This is investment research, not personal financial advice.

Novartis AG (NYSE:NVS) fell 13.9% to $137.70 on 8 September after del-desiran, the lead drug acquired with Avidity Biosciences for $12.0 billion, missed the primary endpoint in Phase III. The decline erased $42.5 billion of equity value in one US session. It also landed after pelacarsen failed a cardiovascular-outcomes study and Novartis paused eight rap-cel cell-therapy trials following three patient deaths (Novartis HARBOR 2026; CNBC 2026; Nasdaq 2026).

The market reaction reaches much further than the latest drug. The lost equity value was 3.5 times the entire cash price of Avidity, while del-desiran is only one of three acquired clinical programs. Novartis still produced $8.9 billion of reported free cash flow in the first half and has five growth medicines absorbing much of the Entresto patent cliff. At $137.70, a conservative owner-cash model implies only about 1.5% annual growth over the next decade.

That makes the 13.9% fall look overdone if it is read as a charge against the failed programs alone. It is less obviously excessive as a verdict on pipeline replacement, acquisition discipline and the premium previously attached to both. The evidence separates those two judgements.

Three setbacks turned a trial miss into a credibility event

HARBOR enrolled about 1,300 adults with myotonic dystrophy type 1. Participants received del-desiran or placebo every eight weeks for 54 weeks. The primary endpoint measured the time needed to open the hand after maximum voluntary grip, a direct test of myotonia. Novartis reported that the study did not show a statistically significant benefit on that measure. The trial had a favourable safety profile, but safety cannot rescue a failed efficacy endpoint. Management is reviewing the full dataset and plans to discuss the program with health authorities (Novartis HARBOR 2026; ClinicalTrials.gov 2026).

This was not an inexpensive internal experiment. Novartis completed the Avidity acquisition in February 2026 after agreeing to pay $72 per share in cash, or about $12.0 billion. Avidity's early cardiology assets were separated before completion. Novartis retained the late-stage neuromuscular portfolio: del-desiran for myotonic dystrophy type 1, del-brax for facioscapulohumeral muscular dystrophy and del-zota for Duchenne muscular dystrophy. Del-desiran was the most advanced proof that Avidity's antibody oligonucleotide conjugate platform could turn muscle delivery into an approved medicine (Novartis Avidity 2025).

The timing magnified the miss. On 4 September, pelacarsen failed to reduce major adverse cardiovascular events in the Phase III Lp(a)HORIZON study, despite lowering lipoprotein(a). Novartis stopped further development in that indication. The company had also paused eight studies of rap-cel in autoimmune disease after three deaths, according to CNBC. Those are different mechanisms, disease areas and safety questions. They do not establish that every Novartis platform is broken. In eight days, however, they removed three pieces of evidence that had supported confidence in the post-Entresto pipeline (Novartis pelacarsen 2026; CNBC 2026).

There is counter-evidence in the same week. Two Phase III REMODEL studies of remibrutinib in relapsing multiple sclerosis met their primary endpoints against teriflunomide. Novartis also reported a favourable safety profile and intends to seek approvals globally. That positive result matters because it shows the research engine did not fail uniformly. It received far less weight from the market than the three disappointments (Novartis remibrutinib 2026).

Avidity moved from pipeline asset to balance-sheet fact

The interim accounts show exactly how the acquisition changed Novartis. First-half investing cash outflow reached $14.8 billion, including $11.7 billion for business acquisitions. The Avidity cash price was $12.0 billion, partly offset by $371 million of cash acquired. Intangible assets rose by $12.7 billion and goodwill by $1.2 billion, mainly because of Avidity and two smaller transactions (Novartis H1 2026).

Funding added another layer. Novartis entered an $11.0 billion bridge loan in February and then issued $10.9 billion of US-dollar bonds plus $2.0 billion of euro-denominated bonds during the half. Total financial debt, including derivatives, reached $47.1 billion at 30 June. Liquidity was $7.7 billion, leaving reported net debt of $39.4 billion, up from $21.9 billion six months earlier. The debt-to-equity ratio rose from 0.72 to 1.12. Second-quarter interest expense increased from $289 million to $462 million (Novartis H1 2026).

The HARBOR result does not mechanically erase $12.0 billion. Del-brax and del-zota remain. Del-zota was submitted to the FDA for accelerated approval in Duchenne muscular dystrophy, and Novartis reported positive Phase I/II biomarker data for del-brax before advancing the program. Del-desiran may still have informative secondary endpoints or responsive subgroups, though neither can be assumed before the full dataset is released.

The accounting asymmetry is still uncomfortable. The cash and debt are fixed. The acquired asset values depend on trial success. A material impairment would not reverse the cash outflow or retire a bond; it would acknowledge that part of the expected future economics did not arrive. The cleanest monitoring line is therefore not whether Novartis records any impairment, but whether cumulative Avidity-related impairment exceeds $3 billion or a second acquired program is discontinued. Either event would show that HARBOR was not isolated.

The commercial portfolio can absorb a failed drug, but not an empty replacement cycle

Novartis is now a focused innovative-medicines company following the Sandoz separation. It concentrates on cardiovascular, renal and metabolic disease, immunology, neuroscience and oncology. Chemistry and biologics remain the established development platforms; gene and cell therapy, radioligand therapy and xRNA carry more of the renewal burden. The business makes money by converting protected medicines into global volume, then replacing them before exclusivity expires.

The second quarter captures that cycle in one line. Volume added 18 percentage points to sales growth. Generic competition removed 14 points, pricing removed another three, and currency added two. Net sales reached $14.4 billion, only 1% higher in constant currency (Novartis H1 2026).

The growth group is substantial. In the first half, Kisqali sales rose 48% in constant currency to $3.2 billion, Kesimpta rose 29% to $2.6 billion, Pluvicto rose 55% to $1.3 billion, Scemblix rose 85% to $995 million and Leqvio rose 64% to $932 million. Together, those five medicines added billions of dollars of annualised revenue momentum. They are the commercial moat in numbers: differentiated products, regulatory approvals, specialist sales infrastructure and manufacturing that can support several launches at once.

The expiry side is just as visible. Entresto fell 48% in constant currency to $2.5 billion in the half. Promacta fell 66% to $363 million and Tasigna fell 59% to $297 million. US sales declined 9% even though group sales rose 1% in dollars. Novartis reaffirmed FY2026 guidance for low-single-digit sales growth and a low-single-digit decline in core operating income before the September setbacks (Novartis H1 2026).

This is why del-desiran matters without being a current earnings event. The failed study does not remove a dollar from 2026 sales. It narrows the set of products available to replace erosion later in the decade. Barclays analysts cited by CNBC had assigned about $5 billion of combined risk-adjusted peak-sales opportunity to del-desiran and pelacarsen. That estimate is not a filed company forecast, but it explains why the market connected two unrelated failures to the 5%-6% average sales-growth ambition through 2030 (CNBC 2026).

Four years of improvement preceded the acquisition shock

The filed history is stronger than the share-price move suggests. For comparability, the table uses continuing-operations figures restated after the Sandoz separation. Sales rose from $42.2 billion in 2022 to $54.5 billion in 2025. Operating income more than doubled to $17.6 billion, while operating cash flow rose to $19.1 billion. R&D remained between $9.2 billion and $11.4 billion a year. These figures are reported; ROIC and leverage are author calculations (Novartis 2022; Novartis 2023; Novartis 2024; Novartis 2025).

USD millions except ratios FY2022 FY2023 FY2024 FY2025 H1 2026
Net sales 42,206 45,440 50,317 54,532 27,521
Operating income 7,946 9,769 14,544 17,644 8,985
Net income, continuing operations 6,049 8,572 11,939 13,967 6,413
R&D expense 9,172 11,371 10,022 11,200 5,587
Operating cash flow, continuing operations 13,039 14,220 17,619 19,144 9,558
Property, plant and equipment purchases 916 1,060 1,366 1,548 not shown
Net debt 7,245 10,183 16,141 21,947 39,400
Author-computed ROIC 8.6% 14.3% 22.9% 24.9% 19.6% annualised
Net debt / annual or annualised OCF 0.56x 0.72x 0.92x 1.15x 2.06x

ROIC uses after-tax operating income divided by average invested capital, defined as equity plus financial debt less cash. The effective tax rate is based on reported income before tax and income-tax expense. That produces NOPAT of $6.7 billion in 2022, $9.2 billion in 2023, $12.7 billion in 2024 and $15.1 billion in 2025. Average invested capital was $77.7 billion, $64.2 billion, $55.7 billion and $60.6 billion, respectively. The H1 2026 figure annualises six-month NOPAT and uses average capital through June.

The rising return measure flatters the operating improvement. Sandoz left the group in 2023, buybacks reduced equity, and acquisition timing shifts the denominator. The H1 2026 decline to an annualised 19.6% is more informative than the peak: Avidity expanded invested capital before contributing revenue. Incremental ROIC is not shown because the Sandoz separation and acquisition make year-to-year changes in book capital a poor proxy for organic reinvestment.

R&D scale is not a moat by itself. Amgen spent $7.3 billion on R&D in 2025, up from $4.8 billion in 2023, and is developing olpasiran against the same Lp(a) biology tested by pelacarsen (Amgen 2025). Novartis spent $11.2 billion in 2025. The peer comparison shows both the depth of competition and the limits of spending as evidence. A research budget earns its return only when clinical endpoints, approvals and commercial adoption follow.

Owner cash remained large while capital allocation became less forgiving

The owner-cash bridge begins with reported cash, not core earnings. FY2025 continuing-operations cash flow was $19.144 billion. Subtracting $1.548 billion of property, plant and equipment purchases gives $17.596 billion of author-computed free cash flow. Novartis itself reported $8.891 billion of free cash flow for H1 2026, down 9%; annualising that amount gives $17.782 billion. Both measures include R&D because research expense already passes through the income statement and operating cash flow (Novartis 2025; Novartis H1 2026).

A stricter owner measure also deducts equity compensation. H1 equity-based compensation added roughly $0.6 billion to equity, implying about $1.2 billion annualised. Deducting that from annualised reported free cash flow leaves about $16.6 billion. This treatment is conservative because Novartis repurchases shares, but repurchases are a use of cash and do not make compensation free.

The bridge does not deduct the entire Avidity purchase as annual maintenance spending. Doing so would turn a discrete transaction into a recurring charge. It does expose the capital-allocation issue separately. In H1, Novartis spent about $14.4 billion on Avidity and two smaller acquisitions, paid a $9.1 billion gross dividend and used $3.1 billion for net treasury-share purchases. Operating cash flow was $9.6 billion. Debt funded the gap.

That sequence matters more after HARBOR. A strong balance sheet allowed Novartis to acquire a late-stage platform and continue distributions. The resulting net debt remains supportable relative to cash generation, at roughly 2.1 times annualised first-half operating cash flow. Yet the cushion is narrower, and the acquisition's first Phase III readout failed before the new debt had been reduced. A FY2026 net-debt figure above $40 billion, or failure to move below $35 billion by H1 2027, would show that cash is not repairing the balance sheet as quickly as the valuation assumes.

The market erased more than the failed assets

At $137.70 and 1.908 billion ADR-equivalent shares, Novartis's market capitalisation was $262,752.296 million, or $262.8 billion. Adding $39.4 billion of June net debt gives an enterprise value of about $302.2 billion. The equity traded at 14.8 times annualised H1 reported free cash flow and 15.8 times the stricter $16.6 billion owner-cash estimate. Enterprise value was about 17.0 times annualised reported free cash flow.

Those are not distressed multiples. They are far below a valuation that requires every pipeline program to work. The 10-year US Treasury yield was 4.78% on 4 September, so an 8.5% equity discount rate and 2.5% terminal growth rate provide a demanding but not extreme framework (FRED 2026). Starting from $16.6 billion of owner cash, that model needs only about 1.5% annual cash growth for ten years to reproduce the $137.70 close. Zero growth gives about $123 per share; 2% gives $143; 4% gives $167; 6% gives $194. These are author calculations, not company outlook.

A simpler sensitivity reaches the same conclusion:

Annual owner cash 14x 16x 18x
$15 billion $110 $126 $142
$17 billion $125 $143 $160
$19 billion $139 $159 $179

The post-move price therefore assumes either limited cash growth, a lower valuation multiple, or some combination. The market removed $42.5 billion in one session. That is $30.5 billion more than Novartis paid for all of Avidity, and del-desiran is not the whole acquired portfolio. Even assigning no value to del-desiran and pelacarsen does not explain the full loss without a second adjustment for acquisition credibility, the remaining pipeline or Novartis's prior sector premium.

This is the strongest case that the reaction overshot the direct evidence. It is not proof that the old price was sound. Before the fall, the equity value was about $305.3 billion, equal to 18.4 times the strict owner-cash estimate. That higher multiple assumed the growth brands and pipeline could bridge patent expiries with less capital-allocation friction. Three setbacks damaged that assumption.

Four cases separate one failed trial from a broken renewal model

The severe downside case values annual owner cash of $12.5-$14.0 billion at 13-14 times, or $82-$103 per share. It assumes patent losses outrun launch growth, Avidity suffers a large impairment and the company keeps spending heavily to refill the pipeline. This case requires more than HARBOR. It needs commercial deterioration and another acquired-program failure.

The bear case uses $14.5-$15.5 billion of owner cash and 14.5-15.5 times, producing $110-$126. Here the current portfolio remains profitable, but external deals carry most of the replacement burden and debt reduction competes with distributions. It treats the September sequence as evidence of a persistent renewal problem.

The base case uses $16.5-$18.0 billion and 16-17 times, producing $138-$160. Kisqali, Kesimpta, Pluvicto, Scemblix and Leqvio offset most generic erosion, while del-zota, del-brax and remibrutinib preserve enough pipeline breadth for the prior valuation framework to survive largely intact. The current price sits at the bottom of this range rather than its centre.

The bull case uses $19-$21 billion of owner cash and 18.5-20 times, producing $184-$220. It requires sustained growth from the five commercial medicines, successful approvals for remaining late-stage assets and visible net-debt repair. That case cannot be inferred from one positive remibrutinib announcement. It needs several years of cash conversion and clinical delivery.

The two variables carrying most of the valuation are the sustainable owner-cash base after the Entresto cliff and the multiple warranted by pipeline reliability. A single drug can change both when it was acquired as proof of a platform. That is why HARBOR removed more value than its near-term revenue contribution, while still struggling to account for the entire $42.5 billion loss.

The anti-thesis starts with correlated execution risk

The over-reaction reading can fail in three ways. First, the setbacks may be less independent than they appear. Del-desiran, pelacarsen and several future assets rely on complex modalities, hard clinical endpoints or narrow patient groups. Large R&D spending and multiple platforms can create diversification on paper while development risk remains correlated through trial design, manufacturing and management's asset-selection process.

Second, the commercial bridge may be weaker than headline growth rates imply. H1 volume added 15 percentage points, but generic competition removed 14. Entresto's decline has moved quickly, and US sales fell 9%. If the growth medicines slow before the erosion base becomes smaller, the $16.6 billion owner-cash starting point will prove too high.

Third, acquisition economics can deteriorate before accounting catches up. Avidity added intangible assets and debt immediately. The remaining programs still carry scientific value, but their cash flows are distant. Del-brax Phase III data are expected in 2028. Del-zota's accelerated path depends on regulatory acceptance and later confirmation. The longer those cash flows take, the more a 4.78% Treasury yield and acquisition-funded interest expense reduce their present value.

The counterweight is the existing business. Five products are growing at 29%-85% constant-currency rates, annual R&D exceeds $11 billion, and reported free cash flow remained close to $18 billion annualised during the patent transition. The equity is not dependent on del-desiran for survival. The debate is about the rate and cost of renewal.

The next two results will test cash before science catches up

The first crux resolves through ordinary reporting. Q3 and FY2026 results will show whether the five growth medicines continue to add more revenue than Entresto, Promacta and Tasigna lose. Two consecutive quarters in which that relationship reverses would convert a pipeline concern into a current earnings problem. Trailing free cash flow below $16 billion would also invalidate the owner-cash base used here.

The second crux is the full HARBOR analysis. A credible subgroup, supportive secondary endpoints or a viable regulatory discussion could retain some del-desiran value. A clean discontinuation would make an impairment more likely. Until Novartis publishes those details, the primary endpoint miss deserves full weight and speculative rescue value deserves little.

The third crux sits inside Avidity's other programs. Del-zota's FDA review is the nearer test. Del-brax is the longer one, with Phase III evidence expected in 2028. Cumulative Avidity impairment above $3 billion, or the loss of a second acquired program, would argue that the acquisition platform was mispriced rather than merely unlucky on its first late-stage result.

Verification is partial because the Finance API sidecar resolved Novartis's identity and filings but had not ingested the 8 September close at the run cutoff. It showed a stale latest close of $158.86 and previous close of $156.24. The article therefore uses the fetched Nasdaq session record: $137.70, down $22.29 from $159.99, or 13.9321%, with a $262.752 billion market capitalisation. The arithmetic reconciles exactly to 1.908150 billion shares. Primary financial figures were checked against the 20-F and H1 filing; author calculations are labelled.

The 13.9% fall was larger than the direct asset loss supported by public evidence. At the same time, Novartis had paid cash, raised debt and attached its medium-term growth case to pipeline renewal before three clinical disappointments arrived in eight days. The close now prices modest owner-cash growth and asks the remaining Avidity programs to prove that HARBOR was one failed trial, not the first audit of a $12 billion mistake.

Source notes

Confidence is high in the identity, trial outcomes, acquisition terms and filed financial history because they were checked against SEC filings, Novartis releases and the ClinicalTrials.gov record. Confidence is lower in the allocation of the $42.5 billion market-value loss among individual drugs and management credibility; no public document provides that split. Novartis has not published the complete HARBOR dataset, secondary-endpoint results, subgroup analyses or a revised carrying value for the Avidity assets. Those omissions limit any estimate of retained del-desiran value. The Finance API market series lagged the covered session, so the price, previous close, signed move and market capitalisation come from the fetched Nasdaq session record instead.

References

  • SEC 2026: SEC company submissions record for Novartis AG and its NVS filings.
  • Novartis HARBOR 2026: 8 September Form 6-K announcing the Phase III HARBOR primary-endpoint miss.
  • CNBC 2026: Independent account of the del-desiran, pelacarsen and rap-cel setbacks and analyst context.
  • Nasdaq 2026: 8 September NVS close, previous close, move, market capitalisation and share-count reconciliation.
  • Novartis pelacarsen 2026: 4 September Form 6-K announcing the Lp(a)HORIZON outcome.
  • Novartis remibrutinib 2026: 1 September Form 6-K reporting positive REMODEL-1 and REMODEL-2 results.
  • Novartis H1 2026: Condensed interim accounts, product trends, cash flow, debt and Avidity purchase accounting.
  • Novartis 2025: FY2025 Form 20-F and comparative financial statements.
  • Novartis 2024: FY2024 Form 20-F and comparative financial statements.
  • Novartis 2023: FY2023 Form 20-F, including Sandoz-separated comparatives.
  • Novartis 2022: FY2022 Form 20-F and year-end balance-sheet figures.
  • Novartis Avidity 2025: Acquisition terms and the three-program neuromuscular portfolio.
  • ClinicalTrials.gov 2026: HARBOR study design, enrolment and endpoint record.
  • Amgen 2025: FY2025 Form 10-K for peer R&D spending and olpasiran context.
  • FRED 2026: Federal Reserve 10-year Treasury yield series used in the discount-rate context.