A $1.5 billion deal was announced, and the buyer's stock fell 8%

On Monday morning, July 20, 2026, Tempus AI (NASDAQ: TEM) said it had signed a merger agreement to acquire cancer genomics company Personalis (NASDAQ: PSNL) at $16.25 per share. On an enterprise value basis, excluding the stake Tempus already owned, that works out to roughly $1.5 billion. Later that same session, Tempus traded down 8.06% to $48.24. Personalis, the company being bought, fell too — some outlets reported a drop of around 12%.

Normally an acquisition announcement sends the target up by roughly the premium and knocks the acquirer around a little. This time both went down. The reason is the deal structure. This is a 100% stock transaction, and the number of Tempus shares Personalis holders receive is set by a floating exchange ratio that only locks in at closing. Translation: if Tempus stock falls, the value in Personalis holders' hands falls with it. The moment Tempus dropped 8%, Personalis shares stopped being a "$16.25 cash ticket" and became a derivative on Tempus stock.

The premium itself was thin. $16.25 is about 5.6% above the prior trading day's close (Friday, July 17) — the official release rounds it to 6%. Typical M&A premiums land somewhere between 20% and 40%, so this is low. The company's counterargument: measured against the unaffected 30-day volume-weighted average price, before rumors moved the stock, the premium is 28%. That's a fight over where you draw the baseline. Pick one benchmark and this is a fair deal; pick the other and it's a bargain-bin price.

Here's the deal, though: the real story isn't the numbers, it's the technology. What Tempus is spending $1.5 billion on is less a company than a single test called NeXT Personal. It takes a few milliliters of blood from a patient who has finished surgery or chemotherapy and hunts for DNA fragments shed by cancer cells still hiding somewhere in the body. Why that's a $1.5 billion asset, and why this market is suddenly this hot, is where we should start.

A Chicago AI company and a Stanford genomics spinout

Tempus AI was founded in 2015 in Chicago by Eric Lefkofsky, best known as a Groupon co-founder. The origin story is personal. When his wife was diagnosed with breast cancer, he found that tumor data lived in silos at every hospital and that clinicians had almost no structured data to work from. That's the company's official telling. Today Tempus runs tumor tissue and blood genomic sequencing (xT, xF and others), AI models trained on that data, and a data-and-applications business that sells the resulting datasets to pharmaceutical companies.

The numbers say it has grown into a real business. First-quarter 2026 revenue came in at $348.12 million, up 36.1% year over year, split between $261.10 million in diagnostics and $87.02 million in data and applications. The company raised full-year 2026 revenue guidance to $1.59-1.60 billion and is targeting roughly $65 million in adjusted EBITDA. It's not a big profitable company yet, but it's in the phase where losses shrink while revenue grows in the 30s. Cash and equivalents stood at $521.17 million as of March 31.

Personalis is a different animal. It was incorporated in Delaware in February 2011 as a Stanford spinout, with next-generation sequencing researchers on the founding roster: John West (first CEO), Euan Ashley, Atul Butte, Russ Altman, and Michael Snyder. It went public on Nasdaq in June 2019, raising $139.80 million net of fees, and for years ran a heavily B2B business supplying tumor profiling and immune profiling services (ImmunoID NeXT, among others) to pharma companies, health systems, and research institutions.

These two are not strangers. In November 2023, Tempus made an equity investment in Personalis and simultaneously signed a commercial partnership to distribute NeXT Personal. The arrangement had Tempus's sales organization selling Personalis's MRD test to U.S. oncologists. Per reporting, that investment left Tempus holding roughly 12.25% of Personalis, and the $1.5 billion figure excludes that stake. Some sources put the total equity value at $1.9 billion, so depending on the yardstick you'll see $1.5B to $1.9B.

Lefkofsky told MedTech Dive that he considered buying Personalis back in 2023 but chose the commercial agreement instead, waiting for the MRD test to mature. His reason for buying now is a market shift: volume has moved "pretty dramatically" toward tumor-informed testing, he said. In other words, Tempus spent two and a half years selling somebody else's product, confirmed the demand firsthand, and then bought the whole company.

What MRD actually is — and what's inside the deal

MRD stands for minimal (or molecular) residual disease. Picture a patient whose tumor has been surgically removed and who has finished chemotherapy, and whose CT or MRI scans now show nothing at all. The catch is that imaging has a physical floor on the lesion size it can detect. Invisible doesn't mean zero cancer cells, and those leftover cells are exactly what come back as a recurrence months or years later.

MRD tests look for them in blood. As cancer cells die or divide they shed fragments of their DNA into the bloodstream — circulating tumor DNA, or ctDNA. The problem is how little of it there is: tumor-derived fragments can be one in tens of thousands of the total cell-free DNA. Hence the tumor-informed approach. You sequence the patient's tumor tissue first to build a mutation fingerprint unique to that person, then search the blood only for that fingerprint. It's not fishing in the open ocean; it's fishing with a photo of the specific fish. That's what NeXT Personal does. Tempus's in-house xM MRD runs the tumor-naive route, using blood alone with no tissue required — so the two don't overlap, they complement.

Why this makes money: the result changes treatment decisions. MRD-positive means high recurrence risk, which supports continuing or escalating adjuvant therapy. MRD-negative supports scaling back chemotherapy the patient doesn't need. Tests that change treatment decisions get reimbursed, and reimbursement is when volume explodes. NeXT Personal already has Medicare coverage across three indications — breast cancer, lung cancer, and immunotherapy monitoring — with more expansion expected. Tempus and Personalis size the MRD opportunity at $20 billion.

The volume is showing up. Personalis's preliminary second-quarter 2026 revenue was $22.4 million on 10,384 clinical tests, up 33% from the prior quarter. Personalis test volume flowing through the Tempus channel alone jumped from roughly 5,500 in Q1 to roughly 9,000 in Q2. That's evidence the partnership was working — and also the exact point where Tempus starts asking why it's splitting margin with someone else on revenue its own sales force built.

Here are the deal terms, per the Form 425 filed with the SEC and the official release.

Item Detail
Announced July 20, 2026 (approved by both boards)
Size ~$1.5 billion (enterprise value, excluding Tempus's existing stake)
Per-share consideration $16.25
Premium ~5.6% to prior close (6% per the release); 28% to unaffected 30-day VWAP
Form of consideration 100% stock; Tempus may elect to pay up to 50% in cash
Exchange ratio Floating, capped at 0.3356 shares; fixed at closing
Cash source Cash on hand plus draws on existing credit facility ($521.17M cash as of March 31)
Walk-away trigger Personalis may terminate if Tempus trades below $46.00
Break fee ~$76.8 million, symmetric both ways
Outside date April 20, 2027
Expected close Late 2026 to early 2027
Conditions Personalis shareholder approval, regulatory approval, customary closing conditions

The number that matters most is 0.3356. Divide $16.25 by 0.3356 and you get about $48.40. Above roughly $48.40, Personalis holders receive exactly $16.25 of value. Below it, the ratio stops at 0.3356 and what they receive is worth less than $16.25. Now look again at that $48.24 close on announcement day — the stock was already trading right at the threshold. That's why Personalis shares sat well below the headline deal price. And the $46.00 trigger is the emergency exit: go under it and the Personalis board can blow up the table.

What each side gets — and the risk baked into the structure

Tempus's side is straightforward: vertical integration. It has been the distribution partner selling somebody else's test, and now it owns the test. The full margin comes in-house, and more importantly the data comes in-house. The core of the Tempus business model isn't diagnostics for their own sake — it's selling the multimodal data that diagnostics generate to pharma. When you can observe the entire patient journey inside one platform, from diagnosis to therapy selection to recurrence monitoring, the quality of what you can sell for biomarker discovery and clinical trial enrollment changes materially. Covering the full MRD spectrum with tumor-informed (Personalis) plus tumor-naive (Tempus xM) is a real second benefit.

Personalis shareholders have a much more awkward position. CEO Chris Hall said that after a thorough process, he's confident Tempus's proposal delivers the greatest value to shareholders and the fastest path for patients. "Thorough process" is also a signal that other bidders were actually approached. Still, a 5.6% premium, all stock, plus a floating exchange ratio is a package that invites complaints. Getting stock instead of cash means carrying Tempus share-price risk for the six to nine months until closing. The up-to-50% cash option belongs to Tempus, not to Personalis holders — and it can only be exercised to the extent the deal still qualifies as a tax-free reorganization.

Pharma customers mostly win. Until now they ordered tumor profiling from vendor A and MRD tracking from vendor B, then stitched the data together themselves. Integrated, they can run trial design through enrollment and monitoring with one vendor. The flip side is that one fewer vendor means less negotiating leverage, which over time can mean price pressure in the wrong direction. For patients, the practical change is that reimbursed MRD testing likely spreads wider and faster. But it's too early to say more access equals better survival. Whether changing treatment based on an MRD-positive result actually improves outcomes is still being proven in large trials, cancer type by cancer type.

The risks stack up like this. First, all-stock plus a floating ratio ties the deal's fate to the acquirer's share price: below $46 Personalis can walk, and anywhere under roughly $48.40 Personalis holders take a haircut. Second, if Tempus wants to max out the cash election it has to lean on cash and its credit facility — adding debt is a real burden for a company still at about $65 million of adjusted EBITDA. Third, closing is at least six months out, and rivals won't sit still during that window. Fourth, dilution: paying $1.5 billion in stock dilutes existing Tempus holders by exactly that much, and the cleanest read of the 8% drop is the market pricing dilution and integration risk immediately.

Precedents — What Worked and What Didn't

The success story people cite most is Exact Sciences buying Genomic Health. Announced July 29, 2019, roughly $2.8 billion in cash and stock, closed that November 8. Exact was heavily concentrated in one product, the Cologuard colorectal screening test. Bolting on Genomic Health's Oncotype DX breast cancer prognostic test turned a screening company into a full-spectrum cancer diagnostics company. How well the integration worked is settled by the ending: on March 23, 2026, Abbott closed its acquisition of Exact Sciences at $105 per share in cash, $21 billion total. Enterprise value multiplied several times over in seven years.

The other useful reference in the same direction is Roche's acquisition of Foundation Medicine. Roche first took roughly 57% of Foundation via a 2015 tender offer, then on June 19, 2018 bought the remaining shares at $137 apiece, about $2.4 billion, making it a wholly owned subsidiary. Take a stake, collaborate, validate, then buy the rest — that sequence is a near-perfect match for Tempus-Personalis. Tempus started with the November 2023 equity investment plus commercial partnership and came back two and a half years later for the rest. The advantage of this pattern is that you validate the target's real demand and operational quality with your own sales force before you write the big check.

The failure case is the textbook: Illumina and GRAIL. Illumina, the dominant player in sequencing instruments, tried to buy multi-cancer early detection startup GRAIL and closed the deal before the European Commission finished its review. The EC issued a record €432 million fine and ordered divestiture. Illumina completed the separation on June 24, 2024, distributing 85.5% of GRAIL to its own shareholders. The irony: on September 3, 2024, barely two months later, the European Court of Justice overturned the EC's jurisdictional theory outright, voiding the fine and the divestiture order. Illumina got the ruling that the order was wrong after it had already broken the company apart. Billions of dollars and several years for nothing.

Three lessons come out of that set. Regulatory risk is not solved by "we'll win on appeal later" (Illumina). Combining complementary test portfolios in adjacent markets does create real value (Exact-Genomic Health). And a minority stake followed by full acquisition meaningfully lowers execution risk (Roche-Foundation). Tempus-Personalis doesn't carry an Illumina-style vertical monopoly problem, and structurally it's following the Roche playbook. The one decisive difference from both success cases: it's paying in its own stock, not cash.

How Rivals Counter

Natera is the clear leader here. Signatera is the de facto standard for tumor-informed MRD, appearing in more than 170 peer-reviewed papers with over 300,000 patients tested cumulatively. In 2025 Natera launched Signatera Genome broadly in the U.S. with substantially higher sensitivity, and it holds reimbursement across multiple solid tumors. From Natera's seat, this deal reads less like a threat than validation: somebody just paid $1.5 billion to enter the category it built. That said, if Tempus bundles the test with its own sales network and data platform, the fight for new accounts gets a lot rougher.

Guardant Health attacks from a different angle. Guardant Reveal is a tumor-agnostic, tissue-free ctDNA MRD test, which wins when tissue is hard to get or results are needed fast. Add Shield, its colorectal screening test, which secured UnitedHealth coverage, and Guardant has its own screening-to-diagnosis-to-monitoring axis. When Lefkofsky says the market has moved toward tumor-informed testing, that's partly a jab at the Guardant approach — though in the clinic the two methods are increasingly used side by side depending on the situation.

Exact Sciences' Oncodetect is a completely different proposition now. When Abbott bought Exact outright for $21 billion in March 2026, Oncodetect stopped being a startup's product and became one line in a global healthcare conglomerate's portfolio. Abbott has framed the deal around leadership in a cancer screening and precision oncology diagnostics market it sizes at $60 billion. That's an opponent with entirely different capital, hospital sales reach, and global regulatory muscle. It's fair to read the Abbott-Exact combination as part of why Tempus is in a hurry to bring MRD in-house right now.

Foundation Medicine continues to serve as the comprehensive genomic profiling (CGP) standard under Roche and has pursued MRD through both in-house development and partnerships. Meanwhile new early-detection entrants like Freenome keep publishing colorectal screening data. The whole "see cancer from a tube of blood" market is heating up at once. What's actually being fought over isn't the narrow MRD category — it's who assembles the entire cancer diagnostics value chain, from screening to diagnosis to therapy selection to monitoring, under one roof. Natera is betting on MRD depth, Guardant on liquid biopsy breadth, Abbott on capital and distribution, Tempus on AI and data.

So What Actually Changes

If you're a developer or data engineer, this is a clean case study in how the data land grab plays out in healthcare AI. Tempus didn't spend $1.5 billion to buy a model. It spent it to buy the longitudinal data that trains models — not one measurement at diagnosis, but repeat measurements at three months, six months, and a year post-treatment. This deal priced the thesis that the moat in healthcare AI is ownership of repeat-measurement data, not the algorithm. Whatever your domain, owning the pipeline that generates time-series data is what ends up separating valuations.

If you're on the enterprise side, particularly managing diagnostics vendors in pharma or biotech, go check your contract renewal dates now. Closing is expected late 2026 to early 2027, but the outside date is April 20, 2027, and until then the two companies operate as legally separate entities. Product line rationalization and pricing changes are both plausible post-integration, so if NeXT Personal is written into an ongoing trial as an endpoint, review your vendor continuity clauses.

If you're an investor, there are exactly two numbers to watch: Tempus at about $48.40 and Tempus at $46.00. The first is where the 0.3356 exchange ratio cap bites. The second is the trigger that lets Personalis terminate. How far below $16.25 Personalis stock trades is a real-time readout of how the market handicaps both deal completion odds and Tempus share-price risk. On the Tempus side, watch the dilution math, how much of the cash election gets used, and how fast MRD revenue actually shows up after integration. Worth noting: even if Personalis walks, the break fee is symmetric at roughly $76.8 million each way.

If you're a regular person — someone who's been treated for cancer, or has a family member in treatment — nothing changes right now. If you're already getting NeXT Personal or a Tempus test, test delivery and coverage stay as they are until the deal closes. Over the medium term, MRD testing likely picks up reimbursement in more cancer types and becomes a more common part of follow-up care. But "detecting recurrence early from blood" does not automatically become "better outcomes." Whether changing treatment based on early detection actually improves survival is still being established cancer type by cancer type, and the risk that results drive unnecessary anxiety or overtreatment is part of the same conversation.

🥄 Three Things You're Probably Wondering

— So what does this mean for me? Not much directly. But if you're in routine follow-up after cancer surgery or chemotherapy, the odds go up meaningfully that within a few years your doctor offers a blood draw to check for recurrence, either alongside imaging or instead of some of it. Reimbursement outside the U.S. is a separate fight and will lag.

— A 6% premium? Couldn't Personalis shareholders vote this down? It's possible. Shareholder approval is a hard condition here, and a low premium plus an all-stock structure is exactly the setup that draws activists and lawsuits. But the CEO says they ran a thorough process, which may mean no other bidder showed up. Too early to call.

— If Tempus stock keeps falling, does the deal die? Not automatically. Dropping below $46 gives Personalis the right to terminate, not the obligation. In practice, parties more often renegotiate terms or raise the cash portion to patch things up. Which way this goes depends on Tempus's balance sheet at that moment, and that's too early to call.

References

Numbers and criteria are as of announcement and may change. Investment calls are yours to make!