Tempus AI Grew Revenue 83% But Is Still Burning $112 Million a Year in Cash

NEW YORK, April 10 —

Tempus AI has beaten earnings estimates for four straight quarters, improving EPS from -$0.24 to -$0.04 in less than a year. Wall Street responded with a consensus price target of $75.43 — 71% above today's $44.16. There's one problem: the company burned $112 million in cash over the trailing twelve months. Nothing about the latest Gilead collaboration suggests that number is about to shrink.

What the Street Believes

The sell-side story on Tempus is simple. Revenue nearly doubled. Losses are narrowing fast. The company sits at the intersection of AI and healthcare — two sectors investors will pay almost anything for. Analysts have set a mean target of $75.43. Cathie Wood's ARK Innovation ETF has been a prominent holder, giving the stock a growth-at-any-price sheen. The thesis: Tempus is building a flywheel where genomic testing generates data, data attracts pharma partners, and pharma partnerships fund more testing. Give it time and the margins will come.

But nobody on the sell side is answering this question: if revenue grew 83% and gross margins are 62.7%, where did the cash go? A company with $1.3 billion in revenue and nearly two-thirds gross margin should be generating cash, or at least approaching breakeven. Instead, Tempus burned $112 million. That gap between reported earnings improvement and actual cash generation is the entire story. The consensus is ignoring it.

What the Data Actually Shows

Follow the EPS trajectory and Tempus looks like it's sprinting toward profitability. Four quarters ago, the company lost $0.24 per share versus a $0.27 estimate. Three quarters ago, -$0.22 versus -$0.25. Two quarters ago, -$0.11 versus -$0.18 — a 40% beat. The improvement looks like operating leverage, like a company growing into its cost structure. Free cash flow tells a different story.

At -$112 million in trailing free cash flow, Tempus is burning cash at roughly the same rate it was when revenue was far smaller. The EPS improvement is driven largely by non-cash items, not actual dollars staying in the business. Stock-based compensation is the usual suspect. It reduces GAAP losses by attracting talent without cash outflows, but does nothing for the bank account. Think of a restaurant that reports lower food costs because the chef agreed to work for equity instead of salary. The P&L improves. The kitchen still needs cash to buy ingredients.

"Tempus will provide real-world evidence analytics to support Gilead's oncology pipeline development."

The Gilead collaboration, announced this week, fits a pattern. Tempus has been signing pharma partnerships at an accelerating pace, each framed as a "strategic collaboration." The critical question: do these deals represent recurring platform revenue — the kind that renews annually and compounds — or episodic data licensing, where Tempus sells access to a dataset once and then needs to find the next buyer? The distinction matters for valuation. A SaaS-like data platform deserves 10-15x revenue. A consulting shop that monetizes proprietary datasets one deal at a time deserves 3-5x. At $44.16, the market is pricing Tempus closer to the former.

Here's the math: $1.3 billion in revenue at 62.7% gross margin produces roughly $815 million in gross profit. To still burn $112 million in free cash flow after generating $815 million in gross profit, operating expenses and capital needs must consume over $900 million a year. Revenue can double and Tempus still won't see positive cash flow unless the cost structure changes fundamentally.

Why This Changes Everything

The consensus $75.43 target implies Tempus will be worth roughly $13 billion. For that to make sense, the company needs to reach positive free cash flow within two to three years. That requires not just continued revenue growth, but a sharp shift in the relationship between revenue and cash. Right now, every additional dollar of revenue costs Tempus more than a dollar to deliver. Growth is not self-funding.

Watch the FCF line, not the EPS line. If Tempus reports another quarter of narrowing GAAP losses while free cash flow stays flat or worsens, that confirms the earnings improvement is a non-cash phenomenon. If FCF loss narrows to -$50 million or better over the next two quarters while revenue keeps growing, the operating leverage thesis is real and the stock is cheap. Until then, the $75 target is pricing in a margin structure that has never existed at this company.

The revenue mix question is equally pressing. If Tempus breaks out its genomics testing revenue from pharma data licensing revenue in future filings, investors will finally see which business is driving growth. If most of the 83% growth came from lumpy pharma deals rather than recurring diagnostic testing, the multiple contracts immediately. Pharma collaborations are project-based. They end. And when they do, the comp gets harder, not easier.

The Bull Case

The bulls aren't wrong about everything. 83% revenue growth is extraordinary at any scale, and $1.3 billion is real scale. The genomics-to-data flywheel, if it works, creates a competitive moat that is genuinely hard to replicate. Tempus holds more clinical data than almost anyone outside of major hospital systems. That data appreciates as the AI models trained on it improve. If the cash burn is a land-grab investment — spending now to lock in data advantages that pay off for decades — then -$112 million is a bargain.

The EPS trajectory, even if partially non-cash, does show the loss rate decelerating. A company going from -$0.24 to -$0.04 per share is doing something right. The question is whether "something right" means approaching sustainable profitability or managing the optics of the income statement. The FCF number says it's closer to the latter. But one strong cash flow quarter could change the narrative entirely.

The Bottom Line

Tempus AI is a $1.3 billion revenue company that cannot generate cash. The four-quarter EPS improvement from -$0.24 to -$0.04 is real on paper, but -$112 million in free cash flow shows the path to profitability is further out than the income statement suggests. Every new pharma "strategic collaboration," including the Gilead deal, should be evaluated not for its headline value but for whether it represents repeatable, platform-driven revenue or a one-time data sale. Until Tempus proves its unit economics work without episodic pharma deals, the $75 consensus target is a bet on a margin structure that doesn't exist yet. The stock gets interesting below $30, where the growth alone might justify the price without requiring a profitability leap of faith. Run the free Tempus AI, Inc. deep-dive →

Basis Report does not hold positions in securities discussed. This is not investment advice.