$TEM: The Medicare Raise Nobody Is Pricing In

The 60 second version

• Tempus AI runs cancer genomic tests that Medicare and insurers pay for one test at a time. It also licenses the deidentified data from those tests to drug makers.
• The shift is that precision oncology is moving past a single tumour profile at diagnosis toward repeated monitoring for residual disease, and reimbursement is finally catching up.
• We place the oncology testing franchise in Flight. The constraint easing is price per test, because the May 29, 2026 FDA approval moves the whole solid tumour DNA portfolio onto one Medicare advanced diagnostic rate.
• The evidence from Q2 2026: oncology volume up 31%, MRD tests up from 6,500 to 9,000 in a single quarter, about $200 million of new data bookings, and the first GAAP profit.
• The next test is Q3 results on November 2, 2026. We want oncology volume growth of at least 28% and at least 11,000 MRD tests.
• Three things would break the thesis: revenue under $400 million, oncology volume growth below 22%, or the 2027 pricing uplift slipping out of 2027.

An oncologist ordering the new version of xT CDx sends one sample, the tumour, and nothing else. Until May 29, 2026, the version that earned the top Medicare rate also needed a blood or saliva sample from the patient. Getting rid of that second tube looks like a small convenience. For Tempus it works as a price increase.

Tempus says the approval moves its whole solid tumour DNA testing portfolio onto unified pricing under Medicare’s advanced diagnostic laboratory test category. Management puts the gain at roughly $200 more per test, or about $85 million a year starting in 2027. No new lab, sales hire or reagent kit is needed to collect it.

The stock still fell about 20% in the month before we wrote this, and at $67.62 the company is worth about $12.2 billion. Our thesis in one sentence: Tempus has become a cancer lab with a regulator approved price rise already on the calendar and a monitoring test growing about 38% a quarter, and today’s price requires only that this keeps going, with little cushion either way.

The fair reading is that this is a central case and not a bargain. The price is not assuming decline. It is assuming success, and the next three months will show whether that success is on track.

The shift

For about a decade, cancer genomics was a one shot purchase. A patient was diagnosed, a sample was sequenced, a report came back with mutations and matched drugs, and the lab got paid once. That model reached the large academic centres years ago. In community oncology, where most American cancer patients are treated, broad profiling has spread more slowly, and the reason has mostly been money and logistics.

Reimbursement was the binding constraint. Medicare’s advanced diagnostic category pays labs a set rate for tests that are unique and clinically proven, and Tempus’s xT CDx was set at $4,500. Only the tumour normal version qualified. Every test run without the matched normal sample got paid less, which left the blended oncology average selling price near $1,850.

Two things changed in 2026. The May approval closed the gap between the tumour only and tumour normal versions. The second change came more quietly: Tempus’s MRD test, NeXT Personal, gained reimbursement in breast cancer, non small cell lung cancer and immunotherapy monitoring. MRD means molecular residual disease, the leftover cancer DNA in blood after treatment. A patient who gets one profile at diagnosis may get monitored a dozen times over the following years.

That second point matters more over five years than the first does over one. A lab that sells one test per patient is in a volume business. A lab that sells recurring monitoring has a patient base that renews.

The company

Tempus runs three businesses that feed one another. The first is oncology testing: xT CDx profiles 648 genes in solid tumour tissue, xF does the same from a blood draw, and NeXT Personal and xM watch for recurrence. The second is hereditary testing through Ambry, bought in 2025, which checks inherited cancer risk. The third, which the company calls Insights, licenses deidentified, outcome linked patient data and builds AI models for drug makers. Customers include $AZN, $GSK, $BMY, $MRK, $BNTX, Daiichi Sankyo and $INCY, all on multiyear contracts.

The company was founded by Eric Lefkofsky, who earlier started Groupon, and he still writes the quarterly letter to shareholders. Those letters read like a founder’s, impatient and specific, and they explain a lot about why the company spent years building data assets before it made money. (An aside: the “AI” in the name has probably cost the stock more than it has helped. It draws in money that trades on chatbot headlines and leaves on the first lumpy bookings quarter. The lab underneath is far less exciting than the label, and far more predictable.)

The detail worth knowing is in the Q2 income statement. GAAP net income of $5.6 million was the company’s first profit, and it included a $98.5 million unrealised gain on investments. Take that out and the operating business still lost money on a GAAP basis. Adjusted EBITDA of $8.0 million against a loss of $5.6 million a year earlier is the cleaner sign of progress. The headline profit is real accounting, but it came from the investment book, not the lab.

The balance sheet changed in Q2 as well. Tempus issued a $460 million convertible note due 2032 with a 0% coupon, and it repaid its term loan and revolver in the first half. Borrowing at zero interest for six years tells you the bond market now treats Tempus as a credit worth lending to, even though the conversion feature is the price of that.

Where it sits on our curve

We place the oncology testing franchise in Flight, and we judge that on the product’s own evidence rather than on genomics as a sector. Flight means the product has crossed the buyer’s threshold, it is being bought at scale and paid for, and the growth now comes from repeat orders and expansion rather than proof that it works. Oncology volume grew 31% year over year in Q2, faster than the 28% in Q1, on a base that is already large. That acceleration on a mature product is what Flight looks like.

There are two buyers with two thresholds. For the ordering oncologist, the threshold is a test that the payer covers and that fits the clinic’s workflow, and the tumour only approval removes the second sample that community practices often failed to collect. For the payer, the number is set: $4,500 under the ADLT rate for xT CDx. Management says that if xF also wins approval, blended oncology pricing rises from about $1,850 toward about $2,600. That last step needs an FDA approval that has not been granted.

On our evidence sequence of demonstration, paid deployment, renewal, expansion and cash, Tempus has reached expansion. The commitments behind it are the new multiyear licences from BioNTech, Daiichi Sankyo and Incyte, added to existing deals with AstraZeneca, GSK, BMS and Merck. Tempus also delivered its first foundation model to AstraZeneca. In the Q2 letter, management said the “vast majority” of 2026 growth is already under contract. Cash is the step not yet reached. Adjusted EBITDA has turned positive, but operating cash flow for the half was still negative.

Is the change repeatable or a temporary surge? The pricing uplift is structural, because CMS sets the rate and it does not get renegotiated customer by customer. Data bookings are a different matter. Roughly $200 million in one quarter is lumpy by nature, and we treat it as a surge until a second quarter confirms it.

The numbers and what the street expects

Q2 2026 revenue was $382.5 million, up 22% year over year, according to the July 30 earnings release filed with the SEC. That is slower than Q1’s 36%, and the slowdown is what the bears point to. Most of it is arithmetic. Q1 still included a period before Ambry was in the prior year comparison, and hereditary testing grew only 5% in Q2 because it was lapping an unusually strong Q2 2025.

The detail under the total tells a different story. Oncology volume grew 31%, up from 28% in Q1. MRD tests rose from 6,500 to 9,000 in one quarter, a 38% sequential jump, and management said only about a tenth of the salesforce sells MRD today. Insights revenue grew 36%. Gross profit grew 26%, faster than revenue, so each dollar of sales is getting more profitable.

For 2025 as a whole, revenue was $1.272 billion, up 83% including Ambry and about 30% organically. The organic number is the one to anchor on. It matches the oncology volume trend closely, which makes us more confident it is the real underlying growth rate.

On the balance sheet at June 30, Tempus held $820.7 million of cash and marketable securities, $599.6 million of it in cash. Against that sit $1.17 billion of convertible senior notes and a $188 million convertible promissory note. Operating cash burn was $80.8 million for the first half, most of it in Q1. Diluted weighted shares were 182.4 million against 173.4 million a year earlier, up 5.2%. That is the real dilution cost of growing this fast, and it comes before any Personalis shares.

Guidance was raised on July 30 to 2026 revenue of $1.595 billion to $1.605 billion, about 25% growth, with adjusted EBITDA of about $65 million. That guide excludes Personalis. For Q3, the company guided to $405 million to $410 million in revenue.

Consensus for Q3, as compiled by MarketBeat, is about $409.4 million in revenue and a loss of about 7.5 cents per share. That puts the street at the top of the guided range, which leaves little room for a soft quarter. Full year consensus is a loss of about $1.38 per share for 2026 and about $1.03 for 2027. We could not confirm the analyst count or the spread between the highest and lowest estimates, and we found no published 2027 revenue consensus.

Over the past ninety days, revisions have moved up, led by the guidance raise itself. Our own 2027 revenue estimate, excluding Personalis, is about $2.0 billion. That is roughly 22% organic growth plus the $85 million pricing uplift. If you take one number from this section, take the oncology volume growth rate, because almost everything else follows from it.

Why it wins

Follow the money. Medicare and commercial insurers pay for each test at a blended price near $1,850 today. Drug makers pay multiyear licence fees for data and models. Each extra test brings real costs: sequencing reagents, lab labour, sample logistics and billing. Each extra data licence costs very little beyond fees owed to the health systems that share records.

The ADLT uplift lands almost entirely in gross profit. A test that costs the same to run and pays $200 more drops nearly all of that $200 to the bottom line. Tempus keeps the benefit because CMS sets the rate. No hospital purchasing department can negotiate it away, and competitors cannot discount it unless they win the same approvals.

That approval position is the moat we trust most. Tempus is the first lab with FDA companion diagnostic approval for both tumour only and tumour normal profiling across 648 genes. A rival would need its own approvals, and the FDA does not rush those.

The harder question is whether a much better, much cheaper base model helps or hurts. We think it cuts both ways, and unevenly. Cheap, capable models make Tempus’s 2.5 million deidentified, outcome linked records more valuable, because a stronger model can get more out of good data, and nobody can download those records. They also weaken any edge Tempus has in its own models and apps, since a drug company with the raw data and an open model may not need Tempus’s software. Our view is that the data is the business and the models are packaging around it.

Personalis adds one more piece. Tempus already distributes NeXT Personal, and buying $PSNL for about $1.5 billion of enterprise value gives it full ownership of the assay that sits at the centre of its recurring revenue plan.

What could go wrong

The strongest alternative explanation is that the growth is lumpier and less structural than it looks. Ambry flattered 2025. Total growth fell from 36% to 22% in a single quarter. Data revenue depends on a handful of very large pharma contracts, AstraZeneca and Pathos especially, and we could not find clear end dates for them. Read that way, the $200 million bookings quarter was one customer acting in one quarter, and the oncology acceleration is a temporary bump from the pricing change getting ordering habits moving.

We take that seriously. If a large AstraZeneca or Pathos contract ended, or bookings stalled while diagnostics growth also slowed, the high margin data revenue that helps justify nearly eight times sales would disappear. The lab alone, at hereditary growth rates, would not support this price. That is the bear case at full strength, and it does not depend on anything going badly wrong. It only needs the data line to stop growing.

The test that separates the two readings is Q3. If oncology volume grows at least 28% and diagnostics revenue growth rises toward it, and bookings come from customers other than AstraZeneca, our reading holds. If volume slows and bookings come mostly from one customer, the alternative wins.

Then there is the $400 million of 2028 uplift. That figure depends on an FDA approval for xF that has not happened. If the approval slips, the 2028 revenue line in our central case shrinks.

Financing is the risk people underrate. At the first half burn of about $40 million a quarter, $820.7 million covers roughly 20 quarters, and adjusted EBITDA has been positive since Q2. That is comfortable for reaching November 2 and the 2027 pricing step. Personalis changes the arithmetic. The deal terms described in the merger filing allow up to half the price in cash, which could mean around $750 million funded from cash and the credit facility. The stock portion could add roughly 6% to 7% more shares at today’s price. At the maximum cash mix, the cushion that looks thick today becomes thin quickly.

A strong operating thesis on a balance sheet that is about to get tighter. We would rather say that plainly than leave it out.

The investment thesis

Tempus is a cancer lab whose price per test has just been raised by regulators and whose monitoring test is compounding, and we think today’s price is justified if oncology volume holds near 30% and the 2027 pricing uplift arrives on time.

The chain runs in four links. The constraint is easing: the May 29 approval moved the solid tumour portfolio onto unified ADLT pricing. Adoption is responding, with oncology volume accelerating to 31% and MRD up 38% in one quarter. Tempus keeps the value because CMS sets the rate and the extra revenue arrives with almost no extra cost. The weakest link is financing, where 5.2% annual dilution and a possible Personalis share issue eat into what each existing share receives.

Now reverse the price. A market value of about $12.2 billion, plus $1.36 billion of convertible debt, minus $0.82 billion of cash, gives an enterprise value near $12.7 billion. That is 7.9 times guided 2026 revenue. If a maturing diagnostics and data business settles at five times sales, the price needs about $2.55 billion of revenue by 2028, or roughly 26% compound growth.

In our downside case, growth slows to 15% and xF approval is delayed, which gives 2028 revenue of about $2.1 billion, worth about $10.5 billion at five times, roughly 17% below today’s enterprise value. In our central case, 2027 revenue is about $2.0 billion and 2028 reaches about $2.55 billion with the xT and xF pricing uplift. That supports today’s price, with very little to spare. The upside case, with 30% growth plus Personalis MRD, reaches $2.8 billion or more in 2028 and an enterprise value between $14 billion and $17 billion.

The gap between the evidence and the price is smaller than the recent selloff suggests. We think the market is pricing Tempus as a data story with a lab attached, and it is wrong about which part is reliable. The pricing uplift is the most certain revenue on the page, and it is close to pure margin.

We hold while the thesis holds and sell when it breaks. The next test is Q3 results on November 2, 2026. Written now, the thesis strengthens on Q3 revenue of at least $410 million, which is 22.7% above last year’s $334.2 million, oncology volume growth of at least 28%, at least 11,000 MRD tests, Data and Apps bookings of at least $100 million from more than one pharma customer, adjusted EBITDA of at least $15 million, and the 2026 guide held at or above $1.60 billion with the $85 million 2027 uplift reaffirmed. It breaks on revenue below $400 million, a 2026 guide below $1.595 billion, oncology volume growth below 22%, MRD tests below 9,000, bookings below $75 million, an adjusted EBITDA guide below $50 million, or the pricing uplift pushed past 2027.

What we are watching

• November 2, 2026, Q3 results before the market opens. Strengthens with revenue of at least $410 million, oncology volume up at least 28%, at least 11,000 MRD tests and bookings of at least $100 million from more than one customer. Breaks with revenue below $400 million or oncology growth below 22%.
• November 2, 2026, the guidance update on the same call. Strengthens if the 2026 revenue guide is held or raised to at least $1.60 billion and the $85 million 2027 ADLT uplift is reaffirmed. Breaks if the guide falls below $1.595 billion or adjusted EBITDA guidance falls below $50 million.
• Late Q4 2026 or early 2027, the expected Personalis close. Strengthens if the final mix uses less than the maximum cash and the share issue stays near 6%. A weaker result would be a cash heavy close that cuts the cushion below about a year of first half burn plus the deal cost.
• Any FDA decision on xF, with no date yet given. Approval would bring the roughly $400 million 2028 uplift within reach. Continued silence into 2027 would push our central case toward the downside.
• Q4 2026 results and the first 2027 guidance, expected early in 2027. Strengthens if 2027 guidance implies at least 22% organic growth before Personalis. Breaks if the ADLT uplift is missing from the guide.

The bottom line

The shift in cancer diagnostics is from a single profile to repeated monitoring, paid for at rates that finally cover the work. Tempus has the approvals, the volume and the data to keep a large share of that, and the May approval handed it a price rise the market has not fully absorbed.

At $12.7 billion of enterprise value, though, the price already assumes that much of this works. Our central case supports it, and only just.

What would change our mind is specific. If oncology volume growth falls below 22% on November 2, or if Q3 data bookings turn out to come almost entirely from AstraZeneca, we would accept that the bears had the better explanation all along.

References

1. Tempus AI, Inc. Tempus Reports Second Quarter 2026 Results. Exhibit 99.1 to current report. US Securities and Exchange Commission. Published July 30, 2026. Accessed October 25, 2026. https://www.sec.gov/Archives/edgar/data/0001717115/000119312526326083/tem-ex99_1.htm
2. Tempus AI, Inc. Q2 2026 A word from our CEO. Tempus AI Investor Relations. Published July 30, 2026. Accessed October 25, 2026. https://investors.tempus.com/static-files/bdc8fe04-0bdd-48b8-b884-ea65d548204d
3. Tempus AI, Inc. Q2’26 Financial Results presentation. Tempus AI Investor Relations. Published July 30, 2026. Accessed October 25, 2026. https://investors.tempus.com/static-files/feb11be8-6e62-4cdd-88c8-2cfdbb2156c3
4. Personalis, Inc. Form 425, transaction overview. US Securities and Exchange Commission. Published July 2026. Accessed October 25, 2026. https://www.sec.gov/Archives/edgar/data/0001527753/000119312526309090/d81168d425.htm
5. OncoDaily. Tempus AI to Acquire Personalis in $1.5 Billion Deal. OncoDaily. Published July 27, 2026. Accessed October 25, 2026. https://oncodaily.com/technology/personalis-548302
6. MarketBeat. Tempus AI (NASDAQ:TEM) Stock Price Down 5.7%: Here’s What Happened. MarketBeat. Published October 16, 2026. Accessed October 25, 2026. https://www.marketbeat.com/instant-alerts/price-tempus-ai-nasdaq-tem-stock-price-down-57-heres-what-happ
7. MarketBeat. Tempus AI (TEM) Q3 2026 Earnings Date, Time and Estimates. MarketBeat. Published October 21, 2026. Accessed October 25, 2026. https://www.marketbeat.com/earnings/reports/2026-11-2-tempus-ai-inc-stock/
8. GuruFocus. Tempus AI Inc (TEM) Q2 2026 Earnings Call Highlights. GuruFocus. Published July 31, 2026. Accessed October 25, 2026. https://www.gurufocus.com/news/8993604/tempus-ai-inc-tem-q2-2026-earnings-call-highlights-revenue-surges-22-to-3825m-raises-fullyear-guidance?mobile=true
9. Yahoo Finance. Should You Be Worried About Tempus AI Stock? Yahoo Finance. Published October 20, 2026. Accessed October 25, 2026. https://finance.yahoo.com/markets/stocks/articles/worried-tempus-ai-stock-110300751.html
10. Business Wire via Stockhouse. Tempus to Report Third Quarter 2026 Financial Results on November 2. Stockhouse. Published October 15, 2026. Accessed October 25, 2026. https://stockhouse.com/news/press-releases/2026/10/15/tempus-to-report-third-quarter-2026-financial-results-on-november-2

Not financial advice.