Pharma buyers and data owners are drug developers that pay for AI-enabled discovery, supply proprietary biology or clinical data, run regulated trials, and own the final commercial drug rights. This node covers the companies where AI can matter only if it changes target selection, molecule design, biomarker work, trial design, or pipeline capital allocation enough to improve revenue, margin, cash flow, or replacement-asset durability. The current basket read is that LLY, REGN, AZN, and MRK have the clearest source-backed AI and data-owner evidence; PFE, GILD, JNJ, and INCY have real exposure, but their equity cases still depend more on portfolio resets, patent cliffs, existing franchises, and clinical proof than on AI adoption itself.
What the stack is: scaled drug developers that combine R&D budgets, proprietary chemistry and biology records, omics and clinical datasets, trial operations, regulatory teams, manufacturing, and commercial rights.
What it does: this layer buys or builds model-assisted discovery work, chooses which targets or molecules enter experiments, decides which candidates enter toxicology or human trials, and owns the label, pricing, reimbursement, and launch path if the asset works.
Main operating pieces: internal discovery labs, clinical and genomic datasets, data-rights governance, molecule and antibody design workflows, biomarker teams, clinical development, regulatory submissions, business-development option rights, manufacturing quality systems, and commercial franchises.
Where it sits: operationally downstream from compute, omics, wet labs, and TechBio vendors, but upstream of final drug revenue; the work sits inside pharma R&D, clinical operations, regulatory affairs, and portfolio-capital-allocation committees.
How the theme uses it: AI-enabled biology reaches company economics only when model output changes a pharma-owned program, partner option, clinical trial, filing, label, or commercial drug enough to improve revenue, margin, cash flow, or patent-cliff replacement.
Terms used later: proprietary data means rights-cleared internal or partner data the company can use for training, validation, or decisions; IND means an Investigational New Drug filing that permits U.S. human testing; CTA is the analogous clinical-trial authorization in many non-U.S. markets; option exercise means the buyer pays to keep or expand rights to a partner program; LOE means loss of exclusivity when patent or regulatory protection fades.
Report boundary: this node is a tactical report layer. It ranks the current value-chain basket, setup labels, confirmation triggers, invalidation levels, and chart provenance. Durable research, claim IDs, raw source registries, and sector thesis maintenance stay in the linked knowledge pages. Price and volume context comes from read-only discovery daily_ohlc through 2026-06-12.
Current Setup
Buyer-owned validation gateAI matters after owned pipeline decisions change.
The node pays when proprietary data, model output, wet-lab evidence, and clinical execution become a named candidate, option exercise, filing, readout, or commercial asset inside a pharma P&L.
Positive proofLLY, REGN, AZN, and MRK have the clearest data-owner evidence.
Look for named programs, milestones, IND/CTA progress, trial starts, or explicit portfolio-prioritization language.
Conversion gateClinical and commercial ownership decides economics.
AI-linked work must survive toxicology, human efficacy, regulatory review, pricing, and launch execution.
Primary constraintMost ticker debates still sit outside AI.
GLP-1 pricing, EYLEA transition, Keytruda LOE, Pfizer reset, HIV durability, STELARA erosion, and Jakafi replacement remain live.
InputRights-cleared data
WorkflowModel + wet-lab decision
GateIND, trial, option, filing
EconomicsRevenue, margin, cash flow
AI-enabled biology needs customers that own proprietary data, clinical execution, regulatory functions, and commercial drug rights. Platform vendors can show model progress, but pharma captures final economics only when AI-supported work becomes better portfolio choices, accepted discovery packages, licensed assets, trial starts, approvals, label expansion, or commercial drugs. The hard evidence is a named program, a paid milestone, an IND or CTA, a clinical readout, a partner option exercise, or a disclosed R&D productivity metric.
Recent source checks show real adoption across the basket. Lilly has TuneLab, a NVIDIA co-innovation lab, and external AI-discovery deals; Regeneron has the Regeneron Genetics Center and EHR-linked data expansion; AstraZeneca has oncology, rare-disease, Evinova, and partner AI evidence; Merck has Mayo Clinic and broader AI/ML discovery work; Pfizer has CytoReason, Boltz, and Chai-related model access; Gilead has Tempus oncology data and AI-enabled discovery partnerships; Johnson & Johnson has Isomorphic and structural-data collaboration evidence; Incyte has a Genesis Molecular AI expansion that uses proprietary experimental data. The next positive proof is an AI-linked candidate reaching a filing, first-human dosing, Phase 2 signal, partner option, milestone receipt, or explicit pipeline-prioritization disclosure.
Most stock-level evidence tests are still outside AI. Lilly depends on GLP-1 net price, supply, capacity, and valuation; Regeneron depends on EYLEA replacement, Dupixent economics, and pipeline scale; AstraZeneca depends on Phase 3 readouts, policy, China, and cash conversion; Merck depends on the Keytruda bridge; Pfizer depends on LOE, debt, dividend coverage, and pipeline credibility; Gilead depends on HIV durability and acquired oncology; Johnson & Johnson depends on Innovative Medicine replacement growth, MedTech margins, and litigation; Incyte depends on post-Jakafi replacement assets. Watch whether AI spend remains a small R&D input or becomes visible in named assets, trial quality, cash flow, and guidance.
Static setup labels and chart thresholds were originally derived from weekly bars through 2026-06-05. Read-only discovery now shows daily_ohlc rows through 2026-06-12, so the old static thresholds are stale and should be refreshed before use as current trading evidence. Fundamental claims route to the AI Biology Platform Validation concept, Health Care sector lane, and linked security lanes.
Basket
This basket is hand-curated from the parent theme, local knowledge coverage, bounded web checks, and read-only discovery coverage. Ranking favors proprietary data ownership and source-backed AI/R&D exposure first, financial materiality and pipeline conversion second, and technical timing third. Roche, Sanofi, Novartis, Novo Nordisk, Bristol Myers, and GSK remain source-only because local security-lane and chart routing are incomplete for this node.
Lilly combines the largest operating momentum in the basket with proprietary data, TuneLab, and NVIDIA lab evidence.
Market cap$845.8B
Next earningsNot confirmed
Latest qtr revenue$19.799B
Role in stack
Lilly is a final drug-rights owner and AI buyer. Its proprietary discovery, safety, clinical, and commercial datasets can feed TuneLab, NVIDIA lab workflows, and partner programs, but conversion requires Lilly-owned candidates, better portfolio choices, or clinical milestones.
Revenue mix
One pharmaceutical segment. Q1 2026 revenue was dominated by cardiometabolic medicines: Mounjaro and Zepbound were $12.822B, and cardiometabolic revenue was about 80% of company revenue.
Latest qtr revenue
Q1 2026 revenue was $19.799B, sourced from the LLY security lane's Q1 2026 10-Q and earnings-release evidence. AI evidence is adoption proof; GLP-1 price, supply, capacity, and cash conversion remain the larger company gates.
Regeneron owns genetics, proteomics, EHR-linked cohort, and antibody-discovery infrastructure.
Market cap$74.9B
Next earningsNot confirmed
Latest qtr revenue$3.605B
Role in stack
Regeneron uses owned genetics, proteomics, EHR-linked data, and antibody discovery to validate targets, define biomarkers, and choose cohorts. The economic gate is whether that data produces internal programs that offset EYLEA pressure and extend Dupixent-era cash flow.
Revenue mix
Q1 2026 revenue was led by collaboration revenue at 52.7% and product sales at 42.6%. Dupixent profit share, EYLEA transition, and internally controlled pipeline scale remain central.
Latest qtr revenue
Q1 2026 revenue was $3.605B, sourced from the REGN security lane and Q1 2026 results evidence. The data-owner edge is strong, but the stock still needs EYLEA HD conversion and pipeline commercial proof.
AstraZeneca has oncology, rare-disease, Evinova, and partner AI evidence tied to a broad clinical pipeline.
Market cap$283.6B
Next earningsNot confirmed
Latest qtr revenue$15.288B
Role in stack
AstraZeneca is a scaled clinical buyer that can convert AI-supported target, antibody, biomarker, or trial-design work into company-controlled oncology and rare-disease programs. The gate is clinical, regulatory, and cash-conversion evidence.
Revenue mix
One IFRS operating segment. Q1 2026 Oncology revenue was $6.798B, about 44% of revenue; Rare Disease was another material growth lane, while policy, China, and partner economics affect conversion quality.
Latest qtr revenue
Q1 2026 total revenue was $15.288B, sourced from the AZN security lane's Q1 2026 results evidence. Latest-quarter revenue is useful for scale, but Phase 3 and label quality decide whether AI/platform work matters.
Merck has Mayo Clinic and internal AI/data-science evidence, with the Keytruda replacement bridge as the main gate.
Market cap$275.1B
Next earningsAug 4, 2026
Latest qtr revenue$16.286B
Role in stack
Merck is a large R&D buyer using AI and multimodal clinical data for target discovery, precision medicine, and trial decisions. The gate is whether these workflows help build non-Keytruda assets before the 2028 U.S. patent cliff.
Revenue mix
Pharma is the dominant business. Q1 2026 sales were heavily anchored by Keytruda/Qlex at about 49% of company sales, so AI and business development must improve the post-Keytruda asset bridge.
Latest qtr revenue
Q1 2026 sales were $16.286B, sourced from the MRK security lane and Merck Q1 2026 materials. Merck's IR events page confirms the next earnings call date as August 4, 2026.
Pfizer has disease-model, foundation-model, and biologics-design partnerships, but the reset story dominates.
Market cap$146.4B
Next earningsNot confirmed
Latest qtr revenue$14.451B
Role in stack
Pfizer buys and partners for disease models, biomolecular foundation models, small molecules, biologics, and ADC workflow improvement. AI value needs to become better target selection, patient selection, filings, pivotal assets, or differentiated data packages.
Revenue mix
Biopharma represented nearly all Q1 2026 revenue, with oncology an important growth area. The equity debate still turns on LOE, debt, dividend coverage, pipeline credibility, and cash conversion.
Latest qtr revenue
Q1 2026 revenue was $14.451B, sourced from the PFE security lane and Q1 2026 company evidence. Pfizer's IR events page did not list a confirmed upcoming earnings date when checked on 2026-06-13.
Gilead has oncology real-world-data and discovery-partner exposure funded by durable HIV cash flow.
Market cap$163.1B
Next earningsNot confirmed
Latest qtr revenue$6.960B
Role in stack
Gilead is a cash-flow-backed buyer where AI and oncology data can improve trial design, patient selection, oncology R&D, or acquired-asset validation. The route is narrower than Lilly or Regeneron because HIV durability remains the main funding source.
Revenue mix
Q1 2026 product sales were led by HIV at about 72% of product sales. Oncology, real-world evidence, and acquired pipeline assets carry most of the AI/data relevance.
Latest qtr revenue
Q1 2026 total revenue was $6.960B, sourced from the GILD security lane and Q1 2026 company evidence. Gilead's IR events page did not show a confirmed upcoming earnings date when checked on 2026-06-13.
Johnson & Johnson has large clinical and structural-data assets, but MedTech and litigation dilute the pharma-buyer read.
Market cap$532.8B
Next earningsNot confirmed
Latest qtr revenue$24.1B
Role in stack
Johnson & Johnson owns a broad clinical, therapeutic, and structural-data estate and has AI discovery collaborations. Conversion requires AI-supported assets or trial decisions that improve Innovative Medicine growth after STELARA erosion.
Revenue mix
Q1 2026 sales were split between Innovative Medicine at about 64% and MedTech at about 36%. Litigation, MedTech margins, STELARA erosion, and net debt dilute the AI-biology signal.
Latest qtr revenue
Q1 2026 sales were $24.1B, sourced from the JNJ security lane and Q1 2026 10-Q evidence. A confirmed next earnings date was not found in the bounded 2026-06-13 check.
Incyte has a more direct Genesis Molecular AI link, but its data-owner breadth is narrower than large pharma.
Market cap$19.7B
Next earningsNot confirmed
Latest qtr revenue$1.273B
Role in stack
Incyte uses proprietary experimental data with Genesis Molecular AI and has higher earnings sensitivity to successful pipeline leverage. The gate is target progression, clinical evidence, and post-Jakafi replacement durability.
Revenue mix
Q1 2026 revenue remained concentrated: Jakafi was about 59.5%, Opzelura about 11.2%, and hematology/oncology ex-Jakafi contributed a smaller but faster-growing launch base.
Latest qtr revenue
Q1 2026 revenue was $1.2727B, sourced from the INCY security lane's Q1 2026 10-Q and financial-update evidence. A confirmed next earnings date was not found in the bounded 2026-06-13 check.
Market caps use read-only discovery instruments.market_cap values queried on 2026-06-13. Next-earnings searches preferred company IR pages; only Merck had a confirmed Q2 2026 earnings event in the bounded check, while the other cards use Not confirmed. Latest-quarter revenue uses Q1 2026 total revenue or sales from the linked security lanes and their company filing or earnings-release evidence.
What Confirms Or Weakens
AreaWhat confirmsWhat weakens or invalidatesWatch next
01Node thesisAI reaches owned assets
What confirms
AI-linked work becomes accepted discovery packages, option exercises, IND or CTA filings, first-human dosing, Phase 2 signal, paid milestones, or explicit portfolio-prioritization decisions.
What weakens or invalidates
Announcements remain platform access, model benchmarks, or broad productivity language with no named program, filing, milestone, clinical decision, or cash economics.
Watch next
LLY TuneLab/NVIDIA programs
REGN genetics-backed assets
AZN readouts
MRK bridge assets
INCY Genesis targets
02Data ownershipUsable proprietary inputs
What confirms
Clinical, genomic, proteomic, image, assay, or experimental datasets are rights-cleared, versioned, auditable, and used in model training, validation, or development decisions.
What weakens or invalidates
Data rights, consent, privacy, quality, or validation claims remain vague; partner data cannot be reused, linked, or traced to R&D allocation.
Watch next
Dataset governance
Partner terms
External validation
Context-of-use evidence
03Economics and pipelineRevenue or replacement proof
What confirms
Pipeline choices, R&D productivity, milestone receipts, launch quality, or commercial assets improve enough to affect revenue, margin, cash flow, or replacement-asset durability.
What weakens or invalidates
AI spend rises while LOE, pricing, debt, dilution, weak launches, or failed readouts dominate the company outcome.
Watch next
Q2 2026 earnings
R&D commentary
Clinical readouts
Regulatory filings
Guidance bridges
04Customer and policyDrug economics stay collectible
What confirms
Payers, regulators, and providers support access enough that AI-originated or AI-assisted assets can convert into net sales, label expansion, or replacement growth.
What weakens or invalidates
Medicare, MFN-style pricing, formulary design, rebates, China access, FDA model-credibility requirements, or trial-integrity issues reduce monetization quality.
Watch next
FDA AI guidance
Medicare/MFN rules
Gross-to-net changes
China compliance outcomes
05Funding and capitalAI spend fits the P&L
What confirms
AI, platform, and business-development spending fit inside cash-flow capacity and help prioritize higher-quality assets without weakening debt, dividends, buybacks, or acquisition flexibility.
What weakens or invalidates
Debt-funded deals, acquired IPR&D charges, capex, dividends, or buybacks crowd out discovery investment or obscure returns.
Watch next
LLY cash conversion
MRK acquisition funding
PFE dividend/debt
GILD oncology M&A
INCY cash deployment
06Stale conditionRefresh trigger
What confirms
Knowledge lanes, earnings-date checks, and read-only discovery coverage remain current through their stated dates, and no material filing, earnings, AI collaboration, regulatory, or clinical update has arrived.
What weakens or invalidates
Roche, Sanofi, Novartis, Novo Nordisk, ABBV, AMGN, BMY, or GSK receives local coverage that changes the basket, or a ranked company discloses material AI-linked clinical or economic evidence.
Watch next
Post-earnings refresh
FDA events
ASCO/ESMO data
AI partnership expansions
Discovery/API refresh
Source Trail
Canonical Thesis
AI Biology Platform Validation for the validation ladder across computational discovery, autonomous labs, diagnostics data, sequencing, and omics platforms.
Health Care for policy, reimbursement, funding, tools, pharma, and biotech customer-budget context.
Market caps use read-only instruments.market_cap values from ../discovery/data/discovery.duckdb, queried with duckdb -readonly on 2026-06-13: LLY $845.8B, REGN $74.9B, AZN $283.6B, MRK $275.1B, PFE $146.4B, GILD $163.1B, JNJ $532.8B, and INCY $19.7B.
Latest-quarter revenue uses Q1 2026 total revenue or total sales from the linked security lanes and their cited company filings or earnings releases: LLY $19.799B, REGN $3.605B, AZN $15.288B, MRK $16.286B, PFE $14.451B, GILD $6.960B, JNJ $24.1B, and INCY $1.2727B.
Earnings-date checks were bounded to company IR/event pages and Nasdaq earnings pages on 2026-06-13. Merck's IR events page confirmed Q2 2026 Earnings Call on August 4, 2026. Pfizer, Gilead, and Johnson & Johnson IR pages did not list a confirmed upcoming earnings event in the captured page text; Nasdaq earnings pages for LLY, REGN, AZN, PFE, GILD, JNJ, and INCY did not provide a credible confirmed future date in this bounded review. Those cards use Not confirmed.
Discovery And Chart Provenance
Read-only DuckDB query against ../discovery/data/discovery.duckdb confirmed daily_ohlc coverage through 2026-06-12 for LLY, REGN, AZN, MRK, PFE, GILD, JNJ, and INCY.
Weekly chart bars should be derived from daily rows by calendar week: open is the first row, high is weekly max, low is weekly min, close is final row, and volume is summed.
The selected-security right rail uses the report API route /api/securities/{ticker}/chart?frequency=weekly&window=3y&as_of=latest, with 20W EMA, 100W EMA where enough weekly history exists, and weekly volume. This page does not embed full OHLC payloads.
Lineage checks for representative tickers LLY and MRK found successful original daily_ohlc ingestion runs with provider source freshness through 2026-05-08, while the canonical table itself contains rows through 2026-06-12.
The local report API was not running during validation on 2026-06-13: curl -fsS 'http://127.0.0.1:8765/api/securities/LLY/chart?frequency=weekly&window=3y&as_of=latest' failed to connect. Right-rail chart hooks remain API-routed and should be checked after the API is started.
Known Gaps
RHHBY, SNY, NVS, and NVO have relevant external AI/data evidence but no local security lane in this workspace at the time of this build, so they remain source-only rather than ranked chart-core names.
ABBV and AMGN have local security lanes and chart coverage, but the merged research pass did not find enough current node-specific AI/data evidence to displace the eight parent-core names.
AI partnership evidence is treated as adoption evidence. Clinical-success evidence requires filings, milestones, trial starts, readouts, approvals, or commercial economics.
Static setup thresholds from the prior build used 2026-06-05 weekly bars and are stale versus the 2026-06-12 daily_ohlc coverage check. Use the right-rail API chart after refresh before treating setup labels or thresholds as current trading evidence.