- What disease does this patient actually have?
- Which treatment is most likely to work?
- Which medication carries the lowest risk?
- Which patients are suitable candidates for a clinical trial?
- How likely is a cancer recurrence?
The biggest mistake I made when I first began researching Tempus was assuming it was simply another diagnostics company. If you only glance at the financial statements, it’s an understandable conclusion. The majority of the company’s revenue still comes from laboratory testing, primarily in oncology, where physicians order molecular and genomic tests to help determine which treatments are most appropriate for individual cancer patients. On the surface, the business appears relatively straightforward: doctors order tests, laboratories process samples, reports are delivered, and Tempus gets paid. It’s a respectable business, but hardly one that immediately screams “next-generation AI platform.”
The deeper I dug, however, the more convinced I became that investors focusing solely on diagnostics are looking at the company through the wrong lens. Diagnostics are certainly important as they fund the business today.. but I increasingly believe they are merely the mechanism through which Tempus acquires something far more valuable. The company’s real asset isn’t its laboratories or even its sequencing capabilities. It’s the continuously expanding collection of proprietary healthcare data generated every time one of those laboratories processes another patient sample.
To understand why that matters, it’s worth stepping back and considering how healthcare actually functions. Modern medicine generates an astonishing amount of information. Every patient produces physician notes, pathology reports, radiology images, genomic sequencing results, laboratory values, prescriptions, treatment histories and, perhaps most importantly, long-term clinical outcomes. The problem has never been that medicine lacks data. Quite the opposite. Healthcare may be one of the most data-rich industries in the world. The problem is that almost none of those data sources communicate with one another. Hospitals operate different electronic medical record systems. Laboratories maintain their own databases. Imaging providers use entirely separate software. Researchers often receive only fragments of the overall clinical picture, while pharmaceutical companies conducting clinical trials rarely have access to the complete longitudinal history of a patient. Information exists everywhere, yet meaningful knowledge remains surprisingly scarce because the data itself remains fragmented.
That observation was the catalyst behind Tempus. When Eric Lefkofsky founded the company in 2015, he wasn’t trying to build another laboratory. His ambition was considerably broader. After his wife was diagnosed with breast cancer, he experienced firsthand how disconnected modern healthcare had become. Despite extraordinary advances in cancer research, physicians were still forced to navigate dozens of disconnected systems to assemble information that ideally should have existed in one unified view. Genetic data sat in one location, pathology reports in another, imaging somewhere else, and clinical histories somewhere else again. The technology to generate information had advanced dramatically, but the technology to organize that information had not.
Coming from the technology sector rather than medicine, Lefkofsky approached the problem differently. Before Tempus, he had already built multiple public companies, including Groupon, Echo Global Logistics and InnerWorkings. Whether investors admired every one of those businesses or not, they demonstrated that he understood something fundamental about modern software companies: data becomes exponentially more valuable when it is standardized, connected and made accessible through applications that customers actually use. Reading through Tempus’ history, I found myself repeatedly thinking about another company that many readers will already be familiar with, good ol' Palantir.
I don’t mean that comparison literally. Tempus and Palantir operate in entirely different industries and have very different economic models. But philosophically, there are undeniable similarities. Palantir’s core innovation was never creating new information. Its value came from connecting thousands of disconnected datasets into a single operational picture that governments and enterprises could actually use. Tempus appears to be attempting something remarkably similar inside healthcare. Rather than generating information for its own sake, the company is trying to create a common operating layer where genetics, pathology, radiology, clinical records and treatment outcomes all become part of the same coherent system.
This distinction also explains why I think the company’s business model is frequently misunderstood. Tempus really operates two businesses simultaneously. The first is the diagnostics operation that most investors immediately recognize. Physicians order molecular tests, Tempus analyzes the samples, and revenue is generated through laboratory services. If that were the entire business, Tempus would simply compete with companies such as Foundation Medicine, Guardant Health or Caris Life Sciences on the quality of its diagnostic assays. That is certainly a competitive market, but it isn’t particularly unique.
The second business is considerably more interesting. Every diagnostic test does something beyond generating immediate revenue. It also creates another proprietary data point that becomes part of an increasingly valuable dataset. Each patient contributes not only genomic information, but physician notes, pathology images, radiology scans, treatment decisions and, eventually, real-world outcomes. Tempus then standardizes those previously disconnected records and links them together into what the company refers to as a multimodal dataset. Over the past decade, this process has expanded into relationships with more than 5,000 healthcare institution sites and over 700 direct healthcare data connections, creating one of the largest proprietary clinical datasets available for AI development and pharmaceutical research.
That distinction may ultimately become the defining characteristic of the company. Most diagnostics businesses monetize a patient once. Tempus has the potential to monetize the information generated by that patient repeatedly. The diagnostic test becomes only the first economic event. The resulting data can subsequently be used to improve AI models, identify patients for clinical trials, support pharmaceutical research, develop predictive algorithms, and build software products that physicians use in everyday clinical practice. In other words, diagnostics fund the creation of a dataset whose value extends far beyond the original laboratory report.
This is where I believe Tempus begins to resemble a platform rather than a healthcare services company. One of my favorite characteristics in any business is the presence of a genuine flywheel.. a mechanism through which success naturally creates more success. Amazon built one through logistics. Visa built one through payment networks. Microsoft built one through software ecosystems. Tempus appears to be attempting something similar through healthcare data. Every diagnostic test expands the dataset. A richer dataset improves the quality of AI models. Better AI products create greater value for pharmaceutical companies and physicians. Those customers, in turn, generate more testing volume, which expands the dataset even further. Unlike many businesses where growth simply produces more revenue, Tempus’ growth has the potential to strengthen the underlying product itself.
Perhaps the most interesting conclusion I reached during this research process is that artificial intelligence may not even be the company’s greatest competitive advantage. AI models are improving at an astonishing pace, but they’re also becoming increasingly accessible.
OpenAI, Anthropic, Google and Meta are all investing billions of dollars into foundation models that will inevitably become available across countless industries. Algorithms alone are unlikely to remain scarce. High-quality proprietary healthcare data, on the other hand, is extraordinarily difficult to replicate. You cannot simply scrape millions of patient records from the internet or recreate decades of longitudinal clinical histories overnight. Building those datasets requires years of trust with hospitals, physicians, laboratories and healthcare systems, together with enormous investments in compliance, infrastructure and data governance. That is why I increasingly believe Tempus’ moat is less about artificial intelligence itself and far more about owning one of the richest proprietary datasets upon which future AI systems can be trained.
One announcement, more than any other, reinforced that belief. In 2025, AstraZeneca agreed to a collaboration worth roughly $200 million to develop oncology foundation models using Tempus’ proprietary data. Pharmaceutical companies do not commit nine-figure sums simply because another company claims to possess valuable information. They do so because they believe that information provides a meaningful competitive advantage. By 2026, Tempus had already delivered the first version of that oncology foundation model, transforming what could have been dismissed as a marketing announcement into tangible evidence that some of the world’s largest drug developers see genuine commercial value in the company’s data assets.
This is ultimately why I think the “Palantir of Healthcare” comparison, while imperfect, captures an important part of the story. Both companies derive their value not from producing data, but from organizing fragmented information into something customers can actually use. Where the analogy breaks down is in economics. Palantir is fundamentally a software company with exceptionally high gross margins. Tempus still operates laboratories, manages biological samples, navigates reimbursement systems and invests heavily in clinical validation. It remains a far more capital-intensive business than enterprise software. Yet that doesn’t invalidate the thesis. If anything, it simply changes the question investors should be asking. The real issue isn’t whether Tempus can become another Palantir. It’s whether the higher-margin Data & Applications business can eventually grow large enough that investors begin valuing the company as a healthcare intelligence platform that happens to own diagnostics, rather than a diagnostics company that happens to sell data. To me, that is the single most important question underpinning the entire investment thesis.
Product Advantage: Why I Think Investors Are Looking at Tempus the Wrong Way
Whenever I begin researching a new company, I always start with the same question.
Is the product actually better?
That might sound obvious, but I think it’s surprising how many investors skip this step entirely. We spend endless amounts of time discussing valuation multiples, quarterly earnings, and price targets, yet the quality of the underlying product often receives far less attention than it deserves. Over the long run, however, exceptional businesses almost always begin with exceptional products. If customers genuinely love what a company offers, revenue growth, margins, and shareholder returns have a much better chance of following.
After spending the last several weeks researching Tempus, I came away believing that the company’s competitive advantage has very little to do with any individual diagnostic test.
In fact, I don’t think the laboratory itself is really the product at all.
The product is the ecosystem.
That distinction may seem subtle, but I think it’s one of the most important observations in the entire investment thesis.
Most diagnostics companies exist to answer a single clinical question. A physician orders a test, the laboratory analyzes a sample, and the results are returned to the doctor. Once that report is delivered, the commercial relationship is largely complete until another test is ordered in the future. The diagnostic itself is both the beginning and the end of the economic transaction.
Tempus approaches the problem very differently.
Every diagnostic test is merely the first step in a much longer process. The laboratory report certainly provides immediate value to physicians treating patients, but from Tempus’ perspective it also creates something considerably more valuable: another high-quality data point that can strengthen the entire platform. Each test contributes molecular information, pathology images, physician notes, treatment decisions, radiology studies and, eventually, long-term clinical outcomes. Rather than allowing those records to remain isolated, Tempus connects them into a standardized multimodal dataset that becomes increasingly useful as additional patients move through the system.
The more I thought about this architecture, the more it reminded me of something we see repeatedly across technology.
The world’s best businesses rarely create value from a single transaction.
They create value from every subsequent interaction.
- Google became dominant because every search improved its understanding of the web.
- Netflix became better as more people watched content.
- Visa became more valuable as more merchants accepted its network.
- Amazon improved because every purchase refined its logistics and recommendation systems.
Tempus appears to be attempting something remarkably similar inside healthcare.
Every patient strengthens the product itself.
That’s an incredibly attractive business model because growth here has the potential to improve the quality of the underlying platform at the same time.
One aspect of the business that particularly impressed me was just how integrated the company has become across the healthcare value chain. Tempus generates molecular data inside its own laboratories, ingests clinical records directly from healthcare providers, digitizes pathology slides and radiology images, structures unstructured physician notes, identifies patients eligible for clinical trials, and increasingly delivers AI-powered software back into physician workflows. Each of these capabilities reinforces the others, creating a system where information continuously flows through multiple layers of the platform rather than remaining trapped in isolated silos.
This is important because artificial intelligence, despite all of the excitement surrounding large language models, is rarely constrained by algorithms anymore.
Increasingly, it’s constrained by data.
Two years ago, many investors believed the companies with the smartest AI models would inevitably dominate every industry. Today, that's no longer the case. Foundation models are improving rapidly, but they’re also becoming increasingly accessible. OpenAI, Anthropic, Google, Meta and several open-source communities are collectively pushing the entire industry forward. The intelligence itself is becoming more widely available.
What isn’t becoming more available is proprietary data.
Healthcare may be the clearest example of this phenomenon. Unlike internet content, patient records cannot simply be scraped from public websites. Longitudinal clinical histories cannot be recreated overnight. Building relationships with thousands of hospitals, laboratories and physicians requires years of investment, regulatory compliance, trust and operational execution.
This is also why I found the AstraZeneca partnership so compelling.
In 2025, AstraZeneca committed approximately $200 million to collaborate with Tempus on developing oncology foundation models using the company’s proprietary dataset. That announcement immediately caught my attention because pharmaceutical companies are extraordinarily disciplined when allocating research capital. They commit that kind of capital when they believe the underlying asset provides a meaningful competitive advantage that would be difficult (or prohibitively expensive) to recreate independently. By 2026, Tempus had already delivered the first version of that oncology foundation model, demonstrating that the partnership was producing tangible results rather than simply generating headlines.
To me, that’s one of the strongest external validations of the entire investment thesis.
Management can tell investors their data is unique.
Sell-side analysts can repeat the story.
But when one of the largest pharmaceutical companies in the world is willing to invest hundreds of millions of dollars to build AI models on top of your dataset, that carries far more weight than any investor presentation ever could.
That said, I also think investors need to be careful not to overstate Tempus’ technological advantage.
Some of the company’s individual diagnostic capabilities are not unique. Foundation Medicine, Guardant Health, Caris Life Sciences, Natera and several other competitors all offer excellent molecular testing products. If this investment thesis depended on Tempus having the single best sequencing technology, I probably wouldn’t find the opportunity nearly as compelling.
Fortunately, I don’t think that’s the thesis at all.
The competitive advantage comes from integration.
It’s the combination of diagnostics, multimodal data, artificial intelligence, physician workflow software, pharmaceutical partnerships and clinical trial infrastructure that differentiates the business. Remove one of those pieces and the platform becomes meaningfully less valuable. Put them together and they begin reinforcing one another in ways that become increasingly difficult for competitors to replicate.
Of course, there are risks to pursuing such an ambitious strategy.
One concern I have is that Tempus is expanding into an extraordinary number of adjacent markets simultaneously. Oncology diagnostics, hereditary testing, digital pathology, radiology, cardiology, pharmacogenomics, clinical trial matching, AI foundation models and minimal residual disease are all attractive opportunities individually. Collectively, however, they create a business that is remarkably broad. That breadth could eventually become one of Tempus’ greatest strengths because each business contributes additional data and distribution to the overall platform. But it could just as easily become a distraction if management struggles to integrate so many different products under one unified strategy. Building an ecosystem is considerably more difficult than building a single successful product.
Overall, though, I came away extremely impressed with what Tempus has built.
I like the company because I think management has spent the better part of a decade quietly assembling one of the richest proprietary healthcare datasets in the world while simultaneously building the infrastructure necessary to monetize that data through software, pharmaceutical research and clinical applications. In my view, that’s a much more durable competitive advantage than simply having another AI algorithm.
If I were assigning a score based solely on product quality, I’d give Tempus 8.5 out of 10.
Not perfect, because the business is still proving that its platform can consistently generate software-like economics.
But unquestionably one of the more differentiated products I’ve come across in the healthcare AI space.
Structural Tailwinds: Why I Think the Industry Is Moving in Tempus’ Direction
One lesson I’ve learned over the years is that investing becomes significantly easier when you own companies swimming with the current rather than against it. Even exceptional management teams struggle when they’re fighting structural headwinds, while average businesses can occasionally produce outstanding returns simply because they happen to operate in the right industry at the right time.
That’s one of the reasons I spend so much time thinking about long-term trends before I even look at a company’s financial statements.
- Is this market expanding?
- Are customers naturally increasing their spending?
- Is technology making the product more valuable over time?
- Or is management constantly fighting against forces outside of its control?
When I look at Tempus, I don’t see one tailwind.
I see several independent trends that all appear to be reinforcing one another.
Individually, each one is meaningful.
Taken together, they create what I believe is one of the more attractive long-term industry setups I’ve come across in recent years.
The first and perhaps most obvious trend is the continued shift toward precision medicine. For most of modern medical history, treatment decisions were largely based on broad clinical guidelines. Patients with the same diagnosis often received essentially the same therapy because physicians simply didn’t possess enough information to personalize care. That model is gradually changing. Advances in molecular biology and genomics now allow physicians to understand diseases at a much deeper level, making it increasingly possible to tailor treatments to an individual’s specific genetic profile rather than relying solely on generalized population data.
This shift may sound incremental, but I actually think it’s transformative.
Every move toward personalized medicine increases the amount of information physicians need before making treatment decisions. More genomic sequencing. More biomarker testing. More pathology data. More imaging. More longitudinal follow-up. Every additional layer of information creates more opportunities for companies capable of collecting, organizing and interpreting that data. Precision medicine increases demand for integrated data platforms that help physicians make sense of increasingly complex clinical information.
The second tailwind is something most people outside the healthcare industry rarely think about, yet it may be one of the most important developments of the past two decades.
The cost of sequencing DNA has collapsed.
When the Human Genome Project was completed in the early 2000s, sequencing an individual’s genome cost millions of dollars and required enormous amounts of time and specialized equipment. Today, that same process costs only a tiny fraction of what it once did. What was previously reserved for research institutions has gradually become accessible to hospitals and clinical laboratories around the world.
From an investment perspective, falling sequencing costs create a remarkably powerful dynamic.
Lower costs improves affordability and dramatically increases the volume of data being generated.
As genomic testing becomes routine rather than exceptional, healthcare systems begin producing exponentially larger datasets. Every newly sequenced patient represents another opportunity to better understand disease progression, treatment effectiveness and long-term outcomes. In many ways, genomics resembles cloud computing twenty years ago. As costs declined, adoption accelerated. As adoption accelerated, entirely new business models emerged that previously wouldn’t have been economically viable.
I believe something very similar is beginning to happen in precision medicine.
The third structural trend is the rapid convergence of artificial intelligence and biological research.
One of the most fascinating observations I’ve heard recently came from David Sinclair, the Harvard Medical School professor whose work focuses on aging and longevity. During a conversation on the Joe Rogan Experience, he described how artificial intelligence enabled his laboratory to complete research in only a few months that would have taken approximately 160 years using traditional scientific methods.
Biology is becoming increasingly computational.
Researchers can now evaluate millions of potential molecular interactions before ever performing a laboratory experiment. Machine learning models are capable of identifying patterns hidden inside biological datasets that would be virtually impossible for humans to detect manually. Drug discovery, biomarker identification, protein folding and treatment optimization are all becoming increasingly dependent on computational analysis rather than trial-and-error experimentation.
That changes the economics of scientific discovery as healthcare itself is becoming increasingly digital.
Only a generation ago, physician notes were handwritten. Pathology slides were examined under microscopes. Radiology films were stored physically. Medical records often existed inside filing cabinets rather than databases.
Today, almost every interaction within modern healthcare generates machine-readable information.
- Electronic health records.
- Digital pathology.
- High-resolution medical imaging.
- Wearable devices.
- Remote patient monitoring.
- Genomic sequencing.
The amount of digital healthcare information being created every single day is staggering.
Yet raw data, by itself, has surprisingly little value.
The real value comes from connecting those different pieces of information into something that produces actionable insight.
That’s where I think Tempus’ strategy becomes particularly compelling.
As more healthcare information becomes digitized, the addressable opportunity for companies capable of harmonizing that information expands naturally. Tempus benefits from a world where hospitals continue generating richer datasets because every new source of structured clinical information increases the potential value of its multimodal platform.
Perhaps the strongest tailwind of all, however, isn’t technological.
It’s demographic.
Almost every developed country is facing the same unavoidable reality.
People are living longer.
Birth rates continue to decline.
Healthcare expenditures continue to rise.
Governments, insurers and healthcare providers are all searching for ways to improve outcomes without allowing costs to spiral indefinitely.
Artificial intelligence has the potential to improve productivity across nearly every part of the healthcare system. Earlier diagnoses reduce treatment costs. Better treatment selection reduces unnecessary procedures. Faster drug discovery lowers development costs. Administrative automation allows clinicians to spend more time treating patients rather than completing paperwork.
Even modest improvements in efficiency could translate into enormous economic benefits when applied across entire healthcare systems.
That’s why I increasingly believe AI adoption within medicine will become an economic necessity.
Of course, none of these trends automatically guarantee Tempus’ success.
Structural tailwinds create opportunity but they don’t eliminate execution risk.
Many companies will attempt to capitalize on these same developments, and several very large competitors already possess significant resources. The existence of an attractive market doesn’t mean every participant will become a winner.
But that’s precisely why I find Tempus interesting.
It can become the platform that physicians, researchers and pharmaceutical companies increasingly rely upon as medicine becomes more data-driven. If healthcare truly evolves into a computational science then the companies controlling the underlying data infrastructure could ultimately become some of the most valuable participants in the entire ecosystem.
That is why, after looking at the broader industry rather than just the company itself, I came away with an even stronger conviction in the long-term opportunity.
If I were scoring Tempus purely on structural tailwinds, I’d give it 9 out of 10.
Not because success is inevitable, but because it’s difficult to imagine a future where medicine becomes less personalized, less data-driven or less dependent on artificial intelligence. Every major trend appears to be moving in the same direction.. and at least for now, that direction seems to favor exactly the type of platform Tempus is trying to build.
Management & Execution: Great Businesses Still Need Great Operators
One of the first things I look for whenever I research a company has nothing to do with financial statements.
Before I even open the income statement or look at revenue growth, I want to know one simple thing:
Is the founder still running the business?
I’ve become increasingly convinced over the years that founder-led companies possess a subtle but important advantage over professionally managed businesses. That’s obviously not true in every case, but founders often think differently than hired executives. They’re usually building something they expect to own for decades rather than simply managing quarterly earnings until their next career move. They’re generally more willing to make unpopular long-term investments, endure periods of depressed profitability, and ignore Wall Street’s obsession with the next quarter if they believe those decisions strengthen the business over the long run.
When I discovered that Tempus was still being led by its founder, Eric Lefkofsky, that immediately caught my attention.
Unlike many healthcare CEOs, Lefkofsky didn’t spend his career inside biotechnology or pharmaceuticals. His background is almost entirely in technology and entrepreneurship. Before founding Tempus, he helped build several companies that ultimately became public, including Groupon, Echo Global Logistics and InnerWorkings. Whether investors admired every one of those businesses is beside the point. Building multiple public companies requires a combination of strategic thinking, operational discipline and capital allocation that relatively few entrepreneurs ever demonstrate successfully.
Innovative businesses require someone capable of raising billions of dollars, attracting exceptional talent, integrating acquisitions, managing regulatory complexity and executing against a vision that may take a decade or more to fully materialize.
And i would say that healthcare AI is probably not a business that rewards impatience... iff anything, it punishes it.
Building trust with hospitals takes years. Developing relationships with pharmaceutical companies takes years. Creating proprietary datasets takes years. Convincing physicians to adopt new technologies takes years. This is not an industry where companies move fast and break things. Success depends on patiently compounding thousands of small operational decisions over very long periods of time.
From everything I’ve seen, Tempus appears to understand that.
One of the things that impressed me most during my research was how disciplined the company has been in building its platform. Rather than rushing to commercialize every possible AI application, management spent nearly a decade assembling the underlying infrastructure first. Laboratories were built. Provider relationships were established. Pharmaceutical partnerships expanded. Massive datasets were collected, cleaned and standardized before management seriously began emphasizing the higher-margin software and data products that now receive much of Wall Street’s attention.
To me, that’s exactly the order I’d want to see.
Far too many companies try to monetize artificial intelligence before they’ve built a meaningful competitive advantage.
Tempus appears to have taken the opposite approach.
First, build the data.
Then build the platform.
Only then begin scaling the software.
That sequencing gives me considerably more confidence that management understands where the long-term value actually resides.
The company’s execution since becoming publicly traded has also been encouraging. Revenue has continued growing at a healthy pace, oncology testing volumes have expanded consistently, adjusted EBITDA has turned positive, pharmaceutical partnerships have continued to deepen, and management has steadily increased the contribution from its Data & Applications business. None of those achievements individually prove that the long-term thesis will succeed, but collectively they paint the picture of a management team that has largely delivered on the milestones it has set for itself.
Where management becomes particularly interesting, however, is through its acquisition strategy.
Some companies acquire businesses because they run out of ideas.
Others acquire businesses because each acquisition strengthens an existing ecosystem.
I think Tempus falls much closer to the second category.
When I looked through the company’s recent acquisitions, I didn’t see a random collection of unrelated assets. I saw management systematically filling gaps within the broader platform.
Ambry Genetics expanded hereditary and germline testing, bringing additional molecular data into the ecosystem. Paige added digital pathology capabilities, allowing Tempus to deepen its presence in AI-assisted pathology workflows. Deep 6 AI strengthened the company’s position in clinical trial matching by making it easier to identify eligible patients for pharmaceutical studies. More recently, the proposed acquisition of Personalis significantly expands Tempus’ exposure to minimal residual disease, or MRD, one of the fastest-growing segments of precision oncology. Viewed individually, each acquisition appears relatively straightforward. Viewed collectively, they all reinforce the same objective: expanding the breadth of data flowing through the Tempus platform while embedding the company more deeply into physician and pharmaceutical workflows.
That strategic consistency is something I always look for.
Good acquisitions should make the existing business stronger.
Great acquisitions should make the entire platform more valuable than the sum of its individual parts.
Of course, that’s also where one of the biggest execution risks emerges.
Acquisitions almost always look compelling on investor presentations.
Integrating them successfully is an entirely different challenge.
Healthcare is already one of the most operationally complex industries in the world. Every acquired laboratory comes with different systems, different data formats, different cultures, different regulatory processes and different customer relationships. Bringing those businesses together into one seamless platform is extraordinarily difficult. Many companies have destroyed shareholder value not because they bought bad assets, but because they underestimated the complexity of integrating them.
For me, the proposed Personalis acquisition will likely become management’s biggest test yet.
Unlike earlier acquisitions, Personalis represents a much larger strategic commitment. If management succeeds, Tempus could substantially strengthen its position in one of the fastest-growing areas of oncology diagnostics while adding another important stream of longitudinal patient data to its platform. If integration proves more difficult than expected, however, shareholders may experience meaningful dilution without receiving the operational benefits management is expecting. Given the size of the transaction relative to Tempus itself, I think this is one of the most important developments investors should monitor over the next several years.
There’s another aspect of management that deserves discussion, even though it doesn’t concern day-to-day execution.
Corporate governance.
Like many founder-led technology companies, Tempus employs a dual-class share structure that gives Eric Lefkofsky voting control far exceeding his economic ownership. According to the company’s proxy filings, he controls nearly 58% of the company’s voting power, allowing him to maintain strategic control even though outside shareholders collectively own most of the economic interest.
Reasonable investors can disagree on whether that’s a good thing.
On one hand, founder control allows management to think long term without worrying about activist investors demanding higher margins next quarter or pressuring the company to sacrifice future opportunities for short-term profitability. For businesses attempting to build platforms over decades rather than quarters, that stability can become a genuine competitive advantage.
On the other hand, concentrated voting control inevitably reduces shareholder influence. If management begins making poor capital allocation decisions, overpaying for acquisitions or pursuing strategies that destroy shareholder value, outside investors have relatively little ability to intervene.
Personally, I don’t view the governance structure as a major concern today.
I’ve listened to multiple earnings calls, investor presentations and interviews with management, and I haven’t encountered any significant red flags. Quite the opposite, actually. Management strikes me as thoughtful, deliberate and appropriately focused on long-term value creation rather than short-term market expectations. That certainly doesn’t guarantee future success, but based on everything I’ve seen so far, execution has been one of Tempus’ strongest attributes rather than one of its weaknesses.
Overall, I came away with a favorable impression of the leadership team.
If I were scoring management today, I’d give Tempus 8 out of 10.
That’s a very strong score.
The reason it isn’t higher has little to do with execution so far and much more to do with the future. The company is entering a phase where integration becomes increasingly important, acquisitions become larger, and expectations rise considerably. Building the platform was difficult. Successfully integrating everything into one cohesive ecosystem may prove even harder.
Fortunately for shareholders, that’s exactly the challenge management now appears to be preparing for.
The Moat: Why I Think Data Is Becoming More Valuable Than AI
If there’s one part of the Tempus investment thesis that I think investors consistently underestimate, it’s the company’s competitive moat.
Whenever people discuss artificial intelligence, the conversation almost always revolves around the models themselves. Which company has the smartest algorithm? Which chatbot performs the best? Which model scores highest on the latest benchmark?
I think that’s the wrong conversation.
Two years ago, having a superior AI model may have represented a durable competitive advantage. Today, I’m much less convinced. Foundation models continue improving at an extraordinary pace, but they’re also becoming increasingly accessible. OpenAI continues releasing more capable versions of GPT. Anthropic is advancing Claude. Google has Gemini. Meta continues investing heavily in open-source models. Every few months the performance gap narrows a little further.
Like PLTR CEO has said on numerous occasions: artificial intelligence itself is gradually becoming a commodity.
That’s not to say innovation has stopped, however. The models will undoubtedly continue improving for years. But as the underlying intelligence becomes more broadly available, the source of competitive advantage begins shifting elsewhere.
Increasingly, I think the scarce asset isn’t the model.
It’s the data.
Healthcare illustrates this dynamic better than almost any other industry.
Building a state-of-the-art language model requires enormous computing power and exceptional engineering talent, but in principle another well-funded company can eventually do the same thing. Building one of the world’s largest longitudinal healthcare datasets is an entirely different challenge. It’s an operational problem, a regulatory problem and, perhaps most importantly, a trust problem.
Every piece of meaningful healthcare data must be collected lawfully, standardized carefully, linked correctly and maintained under some of the strictest privacy regulations of any industry. Even after that work is complete, the information has very little value unless it can be connected into a coherent picture of the patient’s medical journey.
That’s the distinction I think many investors overlook.
- Having millions of patient records isn’t particularly impressive on its own.
- Lots of organizations have millions of records.
- Insurance companies have them.
- Hospitals have them.
- Governments have them.
Connecting this data is the key.
Tempus has spent more than a decade building exactly that capability. Rather than simply storing isolated medical records, the company links a patient’s molecular profile with pathology slides, radiology images, physician notes, treatment history and, critically, long-term clinical outcomes. That creates a longitudinal view of disease progression that is substantially more valuable for both pharmaceutical research and AI development than disconnected datasets could ever be on their own.
I think that’s an important distinction because predictive medicine depends on relationships, not individual data points.
- A single CT scan has value.
- A single genomic sequence has value.
- A pathology report has value.
But when those three pieces of information are connected to the treatment a patient actually received—and then linked to whether that patient ultimately responded to therapy five years later—you begin creating something that is exponentially more useful than the individual components alone.
That is the type of information pharmaceutical companies desperately want.
It’s also the type of information that’s extraordinarily difficult to replicate.
It also creates one of my favorite characteristics in any business.. a self-reinforcing flywheel.
- Every new physician ordering Tempus diagnostics contributes another patient to the dataset.
- Every additional patient improves the quality of future models.
- Better models create more valuable products for pharmaceutical companies.
- Those products generate additional demand.
- Additional demand creates more diagnostic volume.
- And every additional diagnostic test feeds another patient back into the platform.
It’s easy to draw this flywheel on an investor slide.
It’s much harder to spend ten years actually building it.
That, in my opinion, is where Tempus deserves the most credit.
The company’s network itself has quietly become another source of competitive strength. Today, Tempus works with more than 5,000 healthcare institution sites and maintains over 700 direct healthcare data connections. Those relationships didn’t appear overnight. Every hospital integration required technical work, regulatory approvals, legal agreements and, perhaps most importantly, physician trust. Healthcare is notoriously slow to adopt new technology, but once systems become embedded into clinical workflows they also tend to remain there for very long periods of time.
That creates a subtle form of switching cost.
Replacing enterprise software is inconvenient.
Replacing software that has become integrated into clinical decision-making, laboratory workflows and hospital systems is considerably more difficult.
Hospitals don’t switch critical infrastructure lightly.
Neither do pharmaceutical companies conducting multi-year clinical research programs.
One of the strongest validations of this moat, in my opinion, comes from the caliber of Tempus’ customers rather than anything management says during earnings calls. Some of the world’s largest pharmaceutical companies already rely on Tempus’ platform for various aspects of drug discovery, clinical development and data analytics. More importantly, many of these relationships continue expanding over time, suggesting that customers are finding increasing rather than diminishing value in the platform.
That doesn’t mean the moat is unassailable.
If there’s one mistake investors frequently make, it’s assuming that a company with a strong competitive position faces no credible competition.
Tempus absolutely does.
Foundation Medicine benefits from Roche’s enormous oncology ecosystem.
Caris Life Sciences has spent years building its own molecular profiling platform.
Guardant Health is a leader in liquid biopsy.
Natera has established a strong position in minimal residual disease testing.
IQVIA possesses decades of pharmaceutical relationships and one of the largest collections of clinical research data in the world. Each competitor attacks a different part of the healthcare value chain, and none should be underestimated.
Nor should investors ignore the possibility that large pharmaceutical companies eventually decide to build more of these capabilities internally. Drug developers increasingly recognize the strategic importance of artificial intelligence, and many possess both the financial resources and scientific talent to develop proprietary models over time.
For that reason, I don’t think Tempus wins simply because it has good AI.
It wins only if it continues becoming the easiest, most comprehensive and most trusted platform through which healthcare data flows.
Tempus has quietly spent more than a decade assembling one of the richest, most interconnected healthcare datasets in the world.
If my broader AI thesis is correct and proprietary data becomes increasingly more valuable as AI models continue commoditizing.. then I believe this moat has the potential to become significantly stronger over the next decade rather than weaker.
For that reason, I’d score Tempus’ competitive moat 8 out of 10.
It isn’t immune from competition.. but it’s substantial, difficult to replicate and, perhaps most importantly, has the potential to compound over time as the platform continues growing.
Financial Trajectory: A Great Story Still Needs Great Economics
At the end of the day, every investment thesis eventually has to show up in the financial statements.
You can have an extraordinary vision, a charismatic founder and a product that changes an entire industry, but if the economics never materialize, shareholders rarely benefit. Markets are remarkably patient with companies that are investing for the future but that patience isn’t infinite. Eventually every business has to demonstrate that it can convert innovation into cash flow.
That’s why I always spend a considerable amount of time looking beyond the narrative.
- Is revenue actually growing?
- Are margins improving?
- Is the business becoming more scalable?
- Is management creating shareholder value or simply growing for the sake of growth?
When I look at Tempus today, I see a company that is making meaningful progress, but one that is still somewhere in the middle of its journey rather than at the end.
The headline numbers are certainly encouraging.
During the second quarter of 2026, Tempus generated approximately $382 million in revenue, representing 22% year-over-year growth, while management once again raised full-year guidance to roughly $1.6 billion. In an environment where many healthcare companies are struggling to sustain double-digit growth, those are respectable numbers and suggest that demand for the company’s products remains healthy.
Looking a little deeper, however, I think the most interesting story is where that revenue is coming from.
Diagnostics continues to represent the majority of the business and remains the engine funding the entire platform. Oncology testing, in particular, continues to perform exceptionally well, with testing volumes growing more than 30% year over year. That growth matters because every additional diagnostic test expands the proprietary dataset that underpins the broader investment thesis.
As encouraging as the diagnostics business has been, it’s not actually the segment I’m watching most closely.
The business I care about is Data & Applications.
If diagnostics represent the engine collecting data, Data & Applications represent the mechanism through which that data begins generating software-like economics.
That distinction is incredibly important.
Laboratory testing is fundamentally a services business. It requires personnel, laboratory equipment, consumables and ongoing operational costs. Software, by contrast, is inherently more scalable. Once a platform has been built, every additional customer can often be served at significantly higher incremental margins.
That’s exactly why I believe the mix shift inside Tempus matters so much.
During the second quarter, the Data & Applications segment grew approximately 28% year over year, while the higher-margin Insights business expanded by roughly 36%. Those growth rates comfortably exceeded the overall company’s revenue growth and, in my opinion, provide some of the earliest evidence that management’s long-term strategy is beginning to take shape.
If that trend continues over the next several years, the financial profile of Tempus could look dramatically different than it does today.
That’s an important point because I don’t think investors should value Tempus based solely on what the business looks like today.
They should value it based on what the revenue mix may eventually become.
If Data & Applications gradually evolves from a relatively small contributor into one-third or perhaps even half of total revenue, the company’s margin profile could improve substantially. Investors often underestimate how powerful these mix shifts can become. You don’t necessarily need the diagnostics business to become dramatically more profitable if the software and data businesses continue compounding at faster rates.
We’re already beginning to see small signs of that operating leverage.
Gross profit continued expanding faster than revenue, while consolidated gross margins reached approximately 64%, suggesting that the business is gradually becoming more efficient as higher-value products account for a larger share of overall sales. That’s exactly the direction I’d expect to see if management is successfully transitioning from being primarily a diagnostics company toward becoming a healthcare intelligence platform.
Having said all of that, I also think this is where investors need to be careful not to get ahead of themselves.
One of the easiest mistakes investors can make is looking at a single quarter’s GAAP earnings and concluding that profitability has finally arrived.
I don’t think we’re there yet.
At first glance, Tempus reported positive GAAP net income during the quarter, which naturally generated a fair amount of excitement. Digging into the financial statements, however, tells a more nuanced story. A significant portion of that profit resulted from unrealized gains on marketable securities rather than improvements in the underlying operating business. In other words, the headline number arguably paints a more optimistic picture than the core business currently deserves.
That doesn’t invalidate the progress management has made but investors need to distinguish between accounting profitability and operational profitability.
Personally, I’m far less interested in whether a company reports positive earnings for a single quarter than I am in whether the underlying economics are consistently improving.
That’s why I pay much closer attention to operating cash flow and, ultimately, free cash flow.
Today, Tempus still has work to do.
Operating cash flow remains negative, and while the trajectory is improving, the business has not yet demonstrated the type of durable cash generation that characterizes mature software platforms. Again, I don’t necessarily view that as a reason to avoid the company. Tempus is still investing aggressively in growth, acquisitions and platform development. But it does mean the investment thesis remains partially dependent on management successfully executing over the next several years rather than simply harvesting profits today.
Another area I’ll continue watching closely is stock-based compensation.
This is one metric I think investors frequently dismiss too casually.
Companies often exclude stock-based compensation when presenting adjusted profitability metrics, arguing that it’s a non-cash expense.
Technically, that’s true. Economically, I think it’s much more complicated.
Stock-based compensation represents real dilution for shareholders.
Every share issued to employees reduces the ownership percentage of existing investors.
During the first half of 2026, Tempus recorded more than $100 million in stock-based compensation, a figure that increased meaningfully from the previous year. While I understand why rapidly growing technology companies rely on equity to attract exceptional talent, I also believe investors should treat dilution as a genuine economic cost rather than simply ignoring it because adjusted EBITDA excludes it.
Fortunately, the balance sheet appears considerably stronger than it did before the IPO.
Tempus finished the quarter with more than $800 million in cash and marketable securities, providing management with meaningful financial flexibility as it continues investing in the platform and pursuing acquisitions. That liquidity should allow the company to continue executing its long-term strategy without facing immediate financing pressure, although investors should continue monitoring how future acquisitions affect both cash balances and shareholder dilution.
Looking ahead, there are several milestones that would substantially increase my confidence in the long-term financial story.
First, I’d like to see the Data & Applications segment consistently growing above 30% while representing an increasingly larger share of total revenue. To me, that’s the clearest evidence that Tempus is successfully transitioning toward a more software-oriented business model.
Second, I’d like to see operating cash flow and, eventually, free cash flow turn sustainably positive without relying on accounting gains or external financing.
Third, I want to see stock-based compensation gradually become a smaller percentage of revenue as the business matures.
And finally, I want management to demonstrate that acquisitions such as Ambry, Paige, Deep 6 AI and Personalis aren’t simply adding reported revenue but are actually accelerating organic growth across the broader platform through cross-selling and deeper customer relationships.
From a valuation perspective, I also don’t think Tempus appears wildly expensive given its current growth profile.
Consensus expectations currently point toward approximately $2 billion of revenue in 2027. With a market capitalization of roughly $12 billion, investors are paying a price-to-sales multiple that, while certainly not cheap, doesn’t appear unreasonable compared to other companies attempting to build category-defining AI platforms. In other words, I don’t think investors are being asked to pay an absurd valuation for perfection today. Instead, the market appears to be assigning Tempus a premium that reflects significant future potential while still acknowledging the considerable execution risk that remains.
Overall, I came away with mixed but ultimately positive impressions.
- The business is clearly moving in the right direction.
- Revenue growth remains healthy.
- Margins are gradually improving.
- The higher-margin software business is beginning to contribute more meaningfully.
- The balance sheet is strong.
At the same time, I don’t think the financial story is finished.
Cash generation still needs to improve.
Stock-based compensation remains elevated.
The platform still has to demonstrate that it can consistently produce software-like economics rather than simply software-like narratives.
For those reasons, I’d currently score Tempus’ financial trajectory 7 out of 10.
I think the most exciting part of the financial story still lies ahead. If management successfully executes over the next several years and the Data & Applications segment becomes the dominant driver of value creation, I could easily see this becoming one of the strongest aspects of the investment thesis. Today, however, I think the numbers still need to catch up to the vision.
What Could Go Wrong? Understanding the Risks
Whenever I finish researching a company, I deliberately force myself to switch perspectives.
For a while, I stop thinking like a shareholder and start thinking like a short seller.
If someone were trying to convince me not to invest in this business, what arguments would they make? Which assumptions in my investment thesis are most vulnerable? What would have to happen for me to admit that I was simply wrong?
I’ve found this exercise incredibly valuable over the years because it’s surprisingly easy to fall in love with a compelling narrative. Every management team has a great vision. Every investor presentation paints an optimistic future. The difficult part isn’t finding reasons to be bullish. The difficult part is honestly evaluating the risks that could prevent that future from ever materializing.
Tempus is no different.
In fact, despite how optimistic I am about the long-term opportunity, I think there are several risks investors need to take very seriously.
The first is reimbursement.
This may not be the most exciting topic to discuss, but it could ultimately become one of the most important.
Today, diagnostics remain the financial engine supporting the entire Tempus ecosystem. Every genomic test generates immediate revenue while simultaneously feeding new data into the platform. That works extremely well as long as insurers, Medicare and other healthcare payers continue reimbursing those tests at attractive rates. If reimbursement policies become less favorable, however, the impact extends well beyond laboratory revenue. Fewer reimbursed tests mean fewer patients entering the ecosystem, slower data collection and, ultimately, a weaker data flywheel. In other words, changes to reimbursement policy could negatively influence the long-term growth of the platform itself.
The second major risk is acquisitions.
Earlier in this article I explained why I generally like Tempus’ acquisition strategy. Every transaction appears strategically consistent with management’s long-term vision of building a comprehensive healthcare intelligence platform.
That doesn’t mean every acquisition will create shareholder value.
Healthcare integrations are notoriously difficult. Different laboratory systems, different software architectures, different corporate cultures and different customer relationships all need to be combined into one unified organization. It’s one thing to acquire a promising company but it’s something entirely different to successfully integrate it while preserving customer relationships and continuing to innovate.
This is why I believe the Personalis acquisition deserves particularly close attention.
Strategically, I understand exactly why management wants the business. Minimal residual disease represents one of the fastest-growing areas in oncology diagnostics, and adding another stream of longitudinal patient data strengthens the broader platform.
Financially, however, it’s a much larger commitment than previous acquisitions.
If management successfully integrates Personalis, the transaction could significantly deepen Tempus’ competitive position.
If integration proves more difficult than expected, shareholders may find themselves absorbing dilution without receiving the strategic benefits that originally justified the deal.
Competition is another risk that shouldn’t be underestimated.
One criticism I occasionally see from overly enthusiastic investors is the assumption that Tempus somehow has the healthcare AI market to itself. Nothing could be further from the truth.
Healthcare is one of the largest industries in the world, and naturally it attracts some of the largest and best-capitalized competitors.
Foundation Medicine continues benefiting from Roche’s enormous oncology ecosystem.
Guardant Health remains a leader in liquid biopsy.
Caris Life Sciences has spent years developing its own molecular profiling platform.
Natera continues expanding aggressively in minimal residual disease.
IQVIA possesses decades of pharmaceutical relationships together with one of the world’s richest collections of clinical research data.
Each competitor approaches the market from a slightly different angle, but collectively they ensure that Tempus will need to continue executing at a very high level simply to maintain its current position.
There’s another competitive risk that I think receives less attention.
Large pharmaceutical companies are becoming increasingly sophisticated users of artificial intelligence themselves.
Today, many rely on partners like Tempus for data, analytics and model development.
Will that always remain the case?
I’m not entirely sure.
As AI capabilities continue improving, it’s entirely possible that some pharmaceutical companies decide to internalize more of these functions rather than relying on external platforms. Others may diversify across multiple vendors rather than concentrating their research efforts with a single provider.
That doesn’t necessarily invalidate the investment thesis but it simply means Tempus can’t afford to become complacent.
The company must continue proving that working with its platform creates more value than building those capabilities internally.
Another area I continue watching closely is data itself.
Ironically, the very asset that makes Tempus so attractive also creates one of its biggest risks.
The entire platform depends on the company’s ability to legally collect, standardize and commercialize enormous quantities of healthcare information. Privacy regulations continue evolving around the world, public attitudes toward health data may change, and hospitals themselves may become more protective of the information they generate. If access to patient data becomes more restricted or substantially more expensive, the economics of the business could look very different over time.
Financial execution also remains an important risk.
Although I believe the company is moving in the right direction, Tempus has not yet demonstrated the kind of durable free cash flow generation that characterizes mature software platforms. Stock-based compensation remains elevated, acquisitions continue consuming capital and management still has considerable work ahead before the platform consistently produces software-like economics. The story is improving but it’s still a story that needs further financial proof.
Finally, there’s valuation.
Ironically, this may be the risk that matters most for investors buying shares today.
The question isn’t whether Tempus becomes a successful company.. I think it probably will.
The real question is how the market ultimately chooses to value that success.
If investors continue viewing Tempus primarily as a diagnostics business, the valuation multiple is likely to remain much closer to traditional healthcare companies. Diagnostics businesses generally command respectable multiples because they’re capital-intensive, reimbursement-dependent and operationally complex.
If, however, investors increasingly begin viewing Tempus as a healthcare intelligence platform.. a company whose long-term value comes from software, proprietary data and AI-driven applications then the valuation framework changes considerably.
That’s the bet you’re ultimately making as a shareholder.
After spending several weeks researching Tempus, none of these risks are large enough to invalidate my investment thesis.
But they are significant enough that I wouldn’t ignore them.
In fact, I think one of the biggest mistakes investors make is believing that owning a great company means there are no risks.
Every investment has risks.
Tempus still has to prove that it can transform an extraordinary platform into extraordinary economics. Until that happens, I think investors should view the company as a high-potential business that remains very much in the execution phase rather than a fully mature compounder.
Final Thoughts: So Where Do I Land on Tempus?
After spending the last several weeks researching Tempus, I genuinely think it’s one of the most interesting AI companies I’ve looked at in quite some time.
I believe it’s attempting to build something that very few companies in healthcare have successfully built before.. a platform where diagnostics, proprietary data, artificial intelligence and pharmaceutical research all reinforce one another.
The bull case is relatively straightforward. If healthcare increasingly becomes driven by artificial intelligence, computational biology and precision medicine (as I believe it will) then Tempus has a real opportunity to become the data and intelligence layer connecting hospitals, physicians, researchers and pharmaceutical companies. In that world, diagnostics simply become the engine that continuously feeds an expanding proprietary dataset. As the higher-margin Data & Applications business grows, margins improve, cash flow strengthens and the market gradually begins valuing Tempus less like a diagnostics company and more like a software and data platform.
The base case, in my opinion, is still a very attractive business. Diagnostics continue growing, the software business steadily becomes a larger percentage of revenue, margins gradually improve and Tempus develops into a high-quality healthcare technology company. It may never achieve software-like economics comparable to companies such as Palantir, but it still becomes an important player in the future of precision medicine and generates respectable long-term shareholder returns.
The bear case is equally worth considering. Management could struggle integrating acquisitions, reimbursement dynamics could become less favorable, competitors could narrow Tempus’ data advantage and AI adoption across healthcare could progress much more slowly than investors currently expect. Revenue might continue growing, but shareholders may never receive the software-like margins or valuation multiple that today’s bulls are expecting. That’s an important distinction because great businesses don’t always become great investments if expectations become too optimistic.
This ultimately brings me back to the question I asked at the beginning of this article: what exactly is Tempus becoming?
If investors continue viewing the company primarily as a diagnostics business, then today’s valuation may prove entirely reasonable. But if diagnostics ultimately become nothing more than the customer acquisition engine for a much larger healthcare intelligence platform, then I think we’re looking at a very different business than the financial statements currently suggest. Personally, I increasingly believe that’s where the company is headed.
That doesn’t mean I’m rushing out to buy the stock tomorrow morning.
In fact, I’m probably doing the opposite.
The shares have enjoyed a very strong run recently, helped by optimism surrounding the Moderna partnership and a broader re-rating across AI healthcare and biotechnology companies. While I remain constructive on the long-term story, I’ve learned over the years that patience is one of the most valuable qualities an investor can have. Markets rarely move in straight lines. Five, ten and even fifteen percent pullbacks happen regularly, even in outstanding businesses, and those are usually the opportunities I prefer to take advantage of.
My current plan, assuming nothing materially changes in the investment thesis, is to begin selling cash-secured puts around the $60 strike on a meaningful market pullback. If I’m assigned shares at that level, I’d be very comfortable becoming a long-term shareholder.
I want to emphasize the words long-term.
This isn’t a trade for me.
If I eventually buy Tempus, I expect to measure that investment in years rather than quarters.
I also recognize that it’s a more speculative investment than companies like Nebius or NVIDIA. Those businesses benefit from AI infrastructure demand that is already visible today and, in my opinion, easier to model. Tempus requires a greater leap of faith. The pace of AI adoption across healthcare, the commercialization of its data platform and the long-term economics of the business all remain less certain.
Ironically, that’s also where I think the opportunity exists.
The market rarely offers exceptional returns on businesses whose future is already obvious. The most attractive investments often involve identifying companies where the destination appears compelling, even if the path remains uncertain.
That’s exactly how I view Tempus today. I don’t think management has proven every part of the thesis yet. The company still needs to demonstrate stronger cash generation, continue growing its higher-margin software business and successfully integrate its recent acquisitions. But I also believe they’ve quietly spent the better part of a decade building one of the richest proprietary healthcare datasets in the world, and if my broader AI thesis is correct, if artificial intelligence truly transforms medicine, drug discovery and biology over the next decade, then companies controlling that data infrastructure could become some of the biggest beneficiaries of the entire AI revolution.
For that reason, Tempus is officially going onto my buy list.
I’ll be watching every quarterly report closely, paying particular attention to the growth of the Data & Applications segment, operating cash flow, free cash flow and how successfully management integrates Personalis into the broader platform. Unless something materially changes in the investment thesis, I fully expect to become a shareholder the next time the market gives me an opportunity.
Whether Tempus ultimately becomes the “Palantir of Healthcare” remains to be seen.
But after completing this research, I can say one thing with confidence.
It’s one of the most fascinating companies I’ll be following over the next decade.
