Snowflake shares have fallen 35.4% this year, even as the company said it added 740 net new customers in the fourth quarter of fiscal 2026 and pushed deeper into artificial intelligence products. The stock has lagged both the broader Zacks Computer and Technology sector, which rose 10.6%, and the Zacks Internet Software industry, which fell 12.1%.
The drop has come despite a steady rise in usage across the platform. Snowflake said it had 733 customers spending more than $1 million annually in the fourth quarter, up from the prior period, and 56 customers spending more than $10 million a year. Net new customers increased 40% year over year, a sign that demand is still building even as investors question how quickly that growth will translate into returns.
That tension is clearest in the company’s margins. Snowflake said its free cash flow margin took a 150-basis-point hit in the fourth quarter because of the Observe acquisition, and it also faces pressure from the lower-margin profile of new AI products and the infrastructure costs that can come with them. The company has been trying to balance that burden with faster product rollout, launching more than 430 capabilities in 2026 and announcing major updates to Snowflake Intelligence and Cortex Code in April 2026.
Snowflake said Snowflake Intelligence had been adopted by more than 2,500 accounts within three months of launch, while Cortex Code had been embraced by more than 4,400 customers. Those gains matter because the company is still trying to prove that its AI push can hold up against heavier spending from Oracle, Amazon and Alphabet, all of which are expanding their footprint in the AI space. For now, Snowflake is telling the market to expect product revenue of $1.262 billion to $1.267 billion in the first quarter of fiscal 2027, which implies year-over-year growth of 27%, along with an operating margin of 9%.
Analysts are looking for $1.32 billion in fiscal first-quarter revenue and earnings of 32 cents per share, leaving Snowflake with a familiar challenge: show that product adoption can keep climbing fast enough to satisfy a market that has already marked down the stock. The next test is whether that growth can keep outrunning the cost of building the AI business around it.
