Inside TEM Stock’s Strange Year: A 36% Revenue Beat, a 7% Drop, and Cathie Wood Buying the Dip

Right now, when you bring up Tempus AI, you can see a certain type of annoyance on the faces of fund managers in the biotech industry. Typically, it begins with a small frown, followed by a protracted silence and a statement such as “The numbers are fine.” The stock simply won’t act appropriately. TEM has a market capitalization of $8.24 billion and is now trading at $46.18 on Thursday, May 14, 2026.

The income statement does not convey the same information as the 52-week range, which shows a high of $104.32 last summer and a low of $41.73 earlier this spring. Even while the company has continued to increase revenue at a rate of 36% annually and has quietly emerged as one of the more useful pieces of plumbing in American oncology, shares have lost more than half of their peak value.

Fundamentally, Tempus is a Chicago-based healthcare technology firm that was founded on the notion that data should inform cancer treatment in the same way that data informs contemporary advertising. The business creates massive multimodal datasets, uses proprietary AI models to evaluate pathology images, sequences tumor DNA, and grants access to pharmaceutical companies planning precision-medicine trials. The product mix is not typical. Diagnostics, the laboratory testing side, generated $261.1 million in Q1, up nearly 35% year over year, with oncology test volume growth of 28% and hereditary test volume up 54%.

The Data and Applications segment, which monetizes the underlying database to pharma and academic clients, brought in $87 million, growing 40.5%. The Insights product line expanded by 44.1% inside that data business. This type of growth profile would have generated a 20x sales multiple in 2023. The multiplicity has shrunk in 2026.

The actual Q1 report, which was presented on May 5, was impressive by practically all standards. In contrast to expert projections of about $337 million, revenue reached $348.1 million. The projection for the entire year was increased to $1.59 to $1.60 billion. Over the previous year, gross profit increased by 43%. Compared to a $16.2 million loss a year prior, adjusted EBITDA shrank significantly to a mere $2.8 million loss. A record $1.1 billion was reached in the total contract value.

With a net revenue retention of 126%, current clients have significantly strengthened their bonds throughout the previous 12 months. CEO Eric Lefkofsky, who founded the company in 2015 after losing his mother to breast cancer, spoke confidently on the call about the company’s data and modeling business and the deployment of more sophisticated algorithms across the platform. On May 7, the stock fell 7%.

The reasons for the post-earnings selloff sit in the texture rather than the headlines. Revenue growth, while still high, decelerated from the 83% pace shown in Q4 2025. The hereditary genetics business, anchored by the Ambry acquisition that closed in February 2025, posted volume growth of only 7% when adjusted for pre-acquisition comparisons, well below expectations.

Net loss widened to $125.9 million from $68 million a year earlier, swollen by $56.3 million in stock-based compensation and $32.3 million of unrealized losses on marketable securities. Several analysts noted that the deceleration in hereditary testing, combined with continued heavy spending, makes the path to GAAP profitability less obvious than the EBITDA trend suggests. It’s the kind of nuance the market is currently allergic to.

What’s caught the attention of contrarian investors, including Cathie Wood, is the disconnect between the operating results and the share price reaction. During the post-earnings decline, ARK Invest increased its holdings of TEM shares by almost $14.9 million. There’s a sense, talking to healthcare technology specialists, that the broader market is treating Tempus as if it were a money-losing AI hype stock when it is actually a money-losing growth-stage diagnostics company with hospital and pharma relationships that are genuinely hard to replicate. The company has reportedly sequenced more than a million tumor samples since its founding, generating a dataset that, to date, no traditional pharmaceutical company has been able to assemble on its own.

However, the bearish thesis makes sense and is worth considering. With a trailing P/E of minus 26.71, Tempus is still not profitable. The cost of stock compensation is substantial. In a competitive industry for consumer-facing testing, the hereditary genetics slowdown raises concerns about pricing pressure. Additionally, the rotation out of unprofitable growth names through 2026 has severely impacted the larger AI-healthcare category; this rotation has not yet stabilized.

Additionally, there is the straightforward issue of valuing framing. Last summer, the stock was priced at $104 per share, as if every hopeful scenario would come to pass. With limited credit for the optionality of the data company, its current price of $46 is more in line with a basic assumption. Neither of the prices has been clearly correct.

The larger context of healthcare technology is difficult to ignore. With varying degrees of success, Veracyte, Guardant Health, Natera, and Exact Sciences have all struggled with the same paradox: actual clinical adoption, real growth, but clumsy execution and unyielding unprofitability. Tempus, which has the loudest founder and possibly the most unique data asset, is in the midst of that group. Lefkofsky brings credibility and a certain Chicago-style bravado that has historically divided investors. He has established and departed firms like Groupon.

Whether TEM is being misinterpreted by a weary market or whether the deceleration fears develop into anything more systemic will probably be determined over the following few quarters. For the time being, the company continues to increase revenue at a rate that most major healthcare organizations would be jealous of, the stock continues to trade as if investors don’t quite believe it, and the most fascinating investing discussions typically take place in the space between those two images.