AI application giant prints cash at 51% margin while data warehouse burns through hundreds of millions of dollars

• Palantir reported a 63% increase in sales, a 51% operating margin and $540 million in free cash flow.

• Snowflake grew 29%, but had a -27% operating margin and an operating loss of $329 million.

• Palantir trades at 112 times sales with earnings. Snowflake's sales are 17 times higher, even though it is not profitable.

• A recent study identified a single habit that doubled Americans' retirement savings and turned retirement from a dream to a reality. Click here for details.

Palantir (NYSE:PLTR) and snowflake (NYSE:SNOW) reported earnings that revealed two companies attacking AI from both ends of the stack. Palantir posted 63% sales growth and an operating margin of 51%. Snowflake has grown 29%, but its operating margin is -27% and it's wasting cash.

Palantir's third quarter showed the company achieving escape velocity. U.S. commercial revenue increased 121% to $397 million. Artificial Intelligence Platform (AIP) takes customers from bootcamp to 7-figure deals within 2 months. CEO Alex Karp: “We achieved 44% growth in the world's most important market while achieving the highest Rule of 68 in the world. We intend to maintain the paradox of high margins and high growth.”

This Rule of 68 score (growth rate plus operating margin) is exceptional. Palantir generated operating income of $393 million, an increase of 248% year over year, and free cash flow of $540 million. The company's 12-month free cash flow exceeded $1 billion for the first time.

The snowflake quarters seemed solid on the surface. Sales exceeded expectations, reaching $1.21 billion, with product sales increasing 29%. Net revenue retention remained at 125% and the company added over $1 million in customers. However, Snowflake lost $329 million in sales and $288 million net. Although cash flow was positive at $138 million, the path to sustainable profitability remains unclear.

At Palantir, we believe the value of AI lies in production deployment, not model training. Shyam Sankar, CTO: “Models are converging, but the price of inference has come down significantly. This only strengthens our belief that the value is in the application and workflow layers.” AIP automates workflows. One insurance customer used 78 AI agents to reduce underwriting time from two weeks to three hours. Trinity Rail delivered $30 million in revenue impact in three months.

Snowflake owns the data warehousing layer. CEO Sridhar Ramaswamy emphasized, “There is no AI strategy without a data strategy.” Snowflake Intelligence saw strong early adoption. But with competitors like Databricks and cloud providers pushing their own solutions, Snowflake will have to convince customers to consolidate their workloads onto its platform.