Snowflake CEO Frames Risks From Agentic AI and SaaSpocalypse

Forbes reports Snowflake CEO Sridhar Ramaswamy cautioned at Snowflake Summit 26 that the software industry's "SaaSpocalypse" may not be over, saying "We spend a lot of time thinking about what durable value is" as agentic AI lowers the cost of building software. Forbes and Fortune report Snowflake's latest quarter delivered $1.39 billion in revenue, up 33% year-over-year, and Forbes says the company committed $6 billion to AWS for Graviton compute and AI capacity and rebranded its agentic products to CoWork and CoCo. CNBC and Fortune tie the results to a broad software-stock rally in late May. Forbes also cites accounts of Microsoft trimming Claude-powered coding agents as bills rose and Uber's CTO disclosing high agentic-AI token costs, framing consumption-based pricing as central to how vendors absorb volatile AI costs.
What happened
Forbes reports Snowflake CEO Sridhar Ramaswamy used Snowflake Summit 26 to warn that the industry's "SaaSpocalypse" may not be over, asking how companies will define durable value when agentic AI lowers the cost of building software. "We spend a lot of time thinking about what durable value is," he said, per Forbes. Forbes also reports Snowflake rebranded its agentic products to CoWork, for knowledge workers, and CoCo, for developers, renaming its Snowflake Intelligence and Cortex Code surfaces.
The earnings and cloud commitment
Forbes and Fortune report Snowflake's most recent quarter delivered $1.39 billion in revenue, up 33% year-over-year, with Forbes citing remaining performance obligations of $9.21 billion and 779 customers generating more than $1 million in trailing-12-month product revenue. Forbes reports the company committed $6 billion to AWS for Graviton compute and AI capacity, its largest infrastructure commitment. CNBC and Fortune link the quarter to a broad software-stock rally in late May, with CNBC reporting the iShares software ETF rose 21% that month.
Why pricing models matter
Agentic workflows string many model calls together, and per-token billing scales with that volume, creating a two-sided cost problem of rising inference spend and immature agent guardrails. Vendors with consumption-based pricing can convert variable AI usage into revenue, while seat-based models face pressure when individual productivity gains do not justify legacy per-seat premiums. Forbes cites accounts of Microsoft pulling back on Claude-powered coding agents as bills rose and Uber's CTO disclosing high agentic-AI token bills earlier in 2026, illustrating how differently peers are managing the same cost curve.
What to watch
Track adoption curves and per-workload token spend for agentic products, whether more vendors shift from seat-based to consumption-based billing, and large cloud or chip commitments that aim to stabilize unit economics for inference and training. For platform teams, instrumentation of AI consumption and guardrail tooling will increasingly determine the cost-benefit of agentic features.
Key Points
- 1Consumption-based pricing lets vendors align revenue with volatile AI compute usage, a contrast to seat-based models squeezed when per-seat premiums outpace productivity gains.
- 2Snowflake's $1.39B quarter (up 33%) and $6B AWS Graviton commitment, per Forbes and Fortune, show how large cloud deals reshape model-inference unit economics.
- 3Agentic AI lowers the bar to build software but multiplies token consumption, making usage telemetry and governance first-order engineering concerns.
Scoring Rationale
Snowflake's strong quarter ($1.39B, up 33%), its $6 billion AWS Graviton commitment, and the CEO's SaaSpocalypse commentary tie one company's results to a broader debate over consumption pricing and agentic-AI costs that affects enterprise procurement and platform economics. It is a notable, well-corroborated business story rather than a frontier-model or regulatory event.
Sources
Public references used for this report.
Practice with real SaaS & B2B data
90 SQL & Python problems · 15 industry datasets
250 free problems · No credit card
See all SaaS & B2B problems