Exelixis Q2: Cabometyx Tops $570M, Zanzalintinib NDA Is a December 3 Coin Flip

Exelixis is the best single-product biotech you can buy. Cabometyx generated $2.3 billion in 2025 and is growing 10% a year. The company bought back $2.9 billion in stock and retired 93.3 million shares at an average of $31.12. Net income is running at $845 million annualized. At $52.20 and a $12.93 billion market cap, Exelixis trades at 5.1x forward sales and 15.3x GAAP earnings. That is not expensive for a profitable oncology company.

The problem is that Exelixis is priced for two approved drugs, and the second one is not approved yet.

Q2 Results: The Cabometyx Machine Keeps Running

Total revenue hit $628.7 million, up 10.6% from $568.3 million in Q2 2025. Cabometyx U.S. net product revenue was $570.6 million, up 10% year over year. Cometriq added $2.4 million. Collaboration revenue from Ipsen and Takeda royalties came in at $55.7 million, up from $48.2 million, driven by higher ex-U.S. sales of cabozantinib.

GAAP net income was $212.1 million, or $0.82 per diluted share. Non-GAAP EPS was $0.91, beating the $0.83 consensus estimate.

Management narrowed 2026 guidance. Total revenue is now $2.500 to $2.550 billion, down from $2.525 to $2.625 billion at the midpoint. Net product revenue guidance is $2.300 to $2.350 billion. The $50 million midpoint cut is real but immaterial: it reflects the natural tapering of a mature franchise, not a demand problem. Cabometyx volume grew in Q2. The wholesale acquisition cost increase of 3.0% in January helped. RCC, HCC, and NET prescriptions are all growing.

SG&A ran at $147.6 million, up from $134.9 million, driven by marketing spend as Exelixis prepares for the potential zanzalintinib launch. R&D spend was $212.0 million, up from $200.4 million, driven by clinical trial costs across the 6+ pivotal zanzalintinib trials now running. The R&D guidance range was lowered to $825 to $875 million from $875 to $925 million, which suggests some trials are enrolling slower than planned or costs are running below budget. That is a mixed signal: lower spend is better for earnings but implies the pivotal data timelines are stretching.

Exelixis repurchased $311.6 million of stock in Q2 at an average price of $47.85 and has now retired 93.3 million shares since 2023. The board authorized an additional $750 million buyback in May. At the current pace, Exelixis will have bought back roughly 15% of its outstanding float in three years. That is real capital discipline from a management team that could be spending that cash on risky BD deals.

The Zanzalintinib Pipeline: One PDUFA, Six Trials

The zanzalintinib NDA for previously treated metastatic colorectal cancer is under FDA review with a December 3, 2026 PDUFA date. The NDA is based on STELLAR-303, a Phase 3 trial of zanzalintinib plus atezolizumab versus regorafenib. The ITT population met the OS primary endpoint: median OS was 10.9 months versus 9.4 months for regorafenib, stratified hazard ratio 0.80, p equals 0.0045. That is a statistically significant but clinically modest benefit of 1.5 months.

The non-liver metastases subgroup missed. In the NLM group, median OS was 15.9 months versus an undisclosed comparator median, HR 0.83, p equals 0.1185. The NLM subgroup was a dual primary endpoint, so STELLAR-303 technically failed one of its two primary analyses. The FDA accepted the NDA anyway in February 2026, which signals the agency considers the ITT result sufficient. But the NLM miss creates PDUFA risk. AdCom is a possibility. A complete response letter citing the failed NLM endpoint is not the base case but is not zero.

Beyond CRC, the pipeline is dense. STELLAR-304 is the Phase 3 trial in first-line non-clear cell RCC, evaluating zanzalintinib plus nivolumab versus sunitinib. Topline results are expected in H2 2026. Non-clear cell RCC accounts for roughly 25% of metastatic kidney tumors and has no approved first-line therapy. In early-phase data, single-agent zanzalintinib showed a 38% objective response rate and 88% disease control rate in heavily pre-treated clear cell RCC patients (STELLAR-001 expansion cohort). If STELLAR-304 hits on PFS or ORR, zanzalintinib becomes the first drug indicated specifically for first-line nccRCC. That is a genuine unmet-need market and the readout is the most important catalyst between now and the CRC PDUFA.

STELLAR-311 is a Phase 2/3 trial of zanzalintinib versus everolimus in advanced neuroendocrine tumors, enrolling patients regardless of tumor site of origin. STELLAR-316, the MRD-positive CRC adjuvant trial with Merck and Natera, starts mid-2026 and tests zanzalintinib with and without Keytruda QLEX in patients who have molecular residual disease after surgery. The primary endpoint is disease-free survival. MRD-positive CRC is a high-risk population with no approved adjuvant therapy; if zanzalintinib can clear ctDNA, the market is substantial but the readout is 2027 or later.

Merck also initiated LITESPARK-034 in April 2026, testing zanzalintinib plus Welireg versus Welireg alone in second-line or later RCC. And STELLAR-201, a Phase 2 trial in recurrent meningioma, started enrolling in May. Meningioma has no approved systemic therapy, so any positive signal here is a first-in-class opportunity.

Competitive Picture: The RCC Battleground

The RCC market is the most competitive space in oncology. Cabozantinib already competes with Merck’s Welireg plus Lenvima, which showed a 30% PFS benefit over cabozantinib in the CLEAR-009 trial. That PDUFA is October 4, 2026, and we covered the Merck RCC setup in detail (see our Welireg+Lenvima PDUFA analysis). Arcus’s casdatifan posted a 15.1-month median PFS in late-line ccRCC in the ARC-20 trial and is advancing toward a Phase 3 (see our ARC-20 analysis). Merck also has LITESPARK-033 and 034 with zanzalintinib plus Welireg, meaning Exelixis is both competing with Merck and collaborating with them simultaneously.

This is the structural tension in the Exelixis thesis. Zanzalintinib is designed to capture the patients who progress on Cabometyx. If it succeeds in nccRCC, it expands the franchise into an indication where Cabometyx has no label. But if Welireg combinations displace Cabometyx and zanzalintinib in first-line RCC, Exelixis is defending two drugs against a Merck franchise that generated $716 million from Welireg alone in 2025. The Cabometyx franchise has room to grow in NET and the CRC adjuvant setting, but the RCC core is under siege from multiple directions.

Valuation: Priced for Approval

At $12.93 billion and $2.525 billion in guided 2026 revenue, Exelixis trades at 5.1x forward sales. That is higher than Bristol Myers Squibb at 2.6x sales and Gilead at 3.2x sales (we recently broke down both BMY and GILD Q2 results). It is also above the oncology biotech sector median. The premium multiple says the market is already pricing zanzalintinib approval into the stock.

Exelixis is not a large-cap pharma where a single drug is a rounding error. Cabometyx is the entire revenue base. If the zanzalintinib CRC PDUFA comes through on December 3, Exelixis becomes a two-product oncology company with a commercial infrastructure already built and an overlapping prescriber base. The launch would be efficient. Peak sales for zanzalintinib in CRC are probably $500 to $800 million; in nccRCC, another $300 to $500 million; in NET and adjuvant CRC, another $500 million combined. The total zanzalintinib opportunity is $1.3 to $1.8 billion at peak. At 4x peak sales, that is $5.2 to $7.2 billion in incremental pipeline value, or roughly $20 to $28 per share.

The stock at $52 already includes about half of that value. If zanzalintinib is approved, the stock re-rates to maybe $60 to $65 on launch optimism and multiple expansion toward 6x sales. If the FDA issues a CRL citing the STELLAR-303 NLM subgroup miss, the stock drops to $38 to $42, where the Cabometyx franchise alone is worth 3.5 to 4.0x earnings.

Risks

The STELLAR-303 NLM subgroup miss is the most immediate risk. The FDA accepted the NDA despite it, but the agency does not always signal review concerns before the PDUFA date. An AdCom would surface the NLM data as a debate point.

Concentration risk is significant. Cabometyx is the only commercial product generating 91% of revenue. The RCC franchise faces competition from Welireg combinations, casdatifan, and other HIF-2 and TKI combination approaches.

The STELLAR-304 nccRCC readout in H2 2026 is binary. If it misses, the zanzalintinib RCC thesis is dead and the stock gives back most of its pipeline premium. If it hits, the stock rallies ahead of the CRC PDUFA and the early nccRCC data gives the FDA another reason to approve.

Cash allocation discipline is a strength until it is not. Exelixis spent $2.9 billion on buybacks at an average of $31.12 per share. That looks brilliant at $52. At $45, it still looks fine. If the stock drops to $38 on a CRL, the buyback pace will accelerate and provide a floor. But buybacks do not replace pipeline execution. A CRL on December 3 resets the clock by 6 to 12 months and the stock earns its lower multiple for that entire period.

Verdict

Exelixis is a Hold at $52. The Cabometyx franchise is a durable $2.3 billion revenue stream growing at a high-single-digit rate with 15% earnings yields. The buyback program is the best in biotech. But the stock is priced for zanzalintinib approval, and the STELLAR-303 NLM subgroup miss introduces real regulatory risk that the consensus Hold rating does not fully price.

The playbook is to wait for the STELLAR-304 nccRCC data in H2 2026. If it reads out positively before the CRC PDUFA, the stock rerates on two upcoming approvals and you buy the pullback after the Q2 beat. If STELLAR-304 misses, the stock drops to $40 to $42 and you have a better entry point ahead of the CRC PDUFA, which still has a 65% to 75% probability of approval.

Do not buy the Q2 beat. The quarter was fine. The guidance cut was small. The stock at $52 is fair. The binary events between now and year-end determine whether Exelixis is a two-drug oncology platform or a one-drug story with a pipeline question mark.