Revisiting the Tempus AI Short Report

Short reports are a blessing in disguise for any stock you’re holding. Someone with financial motivation and sufficient research capabilities has scoured the internet trying to find every single problem with a an asset you’ve bought with your hard-earned money. They’ve done the work for you. Even the recent short report by Michael Burry on one of the fastest growing disruptive tech stocks out there – Palantir $PLTR – raised a few good questions worth considering.

Sometimes it takes years before we know for certain whether a short thesis was barking up the wrong tree. For Tempus AI $TEM, it’s been nine months since Spruce Point Capital released a scathing short report that the company responded to with a terse reply. At that time, we sent an alert to Nanalyze Premium subscribers pointing out how pressure from their existing investor base would likely force them to correct any major problems. And if there were problems to correct, they would eventually show up in the fundamentals, in particular, revenue growth.

Summarizing The Short Report

Short reports usually contain overwhelming amounts of details to create the illusion that there are problems everywhere you look. That’s why we always start by distilling the report into a bullet list of serious accusations. We’ve done this for the Tempus AI short report as follows:

• Only 2% of revenue stems from AI applications
• Aggressive accounting and financial reporting, revenue round tripping
• Weakness in core genomics and data businesses.

These allegations are reinforced when examining the past of the founder, board members, and other executives with allegations of past associations “with troubled companies that restated financial results.” That’s the short report in a nutshell.

The most important question to ask for any short repot is simply this. If these accusations are true, where would they first start to appear in the financials? Each of the above accusations basically says the same thing. Tempus AI will not be able to keep growing revenues. When that happens, Spruce Point expects shares to fall 50 to 60 percent, which will allow them to profit since they are short the stock. Since the report was issued, shares of Tempus are down 8% compared to an S&P 500 return of +18%. However, revenue growth hasn’t skipped a beat.

The main reason for the 87% growth last year was the acquisition of Ambry Genetics. If we back that out, Tempus was still able to realize 30% organic revenue growth in 2025. Recently announced guidance for 2026 is $1.59 billion in revenue, a slight deceleration in growth, but still strongly in the double digits at 22%. Revenue growth is always the ground truth for any disruptive technology company, and Tempus is still checking all the boxes.

If the short report is to be believed, then we should see them struggle to hit these guidance numbers. As risk averse investors, we want to take the short report accusations seriously and identify metrics – aside from just revenue growth – that we can monitor going forward for signs of trouble. The first accusation is that Tempus AI isn’t doing much AI.

Is Tempus an AI Company?

Appending the letter “AI” to your company name is reminiscent of the dot-bomb times when every company appended “.com” to their name for an immediate boost in valuation. A key difference is back then, most .com companies were aspirational stories. As we always say, the ground truth that demonstrates you’re wielding AI effectively – or any disruptive technology for that matter – is always revenue growth, which Tempus has lots of. Spruce Point is likely referencing a very small segment of revenue that Tempus has specifically labeled “AI” which indeed makes up about 2% of total revenues.

At the highest level, Tempus breaks down revenues into two primary segments – “Diagnostics” and “Data and Applications.” The strong growth seen below for diagnostics – 111% growth in 2025 – reflects the acquisition of Ambry Genetics.

We can then break out “Data and Applications” into three sub segments as follows:

• Insights – linked clinical, molecular, and imaging de-identified data
• TIME – trial matching
• NEXT – proactively identifies and minimizes care gaps

The third bullet point used to be labeled “AI” and represents just a fraction of total revenues. Below you can see the growth of each segment over time.

While not explicitly labeled as “AI,” their “Insights” segment provides loads of big data that companies can access through data licenses and subscriptions. AI algorithms are only as good as the data you feed them, so companies with large proprietary data sets can help their customers make informed decisions. That would qualify as AI in our book. And when the company is posting 30% revenue growth, who really cares if it’s AI or not? It comes down to how sustainable this growth is over time. If customers find value in the data, they’ll keep purchasing it and more companies will want access to the data as well. One key metric to watch for this segment is net retention rate (NRR) which came in at 126% for 2025, down from 140% in 2024.

The second largest segment, TIME, is seeing revenues decline over the years. Tempus says this offering uses AI to match patients with trials, but it’s apparently not adding enough value such that customers want to spend more money on it over time.

Finally, there’s NEXT, the smallest segment but also the one that’s growing the quickest. Still, it only makes up 1.5% of total revenues, so it’s largely meaningless for now. While the short report correctly called this out, we would argue that the entire “Data and Services” segment appropriately describes the “AI” in Tempus AI. Let’s just hope it keeps growing.

The Bigger AI Picture

One outstanding question surrounds how exactly Tempus “does AI” at a more holistic level. The company’s business model involves selling genetic tests to patients, then licensing the data from those tests to pharmaceutical companies. However, that data is also used internally to train Tempus’s algorithms to make better recommendations for doctors.

Let’s say a patient has a cancerous tumor. The patient’s oncologist could order a Tempus test to sequence the DNA and RNA of the tumor and find out how well it will respond to certain treatments such as immunotherapy. If the results of the test aren’t conclusive, Tempus can slap on an additional Immune Profile Score (IPS) test which uses their massive database of previous immunotherapy results to determine how well a patient might respond, giving them a score from 0-100 (higher is better.) Tempus can then provide an easy conclusion: IPS-High or IPS-Low. The former would likely benefit from immunotherapy, and the latter likely won’t. And Tempus has the data to prove that it works.

This creates a positive feedback loop for the company. As doctors order more tests, Tempus captures more data, which means the results get more accurate, meaning doctors will order more tests, and so on. Not only is this beneficial for patients, but it means the tests become more valuable over time, so Tempus is able to generate more revenue per test over time.

It all sounds good on paper, but the proof is always in the revenue growth. The second accusation in the short report says that revenues aren’t all that they seem to be.

Alleged Accounting Issues

The most damning accusation from Spruce Point related to Tempus’s accounting is that they’re artificially inflating their revenues by as much as $600 million. The claim here is that Tempus counts “opt-in” options as revenue. This suggests that some customers haven’t actually paid the full amount of their “contract value,” but Tempus still records this as revenue.

Then there’s the issue of round-tripping, which is an accounting no-no. It basically means that Tempus is supposedly receiving revenue that didn’t actually come from outside the business. In this case, Spruce Point claims that Tempus invested $95 million in a joint venture with SoftBank, then licensed data to the joint venture, recording revenue of $95 million. It also doesn’t help that the joint venture’s CEO resigned after just six months.

For complex issues like this, we typically turn to the company to see what they have to say. As we alluded to earlier, Tempus addressed the short report – kind of – in a post on X.

This response doesn’t inspire much confidence. Rather than actually refuting Spruce Point’s points (Spruce points?), Tempus essentially utilized the logical fallacy of “appealing to the stone.” This is simply a dismissal of an argument as absurd without saying why it’s absurd. It’s what cheerleaders do in our comments section. We would have liked to see a more detailed refutation. Looking elsewhere, the only other rebuttal was from TD Cowen. The investment bank shrugged off the report, saying it contained no new revelations and did not fundamentally change their view on the company.

The accusations surround related party revenues which companies are required to report. Indeed, there is mention of the Softbank joint venture and the related party revenues in the financials which aren’t that substantial on their own. If there are other related party revenues that Tempus AI isn’t reporting, they would be breaking the law. It’s highly likely that ARK Invest read this short report and investigated anything that might threaten their third largest position. Unless there is blatant fraud happening – and we have no reason to believe there is – then this seems like a nonissue. The same holds true for revenue recognition. Accounting rules exist for a reason, and so do audits during which a third party accounting firm looks for such problems before they sign off on a company’s financials.

Ultimately, it comes down to Tempus AI not being able to grow in the future. The short report talks about weakness in their core business offering, but we’re just not seeing that.

Weakness in Core Business

When looking at the consistent double-digit revenue growth on display, you wouldn’t think this is a firm showing any weakness at all. According to Spruce Point, Tempus is masking their organic underperformance in their foundational businesses with inorganic additions. They cite Tempus’s May 2025 earnings report, where the company raised annual revenue guidance by $10 million, from $1.24 billion to $1.25 billion. Spruce Point believes annual revenue should have really improved by over $45 million thanks to the Ambry acquisition and a deal with Pathos AI and AstraZeneca. By raising annual guidance by only $10 million, Spruce Point thinks the company is covering up what would have been a $35 million downward revision. Even if they’re right, Tempus beat that 2025 revenue guidance by $20 million, meaning this is essentially a moot point. Tempus could have been sandbagging guidance for all we know.

Spruce Point is also critical of Tempus’s decision to acquire Ambry, sharing skepticism over the lack of synergies. We don’t disagree, especially when 70-90% of acquisitions don’t realize their intended synergies. At least this is a genetic testing company acquiring another genetic testing company; not a quantum computing firm acquiring a satellite company.

Conclusion

Many folks who are bullish on a stock will immediately dismiss any short report as “FUD” while experienced investors will see these as a blessing in disguise. Still, it’s quite challenging to dissect these unorganized verbose rants filled with all kinds of anecdotal evidence. For Tempus AI, it boils down to revenue growth that may not persist for any number of reasons. And the longer growth persists, the less likely the accusations are true. Perhaps the biggest concern we have is that the Ambry Genetics acquisition won’t lead to the synergies the company hopes for. That will probably be a main focal point when we revisit Tempus AI a year from now.