Grail

Sarah Bull:

Alright Marcus, I know you've been itching to tear this one apart, but hear me out first.

For anyone who hasn't done the work on GRAIL — and I know you have, Marcus, despite pretending you haven't built a 47-tab spreadsheet on this — let me explain what we're actually talking about. GRAIL makes a blood test called Galleri that can detect over 50 types of cancer from a single blood draw. You walk into your doctor's office, they take some blood, and two weeks later you find out if you have cancer signals floating around in your body.

The science is actually elegant. When tumors grow, they shed DNA into your bloodstream — we call it cell-free DNA or cfDNA. The problem is, there's not much of it, and it's mixed in with all your normal DNA. What GRAIL figured out is that cancer DNA has distinct methylation patterns — essentially chemical tags that tell you "this came from a tumor" versus "this came from your liver doing normal liver things." They built machine learning models trained on hundreds of thousands of samples to spot these patterns. And critically, they can tell you not just "you have cancer" but "it's probably in your pancreas" — what they call cancer signal origin prediction.

Why does this matter? Because right now we only screen for five cancers — breast, cervical, colorectal, prostate, and lung. The other 70% of cancer deaths come from cancers we don't screen for at all. Pancreatic, ovarian, liver, esophageal — by the time you have symptoms, you're often stage IV and your survival odds are terrible. GRAIL is trying to flip that equation.

Now, to the stock. I've been in healthcare for fifteen years and I can count on one hand the number of times I've seen clinical data this clean drop into a regulatory pathway this favorable. GRAIL is one of those names.

PATHFINDER 2 showed Galleri detected seven times more cancers when added to standard screening. Seven times. And critically, the positive predictive value came in at 61.6% — that's a massive improvement from the 43-50% range in earlier studies. The false positive rate was 0.4%. These are numbers that make the FDA approval conversation much easier.

The setup for 2026 is as good as it gets for a pre-approval biotech. You've got the PMA submission happening this quarter — management specifically pulled it forward from "first half" to Q1. You've got NHS-Galleri data in mid-year from 140,000 patients that could show actual stage-shift at a population level. And you've got commercial momentum accelerating — 35% volume growth in 2025, guidance for 22-32% Galleri revenue growth in 2026.

The balance sheet concern is off the table. They've got $850 million in cash, runway into 2030. Samsung wrote them a $110 million check at $70 when the stock was trading in the 90s. That's not charity — that's a sophisticated strategic investor saying the risk-reward is skewed.

The market is pricing this like a coin flip on FDA approval. I think the probability is significantly higher than that, and the payoff on approval — Medicare coverage, insurance reimbursement, 10x the addressable market — is asymmetric. This is a $4 billion market cap company going after a $20-30 billion TAM. I'm long.

Marcus Bear:

First of all, it's 52 tabs, not 47. And three of them are just memes I've saved for when this thing blows up.

Look, Sarah makes a compelling case, and I want to be clear — I'm not saying Galleri doesn't work. The science is genuinely impressive. Being able to detect pancreatic cancer before symptoms is the kind of thing that could save hundreds of thousands of lives. My mom had a cancer scare last year and I found myself googling whether I could just buy a Galleri test. So I get it. I'm not a monster.

But there's a difference between "the science works" and "this is a good stock to own at $95."

Let me walk through my concerns systematically.

First, timeline. Sarah says PMA submission in Q1. Great. But FDA review for a novel diagnostic of this complexity? You're looking at 12-18 months minimum. That puts approval in mid-to-late 2027, not 2026. So anyone buying today for "the FDA catalyst" is paying up for something that's potentially two years away. And that's assuming no deficiency letters, no additional data requests, no advisory committee drama. The FDA has never approved an MCED test. They're going to take their time.

Second, the NHS data. Everyone's treating this like it's going to be a slam dunk. But let's remember what happened in May 2024 — NHS looked at the first-year data and said it wasn't compelling enough to accelerate implementation. Their words, not mine. They're waiting for the full three-year results. The primary endpoint is reduction in late-stage cancer diagnoses, not detection rates. That's a much harder bar to clear, and the precedent from other screening trials is that first-round screening often doesn't show stage shift because you're catching prevalent cases. If NHS-Galleri disappoints, this stock gets cut in half.

Third — and this is what keeps me up at night — competition. Sarah didn't mention that Abbott just bought Exact Sciences for $21 billion. Exact Sciences has Cancerguard on the market at $689, sixty bucks cheaper than Galleri. They have a massive primary care salesforce. They have Cologuard relationships with every GI doc in America. Now they have Abbott's global distribution. GRAIL is bringing a knife to a gunfight.

Fourth, the cash burn. Yes, they have $850 million. But they're burning $90 million a quarter. That's fine if approval comes in 2027. But if there are delays? If they need to run additional studies? If payers push back on coverage even after FDA approval? Suddenly that runway doesn't look so comfortable.

I'm not short, but I'm definitely not long. The risk-reward at these levels doesn't work for me.

THE FDA PATHWAY

Sarah Bull:

Okay, let me address the regulatory timeline point because I think Marcus is applying standard FDA math to a non-standard situation.

GRAIL has Breakthrough Device Designation. That's not a participation trophy — it means priority review, interactive engagement with FDA throughout the process, and the agency has already signaled they view this as addressing an unmet medical need. GRAIL has been doing modular submissions — they're not dumping 50,000 pages on FDA's desk in March and praying someone reads it before lunch.

And on the NHS point — yes, the first-year data wasn't enough for an accelerated pilot. But that's actually the right outcome! You don't want NHS making a billion-pound commitment on one year of data. What it tells me is that the process is rigorous, which makes positive final results more credible. The trial was designed for three years of screening precisely because stage-shift takes time to manifest. The NLST lung cancer trial showed the same pattern — no stage IV reduction in year one, significant reduction by the end of the study.

The NHS data isn't a risk, it's an asymmetric catalyst. If it's positive, GRAIL becomes the first MCED test with randomized controlled trial evidence of clinical utility. That's the holy grail — pun intended — for payer adoption. It changes the entire reimbursement conversation.

Marcus Bear:

Did you really just say "holy grail"? I'm docking points for that.

Breakthrough Designation is a marketing term. It gets you more FDA meetings, it doesn't guarantee approval or timeline. Theranos had FDA engagement too. I'm not comparing them scientifically — Galleri actually works — but I am saying that "the FDA likes us" is not an investment thesis.

On the modular submission — fair point. But here's what worries me: GRAIL has never taken a product through FDA approval. They spun out of Illumina, they've been selling Galleri as a lab-developed test under CLIA. The PMA process is a different beast entirely. They're going to face questions about manufacturing consistency, about the AI/ML components of their algorithm, about how they handle indeterminate results, about the software validation. These are solvable problems, but they take time.

And Sarah's point about NHS actually reinforces my concern. She's saying "the trial was designed for three years, so we need three years of data." Okay, but management and sell-side have been hyping this as a 2026 catalyst. If the data comes in June and shows directional improvement but doesn't hit statistical significance on the primary endpoint, what happens? I'll tell you what happens — the stock tanks 40% while everyone argues about whether the data was "good enough" and whether a trend is the same as hitting the endpoint. That's not a risk-reward I want to own.

COMPETITION

Sarah Bull:

This is where Marcus and I fundamentally disagree. He sees Abbott buying Exact Sciences as a threat. I see it as validation.

Abbott paid $21 billion for a company whose MCED test, Cancerguard, is clearly inferior to Galleri. Let me be specific: Cancerguard detects fewer cancer types with lower sensitivity. Their PPV in clinical studies was around 28% versus Galleri's 62%. That's not a rounding error — that's the difference between "more likely than not you have cancer" and "probably a false alarm." If you're a patient getting that call from your doctor, which test do you want?

Here's the thing about diagnostics: physicians adopt based on clinical evidence, not price. If you're a primary care doc and you're going to tell a patient they might have cancer based on a blood test, you want the test with the best data. That's Galleri. Full stop. PATHFINDER 2 is the largest interventional MCED study ever conducted in an intended-use population. Exact Sciences is still recruiting for their equivalent study.

And strategically — if Abbott thought they could build or buy a better MCED test, they would have. They bought Exact Sciences because they needed to be in this market and GRAIL wasn't for sale. That's a signal about how valuable GRAIL's position is.

Marcus Bear:

Sarah's making my point for me. She says "physicians adopt based on clinical evidence, not price." Great — so how is GRAIL doing commercially? They sold 185,000 tests in 2025. Sounds impressive until you realize there are 100 million Americans over 50. They've penetrated 0.2% of the addressable market in five years. Why? Because no one's paying for it except tech executives and their anxious spouses in Palo Alto.

I'm exaggerating, but only slightly. The average American isn't dropping $950 on a blood test their insurance doesn't cover, no matter how good the data is. Commercial traction without reimbursement is like having a Ferrari with no gas — looks great in the garage, doesn't get you anywhere.

The Cancerguard PPV comparison is also a bit misleading. Exact Sciences uses a different methodology — multi-omic biomarkers versus methylation-only. Different studies, different populations, not apples-to-apples. And their 68% sensitivity for the deadliest cancers is competitive.

But here's the real point: competition isn't just about who has the better test today. It's about who can commercialize faster when reimbursement unlocks. Exact Sciences has 4,000 sales reps calling on doctors every day for Cologuard. GRAIL has a fraction of that. When Medicare coverage happens — and it will happen eventually for someone — Exact Sciences can flip a switch and offer Cancerguard to every Cologuard patient. GRAIL has to build that infrastructure from scratch while burning $400 million a year.

First-mover advantage matters, but so does execution. I'd rather own the company with the distribution infrastructure than the company with the slightly better clinical data.

Sarah Bull:

"Slightly better clinical data" — I'm writing that down. When this thing doubles, I'm sending you a framed quote.

VALUATION AND RISK-REWARD

Sarah Bull:

Let's do the math because this is where I get genuinely excited.

If Galleri gets FDA approval and Medicare coverage, the addressable market explodes. There are roughly 80 million Americans over 50 who should be screened annually for cancer. Even at a conservative $500 reimbursement rate and 10% penetration over five years, you're looking at $4 billion in annual revenue. Apply a 5x revenue multiple — conservative for a high-growth diagnostic with recurring revenue characteristics — and you get a $20 billion market cap. That's 5x from here.

Now, I'm not saying that's the base case. But what's the downside? Let's say FDA drags, NHS disappoints, and they burn through cash faster than expected. The floor is probably $30-40 — the value of the clinical data, the IP, the platform, the 140,000 patient NHS dataset. That's 60% downside against 400% upside. Even if you haircut the probability of the bull case to 30%, the expected value math works.

And here's what the market is missing: GRAIL doesn't need to be a standalone company forever. At $4 billion, they're an M&A target. Every big pharma and diagnostic company is looking at MCED. If the clinical data holds up and the regulatory pathway clears, someone writes a check. J&J, Roche, Danaher — take your pick. The Samsung investment isn't just capital, it's a potential acquirer kicking the tires.

Marcus Bear:

Sarah's upside case requires a lot of things to go right. FDA approval, Medicare coverage, physician adoption, patient compliance, manufacturing scale-up, and no competitive displacement — all while burning $400 million a year and having never launched an FDA-approved product. That's a lot of execution risk at a $4 billion valuation.

Let me offer an alternative framework. What's GRAIL worth today, right now, as a going concern without assuming any catalysts? They'll do maybe $180 million in revenue in 2026 at 25% growth. They're deeply unprofitable. They're selling a $950 test that no insurance covers to a tiny fraction of the addressable market. On comps, without FDA approval, this is maybe a $1.5-2 billion business being generous.

So the market is pricing in $2 billion of option value on FDA approval and downstream catalysts. That might be right. But it means you're paying up for optionality that may not materialize for two-plus years. Time decay matters in options, and it matters here. Every quarter they burn $90 million is another quarter the cash runway shrinks and the pressure to raise capital or find a partner increases.

I'd rather wait. Let me see the PMA submission and FDA's initial response. Let me see the NHS data. If all three come in positive, I'll happily pay $150 for this stock. But paying $95 today to own the uncertainty? That's not my game.

Sarah Bull:

You realize that if you wait for all three catalysts to clear, you'll be buying at $200 and telling yourself you were "disciplined," right?

Marcus Bear:

I can live with missing the first double if it means avoiding the 50% drawdowns. My LPs appreciate my hair remaining attached to my head.

CLOSING

Marcus Bear:

Look, I want to be clear — I respect the science here. Multi-cancer early detection could be one of the most important medical advances of the decade. If this works at scale, we're talking about catching pancreatic cancer before it kills you, finding ovarian cancer before stage IV, detecting liver cancer when it's still operable. That's genuinely life-changing. GRAIL has pioneering technology and strong clinical data.

But pioneering companies don't always make pioneering investments. Plenty of great science has destroyed shareholder capital. Plenty of first movers got run over by better-capitalized fast followers.

The setup for 2026 is binary. Two major catalysts — FDA submission response and NHS data — will determine whether this stock doubles or halves. The market is pricing it like a coin flip. Maybe that's right, maybe it's not. But when I'm allocating risk capital, I want to see the cards before I bet. This is a show-me story, and I'm waiting to be shown.

Also, I'm still mad about Theranos. Every time I see a blood test promising to change medicine, I get twitchy.

Sarah Bull:

Marcus brings up Theranos like I'm asking him to invest in a company that faked all their data. Elizabeth Holmes couldn't detect anything with her machines. GRAIL has 175,000 tests completed, peer-reviewed publications in The Lancet, and data from trials involving 200,000 patients across three continents. Slightly different situations.

But fine — Marcus wants to wait for certainty. The problem is, certainty in biotech doesn't come cheap. By the time the FDA response is positive and the NHS data is out, this stock is at $200 and he's telling his PM meeting it's "too crowded."

Here's my closing thought: GRAIL is one of maybe three companies in the world that could transform how we screen for cancer. They have the best clinical data, the most advanced regulatory pathway, and now the balance sheet to see it through. Yes, there's execution risk. Yes, there's regulatory risk. Yes, I might be wrong and Marcus will send me the framed drawdown chart. But at $95, you're being compensated for those risks with asymmetric upside.

I've seen this setup before. Exact Sciences at $15 before Cologuard approval. Guardant at $30 before their growth inflection. The time to buy is when the outcome is uncertain and the potential is enormous. That's GRAIL today.

I'm long, I'm adding on weakness, and I've already picked out the "I told you so" meme I'm sending Marcus when this works.

Marcus Bear:

For the record, my meme folder is better.