The person who harvests the “King of Cancer”

On August 26, the FDA approved Revolution Medicines' RMC-6236, trade name RASONQUE, generic name daraxonrasib. The share price immediately surged to a historical high, with a market value of about $44.5 billion.

A year ago, the market cap of this company was only about $8 billion.

In one year, it gained $36.5 billion in incremental value.

Revolution was not the pioneer in the RAS field. The world’s first approved KRAS inhibitor was Amgen’s sotorasib, in May 2021; the second was Mirati’s adagrasib, in December 2022. While these two companies validated the “RAS as a druggable target” hypothesis in NSCLC, Revolution was still a mid-sized biotech founded on the chemistry of natural products, holding onto SHP2 and mTOR projects, and just starting out on its RAS(ON) pipeline.

Ranked by approval sequence, Revolution was the third to “pick the RAS fruit.” Yet, the market now prices its product over one hundred times the annual sales of Amgen's groundbreaking drug.

Sanofi had terminated its SHP2 collaboration with Revolution as early as June 2023, refocusing R&D priorities. Now, the RAS(ON) pipeline accounts for over 90% of Revolution’s valuation. This $44.5 billion is almost entirely a bet on its RAS success.

How did a latecomer seize the biggest reward?

The Sweetest Fruit Was Always Out of Reach

To answer this, let's first understand the landscape of the RAS target.

The KRAS mutation spectrum is like a disk, with different positions marked on its edge: G12C, G12D, G12V, G12R, G13D, Q61H. Amgen and Mirati, with two covalent drugs, hit the G12C spot.

This was indeed the right fruit. In smoking-related non-small cell lung cancer, G12C is the main druggable mutation, accounting for about 8.38%; Amgen’s CodeBreaK 100 achieved a 36% ORR, Mirati’s KRYSTAL-1 reached 43%. The “undruggable RAS” curse of thirty years was broken.

But shift your focus from lung to pancreatic cancer, and the picture changes drastically.

Pancreatic cancer is the most RAS-dependent tumor, with over 90% of cases carrying KRAS mutations. But its dominant mutation isn't G12C; it's G12D (about 35-42%), followed by G12V (about 20-32%) and G12R (about 14-17%). G12C is only 1-2% in pancreatic cancer.

This is the misalignment: Amgen and Mirati proved target-drug effect but only addressed the edge of the “disk”; the larger, sweeter fruits at the center—G12D, G12V, G12R—remained hanging, untouched.

The reason for this vacancy is a hard one—at the level of chemistry.

G12C was conquerable because of two opportunities: Glycine’s mutation to cysteine provided a strong nucleophilic thiol group for covalent locking, and there was a switch II pocket when GDP-bound (OFF). What about G12D? Aspartic acid’s carboxyl is only weakly nucleophilic, so classical covalent warheads can’t lock on; in the GTP-bound (ON) state, conformational rearrangement collapses the switch region pocket entirely.

This led to an absurd situation: humanity proved RAS can be drugged, yet the deadliest cancer driven by RAS remained untouchable. Pancreatic cancer, still the “King of Cancers,” has a median survival of just over one year for all stages combined. For forty years, countless pharma companies have fallen here, and the “King” had never been truly harvested.

The target-drug relationship is proven, but the sweetest fruit still hangs on the tree. That's the opportunity left for latecomers.

The Hardest Question Is “How To”

There are two types of risk in new drug development, often conflated but fundamentally different.

The first is target validation risk: Will inhibiting this target shrink the tumor?

This is life-or-death; failure means the whole direction is invalidated and billions lost. Amgen and Mirati answered this for the whole industry: It works. They took on the highest-risk challenge and earned their place in history.

The second is feasibility risk: The target is proven, but can you make a molecule to inhibit it?

For RAS, the challenge shifted. By 2021, Amgen eliminated the first type of risk; for G12D and pan-RAS, the feasibility risk instead soared to its peak.

The reason is simple.

G12C’s path to success was “find a mutation-specific covalent handle,” but that handle is missing from G12D and other RAS targets. The surface of RAS(ON) is smooth as a wall: GTP is bound with picomolar affinity, and intracellular GTP is abundant—so competitive displacement is unrealistic. There are no easily accessible covalent residues like G12C. The classic “find pocket, push molecule” approach is dead here.

So the real challenge isn’t “should we drug RAS?” but “what other paths, besides G12C, can work?”

Revolution's answer was a conceptual leap: don’t seek a pocket on RAS, bring in another protein to build the pocket together.

Pocket Makers

This approach has a name: tri-complex inhibitor (TCI), also called “molecular glue” RAS inhibitors or RAS(ON) inhibitors.

The assembly proceeds in two clear, measurable steps. Using daraxonrasib as an example: The drug first binds the abundant intracellular chaperone protein cyclophilin A (CypA), with a dissociation constant of 55.3 nM; this binding reshapes CypA’s surface to create a “neomorphic interface”—one not found in nature. This binary complex then targets GTP-bound RAS, with affinity for G12D of 131 nM, G12V 364 nM, and 154 nM for wild-type.

After these steps, CypA’s volume essentially shields RAS’s effector-binding face in space. RAF, PI3K, and RALGDS can’t bind, and the signal is off.

Structural biology explains its broad spectrum: The CypA-drug-RAS complex leaves an unoccupied groove along the Q61-G12-G13 axis, exactly accommodating the bulky side chains at these oncogenic hotspots; the drug only contacts highly conserved effector lobe residues, untouched by sequence differences among KRAS/NRAS/HRAS. With one molecule, you address multiple mutations and three isoforms.

The tougher breakthrough was with zoldonrasib (RMC-9805): how to target “untouchable” aspartic acid by covalency. The solution: use the tri-complex to bring CypA to RAS(ON), precisely position an aziridine warhead next to Asp12, and let the new interface catalyze covalent bond formation. Modification rate exceeds 95%, second-order rate constant of 89.9 M⁻¹s⁻¹, and under 1% for wild-type KRAS.

It’s not that aspartic acid became reactive—rather, the tri-complex created a private reaction chamber for it.

The route began humbly. The first probe compound bound CypA at only 897 nM, RAS with submicromolar affinity, and its IC50 for RAS-BRAF inhibition was 4400 nM—activity so weak it was almost negligible. Its only merit was to prove “the principle works.”

What bridges the gap from 4400 nM to an approved drug? Structure-guided SAR iteration, macrocycle redesign, and formulation engineering to keep even bRo5 molecules orally bioavailable. But the biggest step was going from “no solution” to “there's a way.”

After the Answer Is Revealed

The “use CypA to make a pocket” strategy was published (two Nature papers in 2024, zoldonrasib's Science paper in 2025); crystal structures entered the PDB, making subsequent process engineering more standardized: designing macrocycles, synthesis, measuring IC50 by TR-FRET, solving cocrystal structures, optimizing PK, and pushing into the clinic.

The path is hard, but can be broken down, outsourced, and accelerated with capital.

Thus, a crop of rivals has sprung up.

Erasca’s ERAS-0015 (oral pan-RAS molecular glue) and ERAS-4001 raised $260 million in a single round; Jacobio’s JAB-23E73; Betta Pharmaceuticals’ BPI-572270 (pan-RAS, non-degrader molecular glue); Arnault’s AN9025; Amgen’s AMG 410; Bayer’s $1.3 billion Kumquat buyout. The same mechanism, same “molecular glue + CypA” model, being reiterated in the chemical languages of a dozen companies.

There is only one party with the original answer, but dozens can copy it. The window of time Revolution bought with its insight is visibly narrowing.

They are not defenseless. Tri-complex platform patents, combination RAS(ON) therapies, and a network of five phase III trials make a real moat. But patents protect compounds, not “ideas”.

First-Mover Trap: The $600 Million Epitaph

The sharpest lesson comes from history.

First-generation KRAS drugs: sotorasib’s estimated global sales for 2025 are $360–400 million, adagrasib’s at $205 million. Combined, two era-defining drugs only sold about $600 million a year.

As the industry quipped, “KRAS accounts for 25% of human cancers, yet delivers less than 1% of oncology drug revenues.”

Amgen was FIC; it broke a forty-year curse and earned a place in textbooks, but didn’t reap commercial rewards. The first to prove a concept is rarely the one who harvests the market.

Today, Revolution sits in the same spot. Its $44.5 billion market cap stands against a consensus of just $100–150 million in sales for 2026. What it holds is a big post-dated check, not cash in hand.

Challengers are ready to play. Among pan-KRAS inhibitors that skip HRAS and NRAS, Jacobio’s JAB-23E73 offers a “me-better” paradigm: In second-line pancreatic cancer, ORR hits 38.5%, efficacy on par with the pioneer; rates of grade-3+ adverse events only 14%, skin rash 14%, and oral mucositis just 1.2% with no grade-3+ events. Against daraxonrasib’s body-wide rash, oral mucositis, diarrhea, even ILD warnings, the difference is clear. In December 2025, AstraZeneca licensed global rights outside China for a $100 million upfront and up to $1.915 billion in milestones—a new export record for a Chinese small-molecule anti-cancer agent.

Some analysts are direct: daraxonrasib’s side effects leave later entrants room for differentiated competition. Same efficacy, better safety; the most classic disruption in pharmaceutical history.

Three Hurdles

Will this post-dated check ever be cashed? Three gates to pass.

First, can first-line replicate second-line results? RASolute 302 doubled median OS from 6.7 to 13.2 months (HR=0.40; 60% reduction in death risk)—a figure never before seen in pancreatic cancer. But RASolute 303 (first-line) has only just begun, and includes RAS wild-type patients. With standard therapy moving earlier, daraxonrasib must prove added benefit, not just as a chemo alternative for intolerant patients.

Second, will NSCLC and CRC deliver? The current valuation rests on pancreatic cancer approval; clinical and approval outcomes in NSCLC and CRC remain uncertain.

Third, reimbursement and access. Priced at $39,800/month (~$478,000/year, ~3.22 million RMB), a privately insured patient’s copay can be as low as $0, but most pancreatic patients are elderly—Medicare Part D will take on most of the cost. Payer stance will directly decide the sales slope.

Another thing: On August 10, 2026, Revolution exclusively licensed its rights for four RAS(ON) inhibitors in countries across China, Southeast Asia, New Zealand and more—markets totalling about 2 billion people—to BeiGene, in exchange for funding and taking on a global phase III. It's leveraging partnerships, but also admitting biotech can hardly conquer the global RAS market alone.

In January of the same year, Merck offered $28–32 billion and was rejected. Three months later, RASolute 302 read out and the stock price soared from $79.65 at the year’s start to $207.68 today.

If they’d sold for $30 billion then, today's $44.5 billion would mean nothing to shareholders. This was the boldest gamble in a decade of RAS, but also the costliest show of confidence.

Genuine Rarities

Back to the original question: why could the latecomer harvest the biggest RAS fruit?

They didn’t get there “first”—but they tackled the “hardest.”

Amgen and Mirati proved RAS can be inhibited, but they plucked the lowest-hanging fruit. Pancreatic cancer—the most RAS-dependent, with the poorest prognosis, and the greatest unmet need—has G12D, G12V, and other broad RAS mutations that require a paradigm leap from “finding” to “making” pockets for therapies to work.

The scarcest thing in drug development is not slogging down an established road, but making the first move in uncharted terrain. Copying compounds, running bioassays, pushing clinical programs: all hard work, but breakable and, to some degree, outsourceable and accelerable with money. Yet the one idea—“use CypA to make a new binding interface”—was worth zero before it was conceived, and $44.5 billion after.

But the story has a long way to go.

Amgen proved a RAS molecule can inhibit the target, and won its place in history; Revolution proved RAS can be tamed by a platform, and won investor expectations. But history and expectations aren't cash. First-generation KRAS, barely reaching $600 million in annual sales after five years, remains a headstone at the track's start—reminding every latecomer: pioneers prove it’s possible, followers get paid; but if you can’t do better than the pioneer, you become the next one left behind.

The pan-RAS fruit left hanging for forty years is finally within reach. Whether it’s truly sweet depends on whether the path from second- to first-line, from pancreatic to lung and colon cancers, can become a true highway.

After all, the greatest reward is never for the first to climb the tree, but for the one who picks the fruit and makes it affordable for all to taste.