Novo Nordisk

A long ownership horizon does not shorten a patient's horizon

Novo Nordisk is a listed pharmaceutical company with a controlling owner that cannot trade control like an ordinary shareholder. Through Novo Holdings, the Novo Nordisk Foundation held all A shares and enough B shares to command 77.28% of votes as of 3 February 2026. The Foundation's A shares cannot be sold while the Foundation exists, and its statutes require it to preserve material influence over Novo Nordisk. Most economic ownership remains outside that control chain.¹

The arrangement can hold research capacity, manufacturing knowledge, and an institutional purpose across product cycles and market shocks. It cannot by itself establish that a medicine works, that promotion communicates its risks, that a patient can afford it, that a worker can contest a restructuring, or that production burdens are borne fairly. Durability is a property of the control mechanism, not proof of the mechanism's consequences.

A patient who needs insulin or an obesity treatment faces a shorter horizon than the owner. The practical questions are immediate: does the medicine help someone with my condition, can I obtain it, what risks must I understand, and who can change the terms? Foundation control matters because it locates authority over a company with unusually consequential products. Patient benefit and access still require their own evidence.

Rival insulin institutions became a three-layer control chain

The lineage began with a licensed scientific opportunity rather than one continuous company. August Krogh returned from Toronto in 1922 with permission to manufacture insulin in the Nordic countries. He joined H.C. Hagedorn and August Kongsted to establish Nordisk Insulinlaboratorium in 1923, placing the enterprise in a foundation arrangement intended to direct profits toward scientific and humanitarian purposes. A 1924 dispute sent Thorvald and Harald Pedersen away to build the rival Novo branch; Novo created its own foundation in 1951 and listed its operating company in 1974 while retaining foundation control.²

The rivalry mattered. The Foundation's history describes different scientific, manufacturing, and cultural traditions rather than a single founder's design. In January 1989 the two foundations announced their merger in what the same account calls the equivalent of an internal “palace revolution”; the operating companies then became Novo Nordisk. In 1999 the Foundation created Novo Holdings and transferred its operating-company shares to it, followed by the 2000 demerger that separated Novo Nordisk from the enzyme business now called Novonesis.² Because this history is written by the Foundation, it is useful for dates, formal lineage, and the Foundation's interpretation—not independent proof of motives, inevitability, or social effects.

The contemporary chain separates three legal roles. Novo Nordisk discovers, develops, manufactures, and sells medicines. Novo Holdings holds and votes the controlling shares. The Foundation owns Holdings, sets the ownership frame, supervises major Holdings decisions, and awards grants. The operating and holding-company boards remain formally responsible for their own operational and strategic matters.³

The top layer is not a public or beneficiary membership organization. Seven of the Foundation's ten directors are elected by the existing board for one-year terms; three are employee-elected. The board sets Foundation strategy, awards or frames grants, supervises Novo Holdings, and elects its own chair and vice chair.³ Danish public authorities supervise legality, but no patient, grant-recipient, public-shareholder, or community constituency appoints those seven seats. Employee representation creates one channel of voice without turning the Foundation into a general assembly of affected groups.

Foundation form changes control, not the burden of proof

An industrial foundation can refuse a takeover, maintain a voting block while public investors supply capital, and retain a corporate purpose after founders and executives leave. Those are real institutional differences. They create the possibility of patient investment and a longer planning horizon; they do not demonstrate that patience caused a particular discovery, that the best projects were selected, or that commercial and philanthropic returns reached the people invoked by the purpose.

Henry Hansmann and Steen Thomsen studied 110 Danish foundation-owned firms and found a robust association between their foundation-governance index and firm performance. Their wider review reported broadly comparable average profitability between Danish foundation-owned and conventionally owned firms. It also identified selection, capital-constraint, and company-size qualifications. The authors explicitly warn that endogeneity, reverse causation, omitted management quality, and other unobserved variables prevent a simple causal reading.⁴

That study supplies a useful check in both directions. Self-appointing boards do not necessarily produce the inefficiency predicted by a simple agency model; neither does an observed association prove that foundation ownership caused Novo Nordisk's research performance, access policies, or grant outcomes. The study is sample-wide and observational, not a Novo Nordisk case evaluation. It also does not test whether patients, workers, communities, or nonhuman life have effective standing inside foundation governance.⁴

The accountability trade is therefore concrete. Control is insulated from the market for corporate control, but the Foundation board is also insulated from ordinary public-shareholder replacement. Its legal and governance disclosures identify supervision, self-election, employee seats, audits, committees, and formal divisions of responsibility.¹³ These mechanisms can discipline power without answering every affected group's question: who may introduce evidence, demand an explanation, appeal a decision, or change the decision maker?

In 2025, the controlling vote settled a board dispute

The control chain became observable when the operating-company board and the Foundation disagreed. At Novo Nordisk's November 2025 extraordinary general meeting, outgoing chair Helge Lund said the parties could not reach a common understanding about governance principles and the future board. Foundation chair Lars Rebien Sørensen said the Foundation had taken the unusual step of seeking greater representation and faster chief-executive succession because it judged that increased competition and slower growth required more urgency. The Foundation wanted a larger board change than the outgoing directors did.⁵

The disagreement did not require persuasion of a dispersed voting majority. Lund said the outgoing board convened the meeting after considering the Foundation's position and its majority of votes. Novo Holdings proposed the new slate, and Sørensen—already Foundation chair and a former Novo Nordisk chief executive—was elected operating-company chair. The four employee-elected Novo Nordisk directors remained. Sørensen said the dual-chair role should be phased out, while declining to guarantee that it would end within 18 months.⁵

The supplementary vote table makes the distribution of acceptance unusually clear. Excluding Novo Holdings, 38.70% of valid votes supported Sørensen, 61.30% abstained, and none voted against. Including the controlling holder, 93.25% supported him.⁵ Abstention is not opposition, but it is also not an independent mandate. The result shows how voting control can settle a governance conflict while leaving economic ownership and acceptance distributed differently.

The Foundation presented its intervention as long-horizon stewardship; the outgoing board described an irreconcilable governance disagreement. Both are participant accounts preserved in the same official record. The episode proves that the Foundation could act decisively, not that the larger change was better for patients, research, workers, public shareholders, or future performance.

The personnel consequences extended beyond directors. Novo Nordisk began a company-wide transformation on 10 September 2025 and reported that around 9,000 employees were let go globally. At the extraordinary meeting, Sørensen said the Foundation supported the transformation and described the layoffs as a way to free resources for other areas, including new research projects.¹⁵ That states management's and the controlling owner's rationale. It does not establish the restructuring's comparative necessity or effects.

Benefit, risk, and promotion must be measured separately

In March 2024, the U.S. Food and Drug Administration approved Wegovy to reduce cardiovascular death, heart attack, and stroke in adults who already had cardiovascular disease and obesity or overweight. In the randomized trial summarized by FDA, more than 17,600 participants received Wegovy or placebo in addition to standard care; major adverse cardiovascular events occurred in 6.5% of the Wegovy group and 8.0% of the placebo group. The same regulatory summary lists a boxed warning and other serious warnings and precautions.⁶

That is strong evidence for a specified indication and population, not a claim of benefit for every user. It does not establish comparative value for every clinical alternative, affordability, long-term adherence, or real-world reach. The benefits and risks must remain together in clinical judgment without being collapsed into either a success story or a rejection of the medicine.

The informational record is not uniformly favorable. In September 2025, FDA sent Novo Nordisk Inc. a warning letter about a 2024 direct-to-consumer video featuring two paid consultants and a Novo Nordisk sales executive. FDA determined that the communication was false or misleading because it omitted or minimized serious risks for Wegovy, Ozempic, and Victoza despite presenting their benefits.⁷ The letter is an official enforcement position about one communication under U.S. drug law, not a final judicial adjudication or evidence that every Novo Nordisk promotion had the same defect.

The two FDA records belong in one account. A medicine can deliver a measured benefit while its promotion inadequately communicates risk. Foundation control does not resolve that tension. Regulatory review, accurate labeling, independent clinical judgment, post-market evidence, and enforceable promotion rules remain separate checks on organizational claims.

Access is made by a system, not by ownership form alone

Patient access exposes another separation between capability and consequence. Michael Fang and Elizabeth Selvin analyzed the 2021 National Health Interview Survey and found that 93 of 495 surveyed U.S. insulin users younger than 65—an estimated 20.4%, with a 95% confidence interval of 16.2% to 25.3%—reported skipping, reducing, or delaying insulin because of cost.⁸ The study used self-reports, had a 50.9% survey response rate, and produced imprecise subgroup estimates. It was not manufacturer-specific and preceded Novo Nordisk's 2024 list-price changes, so it establishes a national access problem rather than Novo Nordisk's share of it.

Novo Nordisk announced in March 2023 that several U.S. insulin list prices would fall in January 2024. The NovoLog and NovoLog Mix 70/30 vial list price was to fall 75%, from $289.36 to $72.34, while specified Levemir and Novolin products were to fall 65%. Associated Press reporting situated the announcement after a rival's cuts and amid public pressure, and a clinician cautioned that list price was one layer of affordability because insurance design and assistance-program eligibility still matter.⁹ The report documents an announced price change and competing interpretations at the time; it is not a post-change outcome evaluation.

Public bargaining later changed another layer. For Medicare's 2027 price year, the Centers for Medicare & Medicaid Services and Novo Nordisk agreed to a $274 30-day negotiated price across Ozempic, Rybelsus, and Wegovy, compared with a 2024 list-price benchmark of $959—a stated 71% discount. CMS reported that 2.282 million Part D enrollees used the products in 2024 and that its process considered manufacturer submissions, patient roundtables, clinical evidence, production and distribution costs, research and development, patents, and therapeutic alternatives.¹⁰

The negotiated price begins on 1 January 2027 and applies within Medicare under program rules; it is not evidence of realized savings yet and does not cover every insured, uninsured, or non-U.S. patient. Together, the rationing survey, list-price cut, and Medicare negotiation show access being made through a chain of company decisions, intermediaries, insurance terms, law, public purchasing, clinical infrastructure, and household means. No reviewed record attributes either price change to a Foundation directive.⁸⁹¹⁰

The same surplus crosses several ledgers

Commercial success created large financial flows. Novo Nordisk reported 2025 sales of DKK 309.064 billion, net profit of DKK 102.434 billion, total dividends of DKK 51.975 billion, and share repurchases of DKK 1.388 billion.¹¹ These audited or statutory company figures establish financial scale and shareholder distribution. They do not allocate the value produced or the costs borne among patients, public payers, researchers, workers, suppliers, communities, animals, and ecosystems.

The workforce ledger moved in the opposite direction during the 2025 transformation. Novo Nordisk reported 69,505 employees at year end, 10% fewer than in 2024, with 57% outside Denmark.¹ Pharmadanmark's later interview with Désirée Asgreen, chair of an employee club then representing about 8,700 leaders and specialists, records months of uncertainty, confidential review of layoff lists, challenges to some individual selections, support during dismissal meetings, and a workplace that remained unsettled.¹² It preserves a worker representative's direct experience and some negotiated process, but it is not a representative survey of dismissed, retained, international, temporary, or supplier workers.

The partner ledger is thinner. Novo Nordisk's value-creation account reports eight obesity-and-diabetes research partnerships or licences and more than DKK 60 billion invested mainly in production capacity during 2025.¹¹ Those figures show that innovation and scale are not produced inside one corporate boundary. They do not disclose partner economics, supplier concentration, contract allocation of risk, indirect labor conditions, or whether local capacity expansion benefited host communities. Claims about partner or community welfare therefore remain limited.

At the Foundation layer, the 2025 annual report records 2,206 grants, DKK 10.9 billion in grant giving, and another DKK 0.9 billion in philanthropic investments. DKK 3.1 billion went through open-competition programs with a reported 16% success rate. The report describes health, sustainability, and life-science initiatives, with Denmark remaining the center of gravity alongside international programs.¹³ These are participant-reported allocations and selection outputs, not independent measures of recipient-defined outcomes, additionality, failures, or whether Foundation control caused the results.

Material costs sit outside both the financial and grant totals. Novo Nordisk reported 2.690 million tonnes of market-based greenhouse-gas emissions in 2025, up 19% from 2024; Scope 3 accounted for 2.507 million tonnes. The company linked the increase to acquisitions, construction, procurement, and expansion and reported no absolute reduction in the year.¹⁴ This corporate inventory has a defined reporting boundary. It is not a full product life cycle and does not trace facility-level pollution, water, land, biodiversity, treatment benefits, or grant-supported ecological outcomes.

The same annual report records 46,869 animals purchased for research in 2025, 96% of them rodents, including 529 dogs and 475 nonhuman primates. Total purchases declined 5%, while use of dogs and nonhuman primates rose in connection with late-stage projects and regulatory studies.¹⁴ Counts make reliance visible but do not measure pain, study severity, lived welfare, replacement possibilities, or research performed beyond the stated boundary. Long-horizon ownership may preserve future scientific capacity; the available evidence does not produce a net intergenerational balance across medicines, institutions, access, emissions, resource use, and animal research.

Different organizational lenses prevent a single verdict

Purpose, mission, and institutional legitimacy asks whether commercial conduct makes the inherited scientific and humanitarian purpose credible in practice. Governance, stewardship, and accountability follows the control chain and the routes for review, while authority, legitimacy, and acceptance distinguishes legal voting power from acceptance by patients, workers, public shareholders, and public institutions. Structure, hierarchy, and scale separates Foundation, holding company, operating company, research partnerships, supply network, and health systems instead of treating them as one actor.

Strategy, competition, and adaptation examines the 2025 intervention and resource shift without assuming that speed proves wisdom. Decision making, judgment, and bounded rationality asks which alternatives and uncertainties each board could see; executive attention, information, and organizational sensing tracks how competition, trial results, access signals, safety information, and worker experience reach decision makers. Measurement, accounting, and control keeps votes, profit, patient reach, grant totals, clinical outcomes, emissions, and animal counts as different measures with different boundaries.

Innovation, entrepreneurship, and renewal concerns the translation of research into medicines and manufacturing capacity. Knowledge, expertise, and professional autonomy asks how scientists, clinicians, regulators, and patients can challenge commercial claims, and learning, quality, and reliability follows how trials, manufacturing controls, field evidence, and enforcement change subsequent work. Long control can preserve accumulated knowledge while layoffs, promotion incentives, or centralized intervention can also suppress or redirect it.

Coordination, communication, and common understanding traces work across laboratories, suppliers, regulators, payers, clinicians, and patients. Delegation, decentralization, and responsibility asks where responsibility lands when access or harm emerges from the chain; cooperation, incentives, and organizational equilibrium examines prices, rebates, contracts, grants, and returns. Work design, productivity, and automation brings the restructuring down to tasks and livelihoods, while culture, informal organization, trust, and voice makes employee clubs, professional challenge, and speak-up channels visible.

Organizational ignorance is not the absence of metrics. It is the distance between aggregate figures and lived effects: company patient counts do not reveal unaffordable prescriptions, an emissions inventory does not reveal a host community, and an animal count does not reveal severity or experience. Preserving those absences is part of truthful measurement, not a reason to stop inquiry.

Comparisons identify alternatives, not ancestry

The attached comparisons are editorial rather than historical genealogies. Newman's Own places an operating company wholly under a private foundation and a statutory philanthropic-business exception; Novo Nordisk instead combines foundation voting control with public economic ownership and a much larger research enterprise. Vanguard removes an outside management-company equity claimant through fund ownership but does not give dispersed fund shareholders direct operating control; it is a contrast in how economic participation and governance are separated.

Zeiss offers a closer scientific-industrial comparison because foundation ownership joins specialist production, research, employee commitments, and science funding. Its disrupted history also cautions against treating a foundation charter as self-executing. Purdue Pharma is a pharmaceutical contrast: it directs attention to the feedback among promotion, evidence, prescribing, owner reward, regulation, and community harm. The comparison does not equate the companies; it tests the claim that benevolent ownership language can substitute for independent consequence evidence.

The organizational-intelligence relation is analytical: can the institution form a truthful model, coordinate judgment, and correct itself when clinical, access, worker, or ecological evidence conflicts with strategy? Benefit for all life widens the beneficiary account beyond owners and grant totals to people needing medicine, workers, communities, nonhuman beings, ecosystems, and later generations. None of these relations claims descent, influence, or endorsement.

Important evidence remains outside the aggregate record

Available evidence establishes the formal lineage and control chain; the 2025 votes and participant rationales; a defined clinical benefit and safety record; one promotion enforcement action; national insulin-rationing evidence; specified price changes; company financial, workforce, partner, emissions, and animal-use figures; a worker representative's account; and Foundation grant allocations. It does not yet establish:

• representative patient experience across income, insurance status, race, disability, diagnosis, geography, discontinuation, and health-system capacity;
• post-2024 insulin affordability attributable to Novo Nordisk's list-price changes, or realized patient and public-payer effects of the 2027 Medicare negotiated price;
• independent grant-recipient outcomes, failed grants, additionality, conflicts, and the deliberations linking commercial returns to philanthropic allocation;
• representative job quality, voice, workload, occupational health, and mobility for dismissed, retained, international, temporary, and supplier workers;
• supplier economics and facility-level community evidence about land, water, housing, traffic, employment, tax revenue, pollution, and cumulative exposure;
• a product life cycle connecting research, production, cold chain, use, packaging, emissions, water, waste, animal-study severity, and replacement; or
• a causal comparison showing which outcomes changed because control rested with the Foundation rather than another owner.

Those gaps constrain any net verdict. Stronger study would pair post-policy utilization and out-of-pocket data with patient-controlled qualitative research; follow one medicine from discovery through payment, use, and discontinuation; add worker-, supplier-, and community-controlled evidence around one production site; and compare Foundation decisions with plausible governance alternatives instead of treating longevity as its own outcome.