After Earnings · Nov 20, 2026 · 31m · 23m saved
With Eric Lefkofsky, Founder and CEO of Tempus AI
Tempus AI holds more than 600 petabytes of matched clinical and molecular data, and it grew by 50 petabytes last quarter alone, Eric Lefkofsky, founder and CEO of Tempus AI, told Austin Hankwitz on After Earnings. Tempus is now paying about $1.5 billion for Personalis to move into blood tests that catch returning cancer early.
Public investors mostly value Tempus as a diagnostics company that sequences patients and bills insurers. Eric says that view ignores the second half of the business, which licenses data to drugmakers and would be worth far more on its own.
"A significant number of investors historically have not appreciated the strength of our data business. They've discounted it to like zero." — Eric Lefkofsky
Eric built Tempus's data pipes into more than 8,000 hospitals, one at a time, and now runs two large clusters of Nvidia chips to train models on what comes back.
The full interview is covered here so you can skip it. 31 minutes of audio, 8 minutes of reading.
Here are the 10 takeaways that matter.
Key Takeaways
Tempus makes money two ways: sequencing cancer patients, and licensing the data to drugmakers
Its dataset is more than 600 petabytes, built through pipes into over 8,000 hospitals
The $1.5B Personalis deal is a bet on MRD, blood tests that find cancer before a scan does
Eric would not be surprised if MRD volume one day passes Tempus's therapy-selection work
AI has made patients smarter but not yet doctors, and Tempus's Hub platform is aimed at the gap
Drugmakers pay because most oncology trials fail, and Tempus data helps design better ones
Eric says investors value the data business at close to zero, while a standalone version would trade like tech
1. Sequencing Plus Data
Eric opened by explaining that Tempus runs two businesses that feed each other. One sequences patients and bills insurers. The other licenses the de-identified results to drug and biotech companies building AI models.
Eric Lefkofsky says Tempus marries sequencing with clinical records
"We make money two ways. We are one of the largest sequencers of patients, particularly cancer patients, in the world... but we also marry that data with clinical data... and we have a very big data and AI modeling business." — Eric Lefkofsky
The sequencing side began with matching patients to targeted cancer drugs. Tempus reported third-quarter revenue of $411.5 million, up 23% from a year earlier, with data and applications revenue up 33%. Adjusted EBITDA rose from $1.5 million a year ago to $25.5 million. Eric tied the data growth to drugmakers building their own large-language-model capabilities over the past three-plus years, which made them hungrier for matched datasets like Tempus's.



