Revolution Medicines

IThe business

Revolution Medicines makes cancer drugs that act on RAS, a family of proteins that tell a cell to grow. When RAS is mutated it stays switched on. The company counts more than 190,000 new RAS-mutant cancers a year in the US: about 56,000 pancreatic (more than 90% of that cancer), 60,000 lung (30%) and 75,000 colorectal (50%).⁵ Revolution’s drugs bind the active form, RAS(ON), by gluing it to a second protein inside the cell, a design the company calls a tri-complex inhibitor.⁵

One drug is approved. On 26 August 2026 the FDA approved RASONQUE (daraxonrasib) for adults with metastatic pancreatic cancer who have had at least one earlier treatment, or who cannot take a multi-drug chemotherapy.⁶ It is a once-daily tablet. The list price is $39,800 for 30 days, which is about $480,000 a year for a patient who stays on it all year C (calculated).⁶ Most will not: pancreatic cancer is lethal, and in the trial half the patients had progressed or died by 7.2 months.⁷

The trial that made the drug

RASolute 302 compared daraxonrasib with the chemotherapy the doctor would otherwise choose, in 500 patients whose cancer had already been treated once. Median survival was 13.2 months against 6.7, a hazard ratio of 0.40. Median time without progression was 7.2 months against 3.6.⁷ Side effects led 1.2% of patients on the drug to stop, against 11.2% on chemotherapy, but the drug is not mild: the label warns of skin and soft-tissue toxicity, mouth sores, diarrhea, bowel perforation and lung inflammation, and 30% of patients had a serious adverse reaction.⁷² The trial was open label, so patients and doctors knew who got which treatment.

The result was announced on 13 April 2026. The approval came 135 days later C (calculated), helped by a priority voucher the FDA granted in October 2025.⁹⁶⁸ In October the company showed that survival was better with daraxonrasib in every RAS subgroup it looked at, including patients in whom no RAS mutation was found.² Those are exploratory analyses, not planned tests.

The first five weeks of sales

Net product revenue in the third quarter was $197.6M, for about five weeks on the market. More than 5,000 patients got the drug.² The company says plainly that this is not a run rate. It includes patients moving over from its free access program, patients who had been waiting for approval, and the first stock put into the pharmacies’ shelves.² The CFO said the channel held under two weeks of stock at the quarter end, and that the gross-to-net discount, the gap between list price and what the company keeps, was at the low end of an expected 20% to 30%; he expects it to rise within that range as the mix of payers matures.¹² About 2,700 doctors prescribed it, and by patient volume about 40% of use was academic and 60% community.¹² Plans covering more than 90% of US pancreatic patients have paid claims, so far through medical exceptions and prior authorization while formal policies are written, and the commercial chief said some of those claims may not be paid once the policies are set.²¹² He also said the access program wound down in October, and he would not say more about new-patient demand before a full quarter.¹² I do not annualize $197.6M. It tells me the drug is wanted; it does not tell me how many new patients start each month once the waiting line is gone.

The company also sells outside the US where law allows: named-patient access in some countries. In October the European medicines committee recommended approval for the same patients, and the European Commission has yet to decide. The company is building teams in Europe, Canada and Japan.² In August it gave BeOne Medicines exclusive rights to four of its drugs in select Asian markets, while keeping Japan, South Korea and everywhere else. The financial terms are not in the filings I read.⁵

What else is in the pipeline

The rest of the value is a set of trials. Daraxonrasib hits many forms of RAS. Zoldonrasib targets G12D, elironrasib G12C and RMC-5127 G12V, one mutation each.⁵ The company is running or starting eight late-stage trials across pancreatic, lung and colon cancer, and I give my odds in the table.

Where it came from

Revolution Medicines was founded in 2014 and is based in Redwood City, California. It bought Warp Drive Bio in 2018, took $50M up front from Sanofi for an earlier project, listed on Nasdaq in February 2020 and merged with EQRx in an all-stock deal in November 2023 that gave it $1.1B of cash.⁸⁴ Until August 2026 it had no product revenue. It had 883 employees at the end of 2025, 673 of them in research.⁸ In August it agreed to lease four buildings of 672,000 square feet for a new headquarters, with base rent of about $2.7M a month once all four are in use in 2028, less $23.1M of abatements through January 2030. The leases take effect only when the landlords close their purchase of the buildings, and end on their own if that has not happened by 29 April 2027. Funds managed by Farallon own a majority of the landlords, and Farallon also owns 6.4% of the company.¹⁴¹⁵

The takeover talk

The Financial Times reported on 8 January 2026 that Merck was in talks to buy the company for $28B to $32B, as pharmaphorum relayed on 9 January. On 26 January BioSpace reported that the Wall Street Journal had said the talks ended over price. Both stories rest on unnamed sources. Neither company confirmed anything in the articles I read. BioSpace, citing Reuters, says AbbVie, also named in early reports, denied them.¹⁶¹⁷ The stock closed at $117.63 on Friday 23 January and $97.78 on 26 January, down 16.9% in a day.¹ I checked EDGAR on the day I wrote this. The company has filed no merger agreement, tender offer or solicitation since the September quarter report, and the only such filings on record date from the 2023 EQRx merger.¹⁸ The company is independent, its shares still trade on Nasdaq, and no bid sets a ceiling on the price. I treat the reported $28B to $32B as a data point about what one buyer would pay, and no more.

IIThe moat

Revolution’s protection is not a law, as it is for a railway or a utility. It is a head start in a new class of drug, patents, and the cost and time rivals need to catch up. That is thinner than it looks from the price, and thicker than for most small companies.

The lead

The company says RASONQUE is the first approved medicine from the class of RAS(ON) inhibitors.⁵ It has five FDA breakthrough designations across its drugs and five orphan designations.² That is a good record of regulators agreeing that the drugs matter. It is not a barrier. The 10-K lists, by target, the companies working on RAS. By my count it names 20 on G12C, 22 on G12D, 6 on G12V and 29 on drugs that block all RAS or all KRAS, with some overlap, among them Amgen, AstraZeneca, Boehringer Ingelheim, Lilly, Pfizer and Roche.⁸ A lead of two or three years is real. A lead of ten is not something I can see.

The patents

The 10-K says patents covering the company’s RAS drugs and platform expire between 2031 (for the older Warp Drive portfolio) and 2045, before any extension, and patents on the companion drugs between 2035 and 2043.⁸ A US patent can be extended by up to five years to make up for time lost in review, but not beyond 14 years from approval, which for RASONQUE would be August 2040.⁸ I do not know which patents cover the approved compound, and the filing does not say. I assume protection lasts to about 2040 and show what happens if it ends three years sooner: the base value falls from $89 to $50 E (estimate). That is the largest single input I cannot check, and it is why I put it in the table of single changes.

What is not protected

• Manufacturing. All of it is outsourced to contract manufacturers, and some are the sole source of a key material.⁸ The company says it expects its stock and planned production to meet demand.⁵
• Price. The US inflation law lets Medicare negotiate prices nine years after launch.⁸ In September 2026 the government published a final rule, called Globe, that makes drug makers pay rebates on part of Medicare-reimbursed units, based on prices abroad. It faces legal challenges, and the company says its effect is likely to be significant and cannot yet be determined.⁵
• One drug. Every dollar of revenue so far is one tablet. Zoldonrasib and elironrasib are earlier.²

My base assumes protection lasts to 2040. I cannot back that with a number from a filing. If I put a one in four chance on losing it by 2037, the base would be about $10 lower; at one in two, about $19 lower C (calculated).

IIIThe competitive advantage

The last section says what protects Revolution. This one is about what it does better than the people chasing it.

A drug with a result doctors can see

The first thing is efficacy. A hazard ratio of 0.40 for survival in a cancer where chemotherapy gives a median of about six months is the kind of number that changes guidelines, and it did. The US guidelines list RASONQUE as a category 1 preferred option after earlier treatment, and as an option “Useful in Certain Circumstances” as first treatment.² It is an oral tablet taken once a day, so it needs no infusion chair.⁶ In the trial, the benefit held in the group with no RAS mutation found.² I treat that last point as promising and unproven, because it came from an exploratory look.

Speed

The company went from topline results on 13 April to approval on 26 August, 135 days, using the FDA’s priority voucher.⁹⁶⁸ It had made the tablets ahead of time, which is why the cost of sales in the third quarter was $0.3M.² Planning ahead like that takes cash: the company had $3.7B when it launched.² It also meant the first sales were large.

A family of drugs, not one

Daraxonrasib hits many RAS forms, in many tumours. Zoldonrasib, elironrasib and RMC-5127 are built to hit one form each, with the aim of using them with chemotherapy and immunotherapy earlier in the disease.⁵ A new class, with a first human trial due in the fourth quarter, is aimed at resistance, the problem that will eventually return to any targeted drug.² If daraxonrasib stops working in a patient, Revolution would like to have the next drug ready. The data to show that works is some way off.

What it looks like on the ground

I do not read much into any one of these lines. The spending line, though, is clear. Selling cost nearly tripled in a year, to $153.3M, and research cost rose 59%.² The company is spending as if it expects the revenue.

Where the advantage is thin

Tolerability is the first. Rash, diarrhea, mouth sores and nausea are common, 30% of patients had a serious reaction, and doctors who treat patients who are already ill will watch the numbers.² Competition is the second. A G12D drug from a larger company with a cleaner profile could take the patients who carry that mutation, the group Revolution’s own zoldonrasib is aimed at.⁸ That is not abstract: OncLive reported on 28 August that a phase 3 of Incyte’s G12D drug INCB161734 in first-line pancreatic cancer, about 588 patients, is enrolling at more than 200 sites, with chemotherapy plus the drug against chemotherapy plus placebo.¹⁹ Price is the third. And a drug that lifts median survival from 6.7 to 13.2 months still leaves half the patients dead by about 13 months, so the first-line trial matters as much as the launch: the money is earlier in the disease.

IVManagement

Who runs it

Dr Mark Goldsmith has been chief executive since November 2014 and chair since 2020. He is 64. Before this he ran Constellation Pharmaceuticals and Global Blood Therapeutics and was a venture partner at Third Rock.¹⁰ The team around him is stable: Jack Anders, who joined in 2018, has been chief financial officer since 2022; Stephen Kelsey has led research since 2017; Margaret Horn is chief operating officer. Anthony Mancini, who ran operations at Genmab, joined in March 2025 to lead the commercial launch, and Alan Sandler, who was chief medical officer at Mirati, joined as chief development officer in September 2025.¹⁰ The lead independent director is Alexis Borisy, who chaired EQRx until the merger.¹⁰ The chair and the chief executive are one person, which the board calls appropriate.¹⁰ The record is good: a company founded in 2014 has an approved drug and a phase 3 program.

How they are paid, and what that pays them to do

Dr Goldsmith’s pay for 2025 was $14.67M: $0.81M of salary, $9.0M of stock options, $3.97M of stock awards and a $0.89M bonus.¹⁰ About 70% of the equity is options and 30% restricted shares by value, and all of it vests with time over four years, with no performance condition.¹⁰ The bonus depends on goals the board sets. Research and development goals carried 55% of the target and scored 87 points; commercial, corporate and organization goals carried 45% and scored 59 points. The total was 146% of target.¹⁰ The board raised the bar in July 2025 to cover a global launch, and says it did not lower any goal.¹⁰

What that pays people to do is hit milestones: start trials, win approvals, raise money. Options pay for a higher share price and nothing else. Neither the bonus nor the equity depends on revenue, profit, return on capital or the share count. In a development company that is common, and until now reasonable. A commercial company with a $10B revenue possibility needs a different set of rewards, and I will look at the 2027 proxy for whether it has one.

How much of their own money is in

On 31 March Dr Goldsmith held 798,456 shares, worth $159.8M at $200.15 C (calculated), and could exercise another 1.94M options within 60 days.¹⁰ All directors and executive officers together held 7.6% including options, mostly through outside investors who sit on the board: Thilo Schroeder of Nextech held 3.8% and Mr Borisy 1.2%.¹⁰ Those are meaningful stakes. The policy bans hedging and pledging, and a clawback rule has applied since November 2023.¹⁰

The 2026 Forms 4 add a second side. Twelve insiders sold 849,438 shares for $140.3M, at an average of $165, and none bought.¹¹ Dr Goldsmith sold 278,879 shares for $45.3M. Of those, 254,010 came from options he exercised this year at $0.54 to $4.73 a share, at a total cost of about $1.1M C (calculated).¹¹ Forty-two of the 43 Forms 4 that report sales cite a Rule 10b5-1 plan, which sets sales in advance.¹¹ About $19.0M of the sales came before the April trial result, near $100 a share, and $48.5M after the August approval C (calculated).¹¹ I read that as normal selling by long-time holders of cheap options, and I do not read it as a view on the stock. Still, it is a different picture from the one in the proxy, and a reader who wants someone with the same exposure as an outside owner will not find it here.

Who else owns it

Schedule 13G filings show FMR at 7.2% and Vanguard Capital Management at 5.13% on 30 September 2026, and, on 30 June, Farallon at 6.4%, BlackRock at 5.7%, Baker Bros. at 5.3% and Janus Henderson at 5.1%.¹⁵ Alphabet’s Form 13F-HR for the quarter ended 30 September 2026, filed on 6 November, lists 2,673,909 shares of Revolution Medicines, valued at $548.8M at quarter end, which is 1.2% of the shares outstanding C (calculated). The filing shows the line under GV Management Company, L.L.C., one of the managers it includes.²⁰²¹ The position has shrunk: the GV line was 4,080,700 shares on 30 September 2025, 3,292,525 on 31 March 2026 and 2,946,619 on 30 June 2026, so the reported holding is down by a third in a year. A 13F shows quarter-end holdings and not the reason, so I cannot say whether that was selling or a transfer.²²²³²⁴

What they did with capital over a full cycle

The company has never earned a profit. It has paid for itself by issuing shares, usually at good moments. Shares outstanding rose from 66.6M at the end of 2020 to 216.3M, 3.2 times.⁴⁵ In round numbers it has raised about $4.7B by selling new shares since the 2020 IPO C (calculated): $2.1B net from underwritten offerings and $0.6B gross from at-the-market sales through 2025, then $1.65B net from the April offering and $0.35B net from at-the-market sales in 2026. The $1.1B of cash that came with EQRx is on top.⁸²⁵⁵ The April 2026 offering raised $1.725B at $142, the day after the trial result was public; the stock closed at $97.78 in January and $200.15 now.²⁵¹ On the evidence, management has been good at raising money at moments of strength. For an owner that is not free: each offering shifts a bigger share of any future profit to new holders, though the sale price and the later returns are what count.

One more point of record. In June 2025 and May 2026 the company chose to pay Royalty Pharma a share of future sales rather than only issue shares. That keeps ownership intact and costs about 4.55% of the first $2B of sales each year, which I model as a cost.⁵ In October it declined an optional third tranche of up to $250M because of its cash.³ I count that as a sign of restraint.

VFinancials

Revolution Medicines has never made a profit. The eight years of numbers tell one story: revenue that came from a partner and then stopped, costs that grew about 19 times, and a share count that kept growing to pay for it. 2026 is the first year with a product.

The pattern is clean. Research cost grew from $51M in 2018 to $987M in 2025, 19 times, and is $1,156M in nine months of 2026.⁴⁵ The extra growth in 2026 is in selling: $365M in nine months after $195M of general cost in all of 2025. Revenue was $12M to $50M a year from the Sanofi program until it ended in 2023, none in 2024 and 2025, and then $197.6M in one quarter.⁸²

What an owner could really take out

The reported net loss for the twelve months to September was $1,871M. The cash an owner would have taken out, after stock pay and after capital spending, was −$1,691M C (calculated). That is 90% of the loss. The loss includes non-cash charges that I do not count: $207.5M in nine months for a rise in the value of the warrants the company inherited from EQRx, and $66.6M of interest on the Royalty Pharma money that is accrued and not paid.⁵ The cash flow statement adds back stock pay, $250.6M in the twelve months. I take it away again, because shares given to employees are paid for by owners C (calculated).⁴⁵ The fairest summary is that about $1.7B a year went out, and I count all of it.

The capital needed to stand still is tiny today: $9.4M of capital spending in the twelve months, because the company leases and uses contract manufacturers.⁴⁵ That will change. The new headquarters is 672,000 square feet, and although the landlords will provide $115.9M for fitting it out, the company will pay about $2.7M a month in base rent once all four buildings are in use in 2028, after abatements end in 2030, and more in build costs, which my model puts at $130M to $150M a year of capital spending in 2028 and 2029 E (estimate).¹⁴

Quality of earnings

• The first sales are not clean. Of the $197.6M, some is stock put on pharmacy shelves, some is patients moving from the free program, and some is a waiting list that cleared. The company says so.² Two customers, both specialty pharmacies, account for 71% and 28% of the revenue and 72% and 28% of the $230.7M of receivables.⁵ Receivables rose by $230.7M in the quarter against revenue of $197.6M, so the cash from the first sales had barely started to arrive C (calculated).
• Cost of sales is a placeholder. $0.3M, because the tablets sold were made before approval and their cost went through research.² I use 8% of revenue E (estimate), a guess for a small-molecule pill made by contractors.
• Stock pay is large. $216.9M in nine months, 14% of operating costs, and the company guides to $270M to $290M for 2026.⁵²
• The loss includes items that are not operating. Net loss was $407.5M in the third quarter and operating loss $373.1M. Interest income was $37.8M and interest expense $33.8M; a $40.7M warrant charge made up the difference.² The warrants expire in December 2026, after which that line ends.⁵
• Taxes. The company has paid no income tax. At the end of 2025 its federal loss carryforwards were worth $315.4M as a deferred tax asset, which is about $1.5B of losses at 21% C (calculated), before the 2026 loss.⁸ I assume $3.0B of usable losses by 2027 and a 24% tax rate on profit after them E (estimate).

The balance sheet

On 30 September 2026 the company held $3,651M of cash and securities, and in October it received $250M from the first Royalty Pharma loan, which pays three-month SOFR (at least 3.5%) plus 5.75% and falls due on 8 October 2032.²³ Against that it owes $500M of 0.50% convertible notes due 2033, which convert at about $198.80 a share, and it has sold $572M of future royalties, which sits on its balance sheet as a liability. Lease liabilities are $151M, and the new headquarters leases are not yet on the balance sheet.⁵²⁶ Equity is $2,401M. The company expects to fund its plans for at least twelve months, and says it may need more money.⁵

The convertible notes deserve a note. They convert at $198.80, and the stock is at $200.15. At the market price I count them as 2.5M shares. In my own values, all below $198.80 except the bull, I count them as $500M of debt.²⁶⁵ The loan has a make-whole clause if repaid early and must be repaid on a change of control.⁵

The share count has risen 3.2 times since the end of 2020, from 66.6M to 216.3M.⁴⁵ Of the increase, 54.8M shares came from the EQRx merger and the rest from offerings, at-the-market sales and options being exercised.⁸ That rise was paid for with real money, and the question for an owner is the price at which the shares were sold, not the count. The April 2026 offering of 12.1M shares was at $142, which was 29% below today’s price C (calculated).²⁵¹ Outstanding options number 17.7M at an average exercise price of $43.13, plus 4.8M restricted share units and about 19.6M warrants, each for 0.1112 of a share at $11.50 a warrant, 2.2M shares in all; the warrants expire on 17 December 2026.⁵ At $200.15 I count 238.6M diluted shares C (calculated), and that means a diluted value of $47.7B against $43.3B on the shares outstanding.⁵

On the price chart there are two jumps. One is the climb through the trial result and the approval: from $97.78 in January to $147 on 14 April, the day after the trial result, and to $221.15 on 27 August, the day after the approval. The other is the day after the Q3 results on 5 November, when the stock fell 11% from $199.70 to $177.65 and then recovered to $200.15 by Monday.¹ I do not read the 5 November fall as news about the company, since the results beat the analysts’ revenue forecast; the filings do not say what moved it.²⁷

VIValuation

At $200.15 the price is 125% above my base value of $89 and 36% below my bull of $315. My margin-of-safety price is $53. The price needs the first-line pancreatic trials and the lung trials to work, and it also needs more patients than I count.

First, how this goes badly wrong

Before any number, here is what would most likely hurt an owner, and I worked back from it. There are three ways.

The first-line pancreatic trials miss. The approved label is for patients who have already had treatment. The larger pools are earlier in the disease.⁵ RASolute 303 has not read out and the company gives no date. I give it a 75% chance of winning. If it misses, my base falls from $89 to $45, and the company would cut costs, which is what my bear case does: $22 E (estimate).

The launch settles lower than the first quarter. Revenue of $197.6M came from five weeks that include a waiting list and stocking, and the company says not to read it as a trend.² What matters is how many new patients start each month and how long they stay on. If each case is 20% smaller in every line, the base is $60; if the discount to list is 35% and not 25%, $70 E (estimate). Medicare rebate rules could add to that.⁵

The cost of waiting. The company spends $2.1B to $2.2B in 2026 and says 2027 will be higher.²¹² In my base, cash falls from $3.9B to about $0.8B before profits arrive. It has $1.0B of optional Royalty Pharma money left, and it has sold shares in every year since 2020.⁵⁴ If it sold $2B of stock at 70% of my base value, the value per share would be about $85.5 instead of $88.8, a fall of about $3 C (calculated). The bigger cost is the habit: each raise is easy, and the discipline of living on what you have is weaker.

None of these needs anything exotic. The first sets my bear. The second and third are why I start from the cash that reaches an owner and not from reported earnings.

What an owner could really take out

I value the cash left after stock pay and after the capital needed just to stand still, not the reported profit. At present that number is negative: −$1,691M over twelve months, against a reported loss of $1,871M. In my base it turns positive in 2030, at $623M, and reaches $4.2B in 2034, against profit after tax of $4.4B E (estimate). The gap there is capital spending and working capital, which is small for a company that rents its buildings and has others make its tablets.

What I count, and what I leave out

I value each year’s owner cash to 2040, then add a terminal value, then add cash and take away debt. The model starts from the September balance sheet and the October loan, includes the fourth quarter of 2026, and weights each trial by my odds. The Royalty Pharma money is a cost taken out of sales and not also a debt. I count no value for the earlier programs (RMC-5127, the new class, two others) and no revenue from BeOne’s territories, but I do charge their research cost, and each $500M a year of research after 2033 takes about $5 a share off the base C (calculated).

How wide the margin gets. My base reaches an operating margin of 58% in 2034, and my bull 73%. Merck earned 32% before tax on $65.0B of sales in 2025.²⁸ The one precedent I know near my base is Gilead during its hepatitis C drugs, at 68% in 2015, 58% in 2016 and 54% in 2017.²⁹ So I am assuming Revolution looks like Gilead in its best years and not like a typical drug maker, and my bull is above anything I can point to. If research and selling costs ran 50% above mine, the 2034 margin would be 46%, my base would be $54, close to my margin-of-safety price, and cash would run out in 2028, so the company would need about $4B of new money E (estimate).

Patients, share and months

Revenue is patients times months on the drug times net price per month. The patient pools are the hardest number on this page. The company gives the number of RAS-mutant cancers; I have cut them by stage, by how many get treated and by line of therapy, and the cuts are mine. Net price is the list price less 25%: $29,850 a month, so a six-and-a-half-month course earns about $194,000 E (estimate).

Read it from the top. The approved label alone makes $2.7B in the US if no first-line label ever arrives, or $1.9B once first-line use takes some of its patients, and every other line is a trial that has not read out. By 2034 I count $7.2B for the US and $9.9B with the rest of the world.

The odds, multiplied out

Each trial is a separate bet, though they share one idea. The base uses my odds as weights.

The five do not move independently. If a RAS drug fails first line in pancreas for a reason that applies to the class, such as toxicity with chemotherapy, the adjuvant and colon trials probably fail too. If it succeeds, odds in the others rise. That is why the bear and bull cases are whole worlds and not sums.

Revenue under each case

The bear peaks at $2.3B in 2030 and fades. The base reaches $4.4B in 2030 and $9.9B in 2034. The bull reaches $24.7B in 2034. For scale, Merck’s cancer drug Keytruda sold $31.6B in 2025.²⁸

The three cases

Bear

RASONQUE does well in the label it has and nothing else works. The first-line pancreatic trials miss, lung fails, the adjuvant and colon trials are dropped, and the company cuts research from $1.9B in 2027 to $350M by 2031. Share of the approved patients peaks at 56%, they stay on the drug six months, deductions are 28% and the price slips 2% a year. Revenue peaks at $2.3B in 2030. Value $22. I give it about one chance in ten.

Base

Each trial is weighted by my odds. The approved label reaches 65% of second-line patients and 55% of those not fit for combination chemotherapy. First-line pancreas is approved in 2029 with 75% odds and reaches half of first-line patients by 2034; lung, adjuvant and colon follow at their odds. Revenue is $4.4B in 2030 and $9.9B in 2034, spending peaks near $3.0B a year and cash falls to $0.8B. Value $89.

Bull

All five trials win, with larger shares and longer treatment, 1.5% price rises and a lower discount to list. First-line pancreas reaches 58% of patients, lung 38% in first line, colon 45%, adjuvant 50%. Revenue is $9.8B in 2030 and $24.7B in 2034. For scale, Merck’s Keytruda sold $31.6B in 2025, and the bull’s 2034 is three-quarters of that.²⁸ Value $315. The product of my odds is 1.6%, and I put the real chance, allowing for shared science, at about one in twenty E (estimate).

No case runs out of cash by itself. The lowest year-end balance is $0.8B in the base and $1.2B in the bear. I do not model any new shares. The company will probably raise anyway, because it has done so every year.

The price against the range

Cash per diluted share is about $16, 8% of the price C (calculated). Enterprise value is $44.1B. Of every $200.15, about $12 is what my base gives the approved label with the full research bill still running, and about $22 is what the bear gives it with the bill cut. The rest is a claim on trials that have not read out.

It helps to put the reported takeover talks beside this. On 8 January the Financial Times reported that Merck was in talks at $28B to $32B.¹⁶ On the 197.0M shares outstanding at the end of 2025, before options and other dilution, that is $142 to $162 a share C (calculated).⁴ Today’s diluted value is $47.7B. Nothing in the filings says a bid was made. But it shows how one large buyer, with lower costs of capital and an existing sales force, may value the company. My 12% rate is for an owner who has no such advantage.

What you must believe at today’s price

Three roads lead to $200.15. One makes everything bigger: 1.8 times my revenue in every year, or $17.7B in 2034. One makes the main trials certain: with first-line pancreas and both lung programs winning, my base gives $127, and the sizes would need to be 1.41 times mine on top. One makes everything certain: with all five trials winning, my base gives $153, and the sizes would need to be 1.23 times mine on top. A 10% discount rate gets to the price with 1.5 times my revenue. None is absurd. Each asks for more than I can see.

Multiply it out. The bull needs five trials to win, about 1.6% on my odds and perhaps 5% with shared science, and then the sizes to be the large ones. A price that is 36% below that bull is a price that already assumes most of it.

Discount rate and sensitivity

The 10-year Treasury yields 4.88%.³⁰ RVMD’s weekly returns have moved 1.33 times the S&P 500’s (SPY) over five years and 1.59 times over two C (calculated).¹³¹ With an equity premium of 5% that gives 11.5% to 12.8%, and I use 12%. I am already weighting the trials by odds, so the 12% is for the market risk of a business whose profits are far away, not for trial risk. The rate matters less than revenue: moving it from 12% to 10% takes the base from $89 to $109, and moving revenue 20% up takes it to $117.

Single changes to the base case

The two biggest levers are the first-line trial and the date exclusivity ends. Three years of earlier loss of exclusivity costs more than the lung programs combined. I cannot check that date, and I show it for that reason.

The margin-of-safety price

My margin-of-safety price is $53, 40% below my base. At that price my base cash flows return 18% a year, six points above my 12% hurdle, and an owner pays 2.4 times the $22 bear value, not the 9 times it costs at the market price. I use a wide margin because the range is wide: $22 to $315 cannot be narrowed by anything I can read. It is a cushion against trials failing, not against everything. In my bear the shares are worth $22, so even $53 loses 59% there. The stock closed at $62.73 on 10 November 2025, within 20% of $53, so this is not a price from another era.¹

Quality and price are two questions

Is it a good business? It has a drug that nearly doubled survival in a cancer with little to offer, a class that may work in lung and colon, and $3.9B of cash. It has also lost $1.9B in twelve months, sold shares every year, and has one product. Is it a good price? At $200.15 the price is above my base, and it pays for trial wins that have not happened. A good business can be a poor purchase at a price that pays for the pipeline in advance. This looks like that to me, with the caveat that my base is probability-weighted and the market may weight the same trials more heavily. I am not saying the shares will fall.

The alternative use of the dollar

A dollar in a 10-year Treasury earns 4.88%.³⁰ At $200.15, my base cash flows return 5.2% a year, which is about the Treasury and well below my 12%. On this site, the Eli Lilly page has a price 91% above my base value there, for a business already earning large profits, and the Iovance page has a price 40% above base, for a launch-stage company with a smaller case.³²³³ Revolution’s gap, 125%, is larger than either. Going the other way, the Booking Holdings page has a price 34% below my base value, for a business that earns money today; its base uses its own discount rate, so the comparison is rough.³⁴ Even with the first-line trial certain, my base gives $103, which is below the price. A reader who puts the five trials at near-certainty and the sizes above mine will reach a different answer. That is a real difference of view and not an error on either side.

What I cannot judge

• The first-line pancreatic trial. I don’t know its result. My 75% could be 55% or 90%.
• How many patients start each month. The first quarter is a waiting list. I can’t see the steady state and the company won’t say.²
• How long patients stay on. Median time without progression was 7.2 months. The average time on drug in real life may be longer or shorter.⁷
• Which patent covers the compound, and when it ends. My value drops by about $39 if it ends three years early.
• Net price. Rebates under the new Medicare rule, deductions to hospitals and the pharmacies’ share are not yet known.⁵¹²
• BeOne’s terms, and the earlier pipeline. Both are unpriced here; they would add to my value, probably by single dollars a share.

Because I can’t know these, I show a range and where the price sits in it, and I do not treat $89 as more precise than it is.

VIIThe case for the price

What most people miss

Three facts cut in the company’s favour and get less notice than the survival number. First, the label names no mutation, and in October the company showed better survival in every genotype subgroup it looked at, including patients with no RAS mutation found, so the pool is not limited by a test. That analysis was exploratory.⁶² Second, the US cancer guidelines already list RASONQUE for locally advanced pancreatic cancer and for KRAS-mutant lung cancer, neither of which the label covers; the company says the guidelines do not change the label.² Third, BeOne is funding and running the colon trial, RASonate 312, on Revolution’s behalf. It is part of the August arrangement in which BeOne also got rights to four of the drugs in select Asian markets; the financial terms are not disclosed.²⁵

The bull case

The case is made best by the company, a fund that held the stock in July and the analysts. Dr Goldsmith said in the third-quarter release that the launch “marked a monumental step forward” and the quarter was “transformational”.² Baillie Gifford’s Edinburgh Worldwide trust wrote in July 2026 that it had held the stock since late 2025, that several sell-side forecasts have daraxonrasib above $10B, and that “The valuation already reflects some of the opportunity”.³⁵ MarketBeat reported on 8 November target increases around the results (HC Wainwright $350, RBC $321, Guggenheim $298), an average target of $263.15, and that Brookline Capital Markets, which assumed coverage on 28 September at $318, cut its 2030 estimate to $17.41 a share from $18.15.²⁷ The big money is first line, in pancreas, lung and colon. If those trials win, sales could be many times the first quarter’s, as in my bull case, $24.7B in 2034. My probability-weighted base is $9.9B, so I am not far from the sell-side figure; the gap is in what I charge for risk and cost.

Against reality

If the first-line pancreatic trial and both lung programs win, my model gives $127 a share at my sizes, not $200.15. The price then needs the launch to come in 1.4 times my pools, months on drug and price; with all five trials winning, 1.23 times. That is how the price is possible: several trials win and the drug does better than I assume. My bull case, $315, has all five winning with larger sizes, and the price is 36% below it. Today’s numbers are far short. The company has one approved use and $197.6M of first-quarter revenue that it says is not a trend, and nothing has read out in first-line pancreas, lung or colon. My base earns $3.55 a share in 2030; Brookline’s $17.41 is near what my bull earns, $19.48 C (calculated).²⁷ The price asks for about 1.8 times my probability-weighted revenue, or $17.7B in 2034.

It helps to multiply it out. The first road needs the first-line trial to win (75%), both lung programs to win (21%) and a launch 1.4 times my sizes, which I put at one in three: about 5% E (estimate). The trials share science, so the true figure is higher, perhaps one in ten, but it is not high.

What a 10x would take. Ten times $200.15 is $2,001.50, which on 243.0M diluted shares at that price is $486B of market value C (calculated). At 20 times earnings that needs $24.3B of net income a year. At a net margin of 55%, above my base’s 44% and near my bull’s, that is $44B of revenue C (calculated). Merck’s Keytruda sold $31.6B in 2025, and my bull case reaches $24.7B in 2034, so a tenfold needs 1.8 times my bull, with a drug as large as Keytruda as the floor.²⁸ Ten times in ten years is 26% a year. Merck’s non-affiliate market value was about $198B on 30 June 2025, so ten times here is 2.5 times that.²⁸ A few winners have done it from a far smaller base: Lilly rose 19 times in ten years, on this site’s page.³² Revolution itself has risen 14 times since its low close of $14.54 on 12 May 2022.¹ My bull case is worth 1.6 times today’s price, and a tenfold needs revenue 1.8 times that bull’s, so I call it close to impossible.

Where I come down

Three things moved me: the survival result, the 135 days from result to approval, and the cash, which means no case of mine needs new money to survive. The first quarter moved my 2026 estimate and not my base, because a quarter with a waiting list in it is not a rate. The price did not move me: it is 125% above my base. What would raise my values is evidence on the first-line trial, since certainty there alone is worth $14 a share, net revenue per patient above the $194,000 I assume for a course, or proof that the patents last past 2040 E (estimate).

VIIIRisks, and what would change my view

The three ways this goes badly wrong are at the start of Valuation. These are the rest.

• Customers. Two specialty pharmacies bought 71% and 28% of all RASONQUE sales in the third quarter and owe 72% and 28% of the receivables. A problem at either would be felt at once.⁵
• One drug and one supplier chain. All production is outsourced, and some suppliers are the sole source of a key material.⁸
• Safety in the real world. The label carries warnings for skin and soft tissue, mouth, bowel perforation and lung inflammation. Trial patients are chosen. Real-world patients are sicker.² The company said early tolerability was “so far so good”.¹²
• Pricing and rebates. The US inflation law allows negotiation nine years after launch. The Globe rule makes manufacturers pay rebates on part of Medicare-reimbursed units, based on prices abroad, and the company calls the effect likely to be significant. It is under legal challenge. The 10-Q also lists a proposed rule called Guard, a Medicaid model called Generous and tariffs on imported brand drugs, in force since 29 September for importers outside a named list, with exemptions, and the company does not say what they would cost it. The commercial chief said some early claims paid by exception may not be paid once formal policies are written.⁸⁵¹²
• Competition. The 10-K names dozens of rivals, some very large. The first approved drug does not always keep the market.⁸
• Money. The company has sold shares every year and has a high-interest loan secured against substantially all its assets, due in October 2032, plus a convertible that may become 2.5M shares. The Royalty Pharma rate also steps up from 2030 to 2041 in any year after sales fell below an agreed threshold, which the filing does not give; the step-up covers only the first $2B of sales and stays in single digits. My model charges a flat 4.55% there, and it would matter most in my bear case.⁵³
• Governance. The chief executive is also chair. Farallon, a 6.4% holder, manages funds that own a majority interest in the landlords of the new headquarters, as the 8-K discloses.¹⁰¹⁴
• Costs of its own growth. The company will have 672,000 square feet of space, and it had 883 people at the end of 2025 and more since. Space and staff, once added, are hard to cut.⁸¹⁴
• BeOne. The company has given up rights in select Asian markets for terms I can’t see.⁵