I want to open with an uncomfortable idea. The company that gave the world a homegrown lung cancer pill and a famous COVID vaccine now writes more of its growth story in other people's laboratories than in its own. In 2025 AstraZeneca crossed €51.95 Bn in revenue, lifted net profit to €9.04 Bn, and posted sixteen positive Phase III readouts . Yet behind that scoreboard sits a quieter truth: a large share of its blockbusters and its boldest bets, from Enhertu to oral GLP-1, arrived through licensing, acquisition, and equity stakes rather than internal discovery alone. In this article I trace how the buying engine actually works, what it tells founders and investors, and where I believe its blind spots lie.

Chapter 1: Sixteen Blockbusters, one machine

In my reading of the FY2025 numbers, three things stand out. First, AstraZeneca is now a genuinely diversified house, with Oncology supplying roughly 44% of revenue while CVRM, Respiratory and Immunology, and Rare Disease each carry meaningful weight . Second, profitability stepped up sharply in 2025, with reported net profit margin rising to about 17% as prior year impairment charges washed out . Third, the company spends like a research business but earns its gross profit increasingly off externally sourced assets, which makes its Return on Research Capital a number worth watching closely.

A short history and the shape of the company today

AstraZeneca was created in April 1999 through the merger of Sweden's Astra AB, founded in 1913, and the United Kingdom's Zeneca Group, itself demerged from ICI in 1993 . The combination was one of the largest European mergers of its time. The next structural leaps came through biologics and rare disease: Cambridge Antibody Technology in 2006 and MedImmune in 2007 for about €13.44 Bn built the biologics arm, while the €34.49 Bn purchase of Alexion, announced in December 2020 and completed in July 2021, created the Rare Disease division now branded Alexion, AstraZeneca Rare Disease . Under Pascal Soriot, chief executive since 2012, the company survived a 2014 takeover approach from Pfizer and pivoted hard into oncology and biologics. From my point of view, the through line of this history is consistent: when AstraZeneca needed a new growth platform, it bought one. That instinct frames everything in the sections that follow. The company today targets an Ambition 2030 of about €70.76 Bn ($80 Bn) in revenue and at least twenty new medicines by the end of the decade, supported by roughly €44.22 Bn ($50 Bn) of planned United States investment and €13.27 Bn ($15 Bn) in China through 2030.

On FY2025 full year numbers, AstraZeneca reported total revenue of €51.95 Bn. Product sales were €49.15 Bn and alliance revenue, the share of partner economics, reached €2.71 Bn, up 39% and a telling signal of how much partnered product now flows through the income statement . The company ended the year with sixteen blockbuster medicines and more than one hundred ongoing Phase III studies.

Therapy area mix FY2025

To place the major brands on a growth share view, I pair each brand's FY2025 euro revenue with its reported actual growth rate as a proxy for market momentum. I treat brands growing above roughly 15% as high growth, and I treat the largest revenue brands as high relative share within AstraZeneca's book. This is a portfolio lens, not a claim about absolute market share in every indication.

Financial health across 2023, 2024 and 2025

Reading the trend, revenue grew from €42.37 Bn in 2023 to €51.95 Bn in 2025, a rise I attribute mainly to Oncology and Rare Disease volume rather than price . The single most important change sits in net profit, which jumped from €6.50 Bn in 2024 to €9.04 Bn in 2025. In my opinion that step is less about a sudden commercial surge and more about the absence of the heavy intangible asset impairment charges that depressed 2024 reported earnings, a point the company itself flags when it notes reported earnings per share benefitted from lower impairments . Gross margin held remarkably steady near 82%, which tells me the underlying product economics are stable even as the revenue base widened. There is now a richer mix of profit sharing partnered products such as Enhertu and Tezspire, where AstraZeneca books revenue but shares economics.

R&D investment and Return on Research Capital

The RoRC figures of 3.9x in 2023, 4.0x in 2024 and 3.5x in 2025 sit at the strong end of large cap pharma, where independent commentary on pharmaceutical productivity typically places returns in a broad two to four times band, with internal rates of return on late stage research often quoted in the low single digit percentages by the Deloitte pharma innovation studies . From my point of view the 2025 dip from 4.0x to 3.5x is not a productivity collapse. It reflects the denominator effect of two very heavy research years in 2024 and 2025 combined with the impairment driven volatility in reported profit. What I find more revealing is that gross profit kept compounding while the company simultaneously poured capital into business development. That tells me AstraZeneca is funding external innovation out of a genuinely productive base, not borrowing against a weak one. Whether that productivity is internal science or licensed science is exactly the question next chapter takes up.

Chapter 2: The four routes to growth

In my assessment AstraZeneca runs one of the most active external innovation programmes in big pharma, and it does so across every modality at once. It licenses-in late stage assets such as Enhertu from Daiichi Sankyo, it acquires platform companies such as Fusion and EsoBiotec, it takes large minority equity positions such as the roughly 44% stake in Cellectis, and it strikes data and artificial intelligence partnerships with Tempus, Pathos and BenevolentAI . The pattern I see is a company that treats the entire external ecosystem as an extension of its own pipeline, with a recent and unmistakable tilt toward China sourced assets and toward artificial intelligence platforms. In the analysis that follows I separate that ecosystem into its component parts, trace the milestones year by year, and then map where the money has actually gone, because I believe the structure of the spending reveals the strategy far more honestly than any mission statement.

Open innovation modalities, current state and historical evolution

I find it useful to separate AstraZeneca's external activity into four buckets, because the company genuinely uses all four.

Inbound innovation: This is the largest and most visible bucket. It includes licensing-in compounds and platforms, acquiring biotech companies outright, acquiring specific assets, and taking equity stakes. The flagship example is the 2019 agreement with Daiichi Sankyo for the HER2 antibody drug conjugate trastuzumab deruxtecan, now Enhertu, worth up to €6.10 Bn with €1.19 Bn upfront, followed in 2020 by a second Daiichi deal for datopotamab deruxtecan worth up to €5.31 Bn . On the acquisition side, the years 2023 to 2025 were unusually busy: CinCor for up to €1.59 Bn, Icosavax for about €0.97 Bn, Gracell for up to €1.06 Bn, Amolyt for up to €0.93 Bn, Fusion for up to €2.12 Bn, and EsoBiotec for up to €0.88 Bn . In my view this run shows a deliberate strategy of buying modality leadership, in radioconjugates, in cell therapy, and in protein vaccines, rather than waiting for internal programmes to mature.

Outbound innovation: This bucket is smaller but real. The clearest example is Evinova, a digital health business AstraZeneca launched in November 2023 to provide digital health solutions to the broader industry, effectively spinning internal capability outward into a commercial platform . AstraZeneca also publishes openly through its scientists, as seen in the peer reviewed characterisation of its oral GLP-1 candidate in Diabetes, Obesity and Metabolism, and it contributes to industry consortia . I would not overstate the outbound side: AstraZeneca is a net importer of innovation, not a net exporter.

Coupled innovation: The Daiichi Sankyo arrangements are the textbook case, since the companies jointly develop and commercialise the antibody drug conjugates worldwide outside Japan and split costs and profits equally . The Tempus and Pathos artificial intelligence collaboration is another, where data, model development and downstream rights are shared among three parties . The December 2025 Jacobio Pharma agreement is structured as rights for AstraZeneca outside China plus a joint China development and commercialisation arrangement, a hybrid that blends inbound licensing with coupled co development.

Platform and ecosystem innovation: Here I place AstraZeneca's venture and equity activity, its growing roster of artificial intelligence platform partnerships, and its data ecosystem. The company has built a recurring set of artificial intelligence relationships, including BenevolentAI for target discovery in kidney and lung disease, Absci for generative antibody design, and more recently Algen, Immunai and BostonGene, alongside the acquisition of Modella AI to bring multimodal foundation models in house . The strategic logic, in my reading, is a flywheel: proprietary data makes AstraZeneca an attractive partner, partners supply algorithms, and the resulting candidates generate more proprietary data.

What I find most telling about this four part structure is the imbalance between the buckets. The inbound bucket is where almost all the committed capital sits, while the outbound bucket is thin and the coupled bucket, though strategically important, tends to involve modest upfront cash relative to its long term value. In my opinion this imbalance is itself a strategic statement. AstraZeneca has decided that its comparative advantage lies in late stage development, regulatory navigation and global commercialisation rather than in originating every molecule internally. It therefore behaves like a buyer and integrator of science first, and an originator second. The Daiichi Sankyo relationship is the clearest illustration of how powerful this can be. AstraZeneca did not discover trastuzumab deruxtecan, yet through disciplined co-development and a global commercial footprint it turned a licensed Japanese molecule into one of oncology's most important medicines, with combined partner recorded Enhertu sales reaching €4.41 Bn in 2025 . That single example, in my view, explains why the company keeps returning to the external market: it has repeatedly proven that it can extract more value from a promising asset than the originator could alone.

A second pattern worth naming is the company's growing comfort with equity as a bridge between a pure licence and a full acquisition. The Cellectis structure, where AstraZeneca took roughly 44% ownership and two board seats while retaining an option to license the resulting products, lets it influence a platform's direction without consolidating it onto the balance sheet. I read this as a sophisticated middle path that gives AstraZeneca strategic optionality at a fraction of the cost of an outright purchase, while keeping the partner independent and motivated. The Syneron Bio peptide deal, where most of the value is contingent on AstraZeneca exercising options, follows the same philosophy of paying for optionality rather than certainty.

Major milestones in open innovation, 2018 to June 2026

Sourcing model, how AstraZeneca finds and integrates external innovation

When I look at where these deals originate, a clear geography of sourcing appears. The company is headquartered in Cambridge in the United Kingdom and runs major research centres in Cambridge, in Gaithersburg in the United States, and in Gothenburg in Sweden, and its published science frequently carries Gothenburg and Cambridge affiliations, as in the oral GLP-1 work . Academic and translational sourcing tends to cluster around these hubs, with the Cambridge biomedical campus and partnerships in the United States and Sweden forming the spine of its university facing research.

The more striking shift, in my opinion, is toward China. AstraZeneca has built global strategic research and development centres in Beijing and Shanghai, and a remarkable share of its 2023 to 2026 licensing came from Chinese biotech: Eccogene in Shanghai for oral GLP-1, Jacobio for Pan-KRAS, AbelZeta for GPC3 CAR-T, and CSPC for monthly weight management, the last of which followed earlier 2024 and 2025 CSPC deals worth €88m and €97m . The January 2026 commitment of €13.27 Bn in China through 2030 reinforces that this is a sourcing strategy, not a series of one off transactions.

On research type, AstraZeneca spreads its external bets across the full discovery to clinic spectrum. It buys computational and artificial intelligence platforms for early discovery, as with Modella, Tempus and BenevolentAI, it licenses clinical stage assets such as elecoglipron and baxdrostat, and it acquires manufacturing and platform capability, as with Fusion's actinium supply chain and EsoBiotec's lentiviral delivery . On governance, the structures vary by intent. Outright acquisitions fold the target into AstraZeneca entirely. Equity collaborations such as Cellectis give AstraZeneca board representation, with two seats and roughly 44% ownership, and an option to license the resulting products before investigational new drug filing, which keeps intellectual property control firmly on AstraZeneca's side . Where the company wants optionality without consolidation, it uses option based licences, as in the Syneron Bio peptide deal where most of the value is contingent on AstraZeneca exercising options.

Startup engagement and corporate venture capital

AstraZeneca does not operate a single, large, publicly branded evergreen venture fund in the way some peers do. Independent investor tracking by CB Insights records an entity described as the AstraZeneca Fund and tallies a history of dozens of investments, more than twenty acquisitions and around twenty portfolio exits, with a recent venture investment into the digital health company Huma at Series D in 2024 . Beyond that aggregate, AstraZeneca's startup engagement runs primarily through corporate development equity stakes tied to research collaborations rather than through a standalone venture vehicle disclosing fund size and cheque sizes. The Cellectis position is the cleanest disclosed example: an initial €71m equity investment plus a €22m research payment, then a further €124m equity tranche, taking AstraZeneca to about 44% ownership and two board seats, with defined milestone payments of €62m to €195m per candidate product across up to ten products.

There is, however, one formal China-domiciled fund vehicle that deserves explicit treatment: the AstraZeneca-CICC Healthcare Investment Fund (also referred to as the AstraZeneca-CICC Healthcare Industrial Fund). AstraZeneca announced it in November 2019, describing it as the first and largest scale healthcare industrial fund the company had ever established . The fund's target size was announced at $1.0 Bn. The co-manager is CICC Capital, the investment management arm of China International Capital Corporation, one of China's leading state backed investment banks, and the stated combination of CICC's capital management experience with AstraZeneca's healthcare expertise is intended to drive innovation across medicine, biotech, diagnostics, medical devices, digital health, and AI in China . According to PitchBook, the fund is classified as a 2021-vintage venture fund, and closed at approximately $313m, well below the $1.0 Bn target. AstraZeneca's own LP contribution within that $313m is not separately disclosed. The fund is listed as still active on AstraZeneca's own partnering pages, where the company continues to describe it as a way for partners to access China market entry support and AstraZeneca's global network.

By the end of 2022, press reported that the fund had completed investments in more than ten companies, including Abbisko Pharmaceutical, Sibiman Biotechnology, Tiankeya, Etuo Pharmaceutical, DeepInsight Medical, Zhizhong Medical, Claritz Medical, Rgenta and Anxuyuan, spanning biomedicine, medical devices, diagnostics, AI-enabled drug discovery, and digital health. Anxuyuan is a fourth-generation long-read gene sequencing company, and the AZ-CICC fund co-led its Round B financing of approximately $100m alongside Yunfeng Fund in June 2022. CB Insights as of mid 2025 tracks eleven disclosed investments under the AstraZeneca CICC Healthcare entity, with the most recent being a March 2025 seed round in Danatlas. Two additional 2025 investments are publicly confirmed through press releases: in September 2025 the fund led YolTech Therapeutics' $45m Series B, where YolTech is a clinical stage in vivo CRISPR gene editing company with its lead programme YOLT-201 in Phase IIa for transthyretin amyloidosis, the first such programme to enter clinical trials in China; and in December 2025 the fund co-led a financing round in Syneron Bio alongside AstraZeneca's own direct corporate investment, in the same company with which AstraZeneca had signed a $3 Bn plus macrocyclic peptide licensing deal in March 2025.

That Syneron example is precisely what makes the AZ-CICC fund analytically significant beyond its disclosed size. AstraZeneca is using it as both an early stage China scouting vehicle and a relationship deepening mechanism that can run alongside or ahead of direct BD&L engagement. Sibiman Biotechnology is an earlier instance of the same dynamic: Sibiman became a global AstraZeneca partner in cell therapy, with the fund investment explicitly enabling a tripartite cooperation between Sibiman, AstraZeneca and hospital networks in China.

From my point of view this fund is structurally distinct from AstraZeneca's direct balance sheet equity deals such as Cellectis or Huma. AstraZeneca participates as industrial limited partner rather than controlling capital allocation, and the fund operates under CICC Capital's investment framework, which is why it does not appear as a conventional CVC vehicle in most Western databases. The correct analytical framing is that the AZ-CICC fund is the earliest layer of AstraZeneca's China sourcing stack, sitting below the direct licensing deals such as Eccogene, Jacobio, AbelZeta and CSPC, and providing visibility into companies before they are large or de-risked enough to attract a direct AstraZeneca commercial agreement. It should be read alongside the 2019 Shanghai Global R&D Centre and AI Innovation Centre announcements as a three-part, simultaneous commitment to embedding AstraZeneca structurally inside China's innovation infrastructure, not as a peripheral financial investment.

Non traditional partners, hospitals, patient groups, digital health and artificial intelligence

AstraZeneca's non-traditional partnering has expanded fastest in digital health, artificial intelligence, and patient-centered engagement. On the digital and data side, the company launched Evinova in 2023 as a digital health business, partners with Tempus for de-identified oncology data and model building, and invested in Huma's remote monitoring platform. On artificial intelligence and machine learning, the roster is long and growing: BenevolentAI from 2019, Absci for generative antibody design, Verge Genomics, Immunai, Algen, BostonGene, the Tempus and Pathos foundation model collaboration, and the Modella acquisition.

On patient advocacy and registry engagement, AstraZeneca has built a more substantial infrastructure than earlier versions of this analysis recognised. The company operates the Global Hypophosphatasia Registry, an international observational registry sponsored by Alexion (AstraZeneca's Rare Disease unit) and collecting long-term epidemiology and burden of disease data. It holds membership in Vivli and commits to sharing anonymised patient level data and core clinical trial reports from AstraZeneca sponsored studies on the Vivli platform. The company has formalised clinical trial transparency commitments via astrazenecaclinicaltrials.com and registration on ClinicalTrials.gov, EU CTR and other global registries, and complies with EFPIA and ABPI transfers of value disclosure frameworks across Europe and the UK.

On patient engagement, Alexion has partnered with patient communities to embed patient perspectives directly in the biopharmaceutical development process and rare disease research pathways, a collaboration described in peer reviewed literature. Beyond registries, AstraZeneca operates the Patients Association partnership in the UK, designed around co-creation of services and initiatives with patients, and the Healthy Heart Africa programme (which has screened more than 10.8 million people for elevated blood pressure as of March 2024 and now includes chronic kidney disease). The Cancer Care Africa initiative, launched in November 2022, advocates for policy change and improved cancer screening across African countries. The Young Health Programme, co-led with Plan International and Johns Hopkins, targets prevention of non-communicable diseases in youth aged 10-24 through advocacy and awareness across multiple countries.

More recently, AstraZeneca has convened patient advocacy summits as a structured engagement mechanism. The AstraZeneca Patient Advocacy Leaders Summit 2025 in Cambridge focused on cancer care pathways and patient-driven insights. The Hematology focused Patient Advocacy Leaders Summit at the European Hematology Association Congress 2026 convened more than 35 patient advocacy leaders from multiple countries. A separate rare disease patient advocacy summit was held in April 2026 in the Middle East. The company sponsors patient advocacy forums including the BIO Patient Advocacy Changemakers Event 2025 and the Healthcare Advocate Summit 2025 in the United States. These programmes, combined with country level transfers of value disclosures, indicate that AstraZeneca's patient advocacy and registry engagement is now substantially more disclosed and formalised than it was five years ago. The primary gap remains in named hospital network data partnerships, where specific multi-year data sharing agreements of the kind some healthcare systems publicly disclose remain undisclosed at the same level of institutional detail.

From my point of view the heat map tells the real story of where AstraZeneca's open innovation centre of gravity has moved. The artificial intelligence row darkens sharply from 2023 onward and peaks in 2025, while hospital and patient facing rows remain comparatively light in public disclosure.

Chapter 3: Where the Money Went

When I aggregate committed capital across the disclosed deals from 2019 to mid 2026, three patterns jump out. First, Oncology dominates the externally sourced portfolio once the one off Alexion megadeal is set aside, taking around 60% of committed capital. Second, the capital is barbelled: a small number of very large commercial stage bets, namely the Daiichi antibody drug conjugates and Alexion, sit alongside a long tail of earlier stage acquisitions and licences. Third, the modality mix has shifted from licensing in the late 2010s to outright acquisition in the early 2020s and, most recently, toward China sourced licences and artificial intelligence collaborations. I define committed capital throughout as upfront payments plus disclosed maximum milestones.

AstraZeneca Committed Capital by Therapy Area

Across the disclosed set, total committed capital reaches roughly €65.1 Bn, but that figure is heavily skewed by the €34.5 Bn Alexion acquisition. Including Alexion, Rare Disease takes about 54% of committed capital. The more useful lens, in my opinion, strips out that single structural deal to reveal the ongoing bolt on strategy, where Oncology leads decisively.

The reading I take from this is that AstraZeneca uses two different external playbooks. Rare Disease was entered through one transformational acquisition and has since been topped up only modestly, as with Amolyt. Oncology, by contrast, is fed by a continuous stream of mid sized deals across antibody drug conjugates, cell therapy and radioconjugates. CVRM is the rising third pillar, driven almost entirely by the obesity and cardiometabolic licences from Eccogene and CSPC . Respiratory and Immunology and Vaccines together account for a smaller slice of committed capital, which tells me these areas are managed more through internal development and selective bolt ons than through large external bets.

I find the Ansoff lens useful for separating where AstraZeneca is deepening versus diversifying. In my view AstraZeneca's external programme is weighted toward product development, that is, bringing new modalities to the cancer and cardiometabolic patients it already understands, rather than true diversification. The exceptions, Alexion and the obesity push, are precisely the moves that reshaped the company.

I want to dwell on what this oncology concentration means in practice, because it is the single most important pattern in the externally sourced portfolio. AstraZeneca has effectively built a modality collection in cancer. Through Daiichi Sankyo it holds the leading antibody drug conjugate franchise. Through Fusion it added radioconjugates and an actinium supply chain. Through Gracell, Cellectis, Neogene and EsoBiotec it assembled a cell therapy stack spanning autologous, allogeneic and in vivo approaches . No competitor has assembled quite this breadth of next generation oncology modalities through external means in such a compressed window. In my opinion that breadth is both a strength and a concentration risk, because it means a meaningful share of AstraZeneca's future oncology growth depends on the company's ability to industrialise several distinct and still maturing technologies at once.

The CVRM story is different in character. Here the external capital is concentrated in a single thesis, obesity and cardiometabolic disease, and it arrived late relative to the market leaders. The Eccogene oral GLP-1 licence and the much larger CSPC weight management licence together represent the bulk of CVRM committed capital, and both are early stage bets on a crowded field . From my point of view this is the highest variance part of the entire external portfolio: the prize is enormous, the field is competitive, and AstraZeneca's assets are still years from the market.

External Deals by Development Stage

Two commercial or approved stage deals, Alexion and the Daiichi pair, soak up the overwhelming majority of committed capital, while the larger number of deals by count sit at Phase I and preclinical. In my interpretation AstraZeneca is willing to pay megadeal prices only for de-risked, commercial scale assets, and prefers to access novel modalities cheaply and early, accepting clinical risk in exchange for lower upfront cost.

AstraZeneca committed capital by modality

This infographic captures the strategic rotation cleanly. The late 2010s were the licensing years, anchored by Daiichi Sankyo. The year 2021 was the acquisition spike, driven entirely by Alexion. The years 2023 and 2024 spread capital across acquisitions and licences as the company bought modality platforms. And 2025 and early 2026 mark the return of licensing, this time predominantly from China and bundled with artificial intelligence collaborations. From my point of view this rotation reflects both opportunity and discipline: AstraZeneca leans on whichever instrument offers the best risk adjusted access to a given capability in a given year.

I also read the modality rotation as evidence of pricing discipline. In years when private biotech valuations were stretched, AstraZeneca leaned on structured licences and equity stakes that limited upfront cash. When valuations softened and attractive clinical stage companies became available, it shifted to outright acquisition. The recent pivot to China sourced licensing, in my opinion, is partly a valuation arbitrage: Chinese biotech assets have frequently been available at lower upfront cost than comparable Western assets, and AstraZeneca's deep operational presence in China lets it underwrite and integrate them more confidently than most peers. This is not opportunism for its own sake; it is a company matching the instrument to the market conditions of each year.

The progression chart underlines why the commercial stage bets dominate committed capital: the Daiichi antibody drug conjugates have already advanced from early clinical to approved and are now central revenue drivers, with combined Enhertu sales recorded by both partners reaching €4.41 Bn in 2025. The earlier stage cell therapy and radioconjugate assets remain in early clinical development, consistent with the lower price AstraZeneca paid to access them.

Before turning to the failures, I want to draw out one more pattern that the committed capital data makes visible: the relationship between deal size and clinical stage is remarkably consistent. The two largest commitments by far, Alexion and the Daiichi Sankyo pair, were both placed on assets that were either already approved or in late clinical development. The smallest commitments, Neogene, Cellectis and the early cell therapy plays, were placed on preclinical platforms. In my reading this is not accidental but a deliberate risk pricing rule that AstraZeneca applies with discipline. The company is willing to pay a premium only when the science is substantially de-risked, and it insists on a steep discount when it is buying optionality on unproven biology. For a founder or an investor trying to predict how AstraZeneca will value an asset, this is the single most useful heuristic I can offer: the price the company is prepared to pay scales sharply with the maturity of the clinical evidence, and far less with the theoretical size of the eventual market.

There is also a timing signal embedded in the modality and stage data that I think deserves emphasis. The acquisitions of 2023 and 2024 clustered around assets that could feed AstraZeneca's Ambition 2030 growth window, which runs to the end of the decade. Radioconjugates from Fusion, cell therapy from Gracell and EsoBiotec, and the rare disease bolt on of Amolyt all fit a portfolio that needs new growth drivers as the older oncology and CVRM franchises mature. From my point of view the external programme is therefore not a series of opportunistic purchases but a structured attempt to assemble the next generation of growth assets on a defined timeline, with the modality collection in oncology serving as the primary hedge against the patent erosion that will eventually reach Tagrisso, Calquence and Farxiga.

Assets that failed and the write offs

Open innovation is not a story of uninterrupted wins, and I think it is important to show the misses. Within the period under review, the most concrete disclosed setback tied to a partnered asset is the ceralasertib programme. In the FY2025 results AstraZeneca disclosed that the LATIFY Phase III trial of ceralasertib combined with Imfinzi in post immunotherapy non small cell lung cancer did not meet its primary endpoint. AstraZeneca's research and development expense lines across 2024 and 2025 also absorbed intangible asset impairment charges that materially depressed reported 2024 earnings, which the company explicitly cites when explaining the 2025 rebound in reported earnings per share. The precise asset level write off values for individual discontinued programmes are not always separately disclosed.

Taken together, this tells me AstraZeneca runs a high volume, oncology weighted external portfolio with a deliberate choice between a few transformational commercial deals and many early stage modality bets, and that it is honest enough in its disclosures to let an outside analyst see both the Enhertu style triumphs and the LATIFY style disappointments.

Chapter 4: Compressed timeline, concentrated bets

When I map every disclosed deal onto a grid of development stage against strategic distance from AstraZeneca's core, a clear shape appears, and so does a clear gap. The company is dense in the cells that combine oncology with early to mid stage, adjacent modalities. It is thin, in my reading, in the cells that would represent late stage, expansionary bets outside oncology, particularly a de-risked Phase III asset in the cardiovascular and metabolic expansion it is so visibly chasing. That gap is the strategic challenge: AstraZeneca has bought its way to oncology depth but is still buying its way into cardiometabolic breadth largely at early stage and high clinical risk.

Reading the grid, the upper left and upper right quadrants tell the story of AstraZeneca's past, and the lower middle tells the story of its ambitions. The Commercial or Approved row is sparse but enormous in capital: Alexion in the New Domain column and the Daiichi antibody drug conjugates in the Adjacent column account for the bulk of all committed capital. These were the bets that redefined the company, and AstraZeneca has not attempted a comparable late stage, new domain entry since 2021. In my opinion that is a deliberate pause rather than an oversight, because integrating Alexion and the Daiichi programmes consumed enormous organisational bandwidth.

The dense cells sit in the Phase I and Preclinical rows under the Adjacent column, almost entirely coloured for oncology: Gracell, Fusion, AbelZeta, Neogene, Cellectis and EsoBiotec. This is where AstraZeneca is doing its most active shopping, and it reveals the real engine of the strategy. The company is accumulating early stage, adjacent modality oncology assets, antibody drug conjugates, radioconjugates, and cell therapies, at relatively modest individual prices, betting that its development and commercial machine can carry a few of them to approval the way it carried Enhertu. From my point of view this is a sound, repeatable model, but it concentrates risk in a single therapy area and a single thesis: that AstraZeneca's internal development capability is good enough to de-risk other people's early science.

The gaps are where I would focus a strategic conversation. The Phase II or III by Expansionary cell is empty, and I have flagged it. AstraZeneca is making a very loud push into obesity and cardiometabolic disease, but every one of those bets, Eccogene's elecoglipron and the CSPC weight management portfolio, sits at Phase I or earlier . In a field where competitors already hold approved and late stage assets, that is a meaningful exposure. The company is paying licence prices for early stage cardiometabolic optionality rather than acquiring de-risked late stage assets, which means its expansion thesis depends heavily on early molecules reading out well over the next several years. The second gap, Phase II or III by New Domain, is empty because AstraZeneca has not entered a genuinely new therapeutic domain at late stage since Alexion. Whether that is prudence or a missed window depends entirely on one's view of where the next durable growth platform will come from once oncology and cardiometabolic mature.

There is also a sourcing concentration risk that the grid does not colour but the underlying deals make plain. A large and growing share of the expansionary and core stage pills, Eccogene, Jacobio, AbelZeta and CSPC, originate from China . That concentration brings genuine pipeline value at attractive prices, but it also ties a meaningful slice of AstraZeneca's future growth to one country's biotech ecosystem and to the geopolitical and regulatory weather around it. The AZ-CICC Healthcare Investment Fund, established in 2019 and still active in 2025 and 2026 with lead investments in YolTech and Danatlas, deepens this China specific sourcing infrastructure further: by the time a Chinese biotech reaches the stage of a direct AstraZeneca BD&L agreement, it may already have been tracked, funded, and validated through the AZ-CICC layer. That is a competitive advantage in access, but it is also a compounding of the single country exposure that the grid already shows at the licensing level.

I also want to test the grid against a counter argument, because I believe the strongest analysis anticipates the rebuttal. A defender of AstraZeneca's current shape would say that the empty late stage expansionary cell is a feature, not a bug. Acquiring a de-risked, late stage cardiometabolic asset today would mean paying a full, competitive price into a field where Novo Nordisk and Eli Lilly already hold approved products, and the returns on such a purchase could be thin. By licensing early, as it did with Eccogene and CSPC, AstraZeneca preserves the upside of a differentiated oral or monthly mechanism while keeping its upfront exposure modest. On that reading, the empty cell reflects price discipline rather than strategic timidity. I find this argument genuinely persuasive on cost, and I would not characterise the gap as a clear error.

Where I am less persuaded is on time. The difficulty with a portfolio of early stage expansionary bets is that they all read out on roughly the same multi-year horizon, which means the company's most important new growth narrative is concentrated in a single window of clinical risk. If elecoglipron and the CSPC portfolio disappoint in pivotal trials, AstraZeneca has no late stage fallback in that cell to cushion the miss, and the cost of acquiring one will likely have risen in the interim. The same logic applies to the dense Phase I oncology cluster: the strategy works only if the company's development engine converts a sufficient number of these early modality bets into approvals before the older franchises erode. In my opinion this is the real tension at the heart of AstraZeneca's open innovation model. It has industrialised the sourcing of early science brilliantly, but it has loaded an unusually large share of its future onto its own ability to de-risk that science on schedule. The grid does not show a company that has made a mistake. It shows a company that has made a concentrated, time bound bet on its own execution.

In my view the central strategic challenge for AstraZeneca is therefore twofold: convert its dense early stage oncology shopping into approvals before the patents on Tagrisso, Farxiga and the older franchises erode, and de-risk its cardiometabolic expansion, ideally by adding at least one later stage, lower risk asset to fill that empty expansionary cell before competitors close the window.

Chapter 5: The Heuristics of deal science

Read separately, the four preceding chapters tell four honest but partial stories. Read together, the numbers disagree with each other in ways that no single chapter surfaces on its own. The financial trend, the modality mix, the therapy area allocation and the complementarity grid were each built from the same underlying deal set, and when I overlay them against one another rather than against the narrative each chapter offered on its own, different relationships emerge that contradict or complicate the strategic story AstraZeneca and this piece have both told so far.

Oncology is overweighted 1.4x against its own revenue share, while Rare Disease is underweighted 5x

Committed external capital ex-Alexion assigns Oncology 60.4% of new spend against a 43.6% share of FY2025 revenue, a 1.4x overweight (author's calculation; not an official company reconciliation). Rare Disease sits at the opposite extreme: it generates 15.5% of FY2025 revenue, the second largest therapy area by that measure, but received only 3.0% of ex-Alexion committed capital, the entire share represented by the single Amolyt bolt on. That is roughly a 5x underweight. Respiratory and Immunology is underweighted too, at 8.8% of committed capital against 15.1% of revenue, a 0.6x ratio, while Vaccines and Immune Therapies is modestly overweighted at 1.5x on a small base.

The pattern this reveals is not diversification but compounding. If AstraZeneca were using external capital to balance a lopsided revenue base, the underweighted categories would be attracting the new money. Instead the opposite is happening: the segment already generating the largest revenue share is also absorbing the largest share of forward looking commitments, while the second largest revenue segment is being left to run almost entirely on the Alexion platform it already owns. Three years from now, on this trajectory, the revenue mix should be more oncology concentrated than it already is today, not less. Whether this is strategically envisioned is anyone's guess.

Alliance revenue could be soon a strategic pillar

Total FY2025 revenue rose 4.0% year over year, from €49.96 Bn to €51.95 Bn, a headline that reads as modest, steady growth. Alliance revenue, the partner economics AstraZeneca books from co-developed assets such as the Daiichi Sankyo antibody drug conjugates, grew 39% to €2.71 Bn over the same period. Back-solving from that disclosed growth rate implies FY2024 alliance revenue of approximately €1.95 Bn, which in turn puts FY2024 product sales at roughly €48.01 Bn against the disclosed FY2025 product sales figure of €49.15 Bn, a growth rate of about 2.4% (author's calculation; not an official company reconciliation, and subject to minor rounding introduced by currency conversion and unbroken other income lines).

The gap between those two growth rates, 39% against 2.4%, means the partnered slice of the business grew roughly sixteen times faster than the organically sold product base. This complicates the headline 4.0% revenue growth figure considerably. A reader who takes that 4.0% at face value would conclude AstraZeneca's underlying commercial engine is healthy and broadly diversified. The decomposition says something narrower: nearly all of the FY2025 increment traces to assets the company did not originate, and the organic core, the medicines AstraZeneca discovered and commercialised entirely on its own account, barely grew at all.

Acquisition's share of annual committed capital moved from 100% to zero in five years

In 2021, the Alexion year, acquisition accounted for 100% of that year's committed capital, because Alexion was the only disclosed deal. In 2024, the year of Fusion, Amolyt and Gracell, it again accounted for 100%, this time across three separate transactions rather than one. In 2025 that share collapsed to 15.9%, with licence, collaboration and equity structures taking the remaining 84.1% of that year's committed capital (author's calculation from the modality figures in Chapter 3). By the first half of 2026, acquisition's share had fallen to zero: every euro of committed capital in that window, all €4.72 Bn of it, moved through licence and asset deal structures via CSPC and AbelZeta. AstraZeneca has not fully consolidated a single external asset onto its balance sheet in 2026 to date, a first anywhere in the disclosed dataset going back to 2019. Two consecutive years of 100% acquisition share followed by a swing to zero is not a gentle drift; it is a structural exit from full ownership as the default instrument, timed precisely to the years when China sourcing and option based structures both intensified.

The China valuation arbitrage claim holds for two deals and breaks on the third

Eccogene's upfront payment of $185m against a total deal value of up to $1.8 Bn is a 10.3% upfront ratio. Jacobio's $100m upfront against up to $1.9 Bn is a 5.3% ratio. Both fit the cheap optionality thesis this piece advances elsewhere when describing China sourcing as a valuation arbitrage: pay a small amount up front, retain the option to pay far more only if the science reads out. CSPC does not fit that pattern. Its $1.2 Bn upfront against a total value of up to $3.5 Bn is a 34.3% ratio, more than three times Eccogene's and more than six times Jacobio's, and in absolute terms that upfront payment is larger than Gracell's $1.0 Bn and close to CinCor's roughly $1.3 Bn, the two most expensive Western bolt on acquisitions of 2023 and 2024 (author's calculation). This is the one place in the entire dataset where a deal's country of origin and its pricing structure directly contradict the thesis built around it. The most recent and largest China sourced deal is not priced like an early option on unproven biology. It is priced like a de-risked Western asset with a known commercial ceiling, which raises a question the rest of this piece has not asked: whether the arbitrage window that justified Eccogene and Jacobio at their much smaller upfront commitments is already closing as Chinese assets mature and competition for them intensifies.

So What?

These relationships, read together, point to one claim: AstraZeneca's capital is not diversifying the company, it is concentrating it, and the concentration is accelerating faster than the headline growth numbers suggest. Oncology absorbs 1.4 times its revenue share in new external capital while Rare Disease absorbs a fifth of its revenue share, meaning the revenue mix three years from now will be more concentrated than it is today, not less.

Alliance revenue, the clearest fingerprint of externally sourced assets, grew roughly sixteen times faster than the organic product base in 2025, which means the 4.0% headline growth figure conceals a business whose own discoveries barely moved the top line. Acquisition, the modality that consolidates an asset fully onto the balance sheet, went from 100% of annual committed capital in two separate years to zero in the first half of 2026, replaced by licence and option structures that preserve seller independence and AstraZeneca's own exit flexibility alike.

The one China deal priced like a mature Western asset, CSPC, arrived in the exact window that acquisition activity disappeared entirely, which suggests the company is now willing to pay full price for external optionality rather than full ownership of anything. None of this reads as opportunism. It reads as a single, coherent bet: that AstraZeneca's own development engine, not its balance sheet, is now the constrained resource, and that the company would rather rent access to more science through smaller, option heavy structures than own less of it outright. The RoRC dip that looked like fading productivity in Chapter 1 is, on this reading, the fingerprint of that same bet. Research and development spending growth stalled to 0.3% in 2025 in precisely the year external committed capital rose 34.5%, which means AstraZeneca is now financing a growing share of its next decade through contracts with other companies' laboratories rather than expansion of its own.

Methodology and Disclaimer

This is a personal analytical perspective on the company's external innovation strategy based exclusively on publicly available information (SEC filings or equivalent, press releases, investor disclosures) current as of June 2026. This is NOT financial, investment, legal, or strategic advice. It does not constitute a recommendation to buy, sell, or invest in any company, security, or asset. Before making any decisions, readers must consult qualified financial advisors, investment professionals, and legal counsels.

While I have cross checked sources and taken care to ensure accuracy, errors and omissions are possible. The onus of final verification lies entirely with the reader. I assume no liability for any losses, damages, or consequences resulting from reliance on this content. Drug development is inherently uncertain; all forward looking statements about pipeline progression, market potential, or strategic outcomes are subject to significant risk and may not materialise. I have no financial interest in, affiliation with, or endorsement relationship with the company or any entities mentioned herein.

Feedback, corrections, and alternative perspectives are welcome. If you would like to collaborate or contribute or even borrow some analytical piece from this post, write to info@kletthamerinsights.com.

A note on currency: All financial figures are presented in euro. Income statement items are converted from United States dollars using each year's average exchange rate (1.0813 for 2023, 1.0824 for 2024, 1.1306 for 2025), and the 2025 average rate is also applied to the multi year committed capital comparison for consistency. EBITDA figures are analytical estimates equal to reported operating profit plus estimated depreciation and amortisation and should be read as indicative. Committed capital is defined as upfront payments plus disclosed maximum milestones, and many such milestones may never be paid.

Sources used in this article: