agreement is important in that context. Reuters reported that the deal is designed to use AWS Graviton processors and AI infrastructure while deepening product integrations around generative and agentic AI. That tells investors Snowflake is preparing for heavier workloads, not just lighter experimentation.
There is still a cost question underneath all this. AI compute is expensive, and software companies cannot pretend the bill does not exist. Vendors building on foundation models must decide whether to absorb those costs, bundle them into premium tiers or pass them through more directly. Consumption pricing makes the economics cleaner, but it also makes customers more aware of what each action costs.
That awareness may become a competitive advantage. If a company can show customers where AI creates measurable value, charging for use feels reasonable. If it cannot, usage pricing starts to look like a meter running in the background. Snowflake's argument is that data infrastructure sits close enough to the work being done that customers can see the value in the consumption.
This is also why the broader SaaS market should pay attention. Salesforce, ServiceNow, Microsoft and a long list of younger AI startups are all dealing with the same question in different forms. The old bundle of software plus seats is being pulled apart by agents that can perform tasks across multiple systems. The companies that adapt pricing to actual work done will have more room to grow. The companies that only rename seats as AI seats may face a tougher conversation.
Snowflake has not settled the debate for everyone. One strong quarter does not remove cyclicality, competition or the risk that customers become more careful with AI spending. But it does show that enterprise buyers are willing to spend when AI is connected to real data workloads and when pricing follows usage rather than org charts.
The next thing to watch is whether Snowflake can keep turning AI adoption into durable consumption without losing margin discipline. If it can, the market will have a clearer model for the AI software era. Not every company will be able to copy it, and that is exactly the point.
Also read: Rivian says AI will make CarPlay less important in its EVs • Rivian is betting AI can make CarPlay feel unnecessary • Stable Audio Studio brings local AI sound generation into focus
