Global AI in Drug Discovery Market Outlook to 2032

Executive Summary

The global AI in drug discovery market — defined as the full value chain of AI platforms applied to drug research (generative chemistry, foundation models, physics-plus-ML simulation, geometric deep learning, structure prediction, lab automation), AI-enabled discovery services, AI biotech companies with proprietary therapeutic pipelines, dedicated AI software-as-a-service for pharmaceutical research, plus the AI-driven portion of big pharma R&D spend including NVIDIA-anchored supercomputing infrastructure — is estimated at approximately US$5.5 billion in 2025 and is projected to reach approximately US$28 billion by 2032, expanding at a CAGR of 25–27 percent over the forecast period. The market has crossed a structural inflection in 2025 from technology promise to clinical proof, but AI-driven drug discovery remains a high-risk frontier rather than a proven commercial pathway — no AI-discovered drug has yet reached FDA approval, and several high-profile clinical failures (Recursion's REC-994 and REC-2282 discontinuations) underscore the structural risk.

Three forces define the trajectory through 2032. First, the industry's first proof-of-concept clinical validation of AI-driven drug discovery was published in Nature Medicine on June 3, 2025 — Insilico Medicine's Rentosertib (ISM001-055, also known as INS018_055), a first-in-class TNIK inhibitor developed using Insilico's generative AI platform Pharma.AI for idiopathic pulmonary fibrosis (IPF). The Phase IIa GENESIS-IPF trial (71 patients, 12 weeks, 22 sites in China) showed patients receiving 60 mg QD Rentosertib experienced a mean +98.4 mL forced vital capacity improvement versus -20.3 mL decline in the placebo group — a clinically meaningful efficacy signal that Insilico is advancing to Phase IIb in 2025–2026. Second, Isomorphic Labs (Google DeepMind spin-out from 2021, AlphaFold IP holder) raised US$600 million in April 2025 in its first-ever external funding round (Thrive Capital-led) — bringing total partnership and funding value to approximately US$3.6 billion when combined with existing collaborations with Novartis and Eli Lilly (worth nearly US$3 billion combined). Isomorphic prepared to launch first human trials of AI-designed oncology drugs in 2025–2026, representing the first commercial test of the AlphaFold 3 protein-DNA-RNA-ligand prediction platform that Google DeepMind and Isomorphic released in 2024. Third, the big pharma plus NVIDIA supercomputing partnership wave reaches commercial scale: Eli Lilly-NVIDIA announced "the most powerful supercomputer owned and operated by a pharmaceutical company" in October 2025 using NVIDIA's BioNeMo platform; Roche-NVIDIA launched a large-scale AI factory partnership 2025; AstraZeneca reports over 50 percent faster target drug design and validation in early discovery where AI has been applied; Sanofi-OpenAI-Formation Bio collaboration (2024) plus emerging Pfizer, GSK, BMS, and Moderna AI programmes collectively position big pharma as anchor demand-side participants.

For investors, pharma executives, biotech founders, and policymakers, the implication is that AI drug discovery has crossed from "interesting science" to commercial inflection, but the path from technology to FDA-approved drug remains a 10–15 year journey with structural attrition risk at each phase. The 2026–2030 period is the decisive window where (a) Insilico Phase IIb plus emerging Isomorphic Phase 1 oncology trials test the AI-design-to-clinical-success pathway, (b) Recursion plus other AI biotechs face make-or-break inflection points on remaining pipeline assets, (c) big pharma-AI partnerships either translate to commercially-meaningful drug approvals or are reassessed for ROI sustainability, and (d) the first AI-discovered drug FDA approval (forecast 2027–2029 base case) validates or challenges the structural promise of the category.

Market Overview

Definition and Scope

This report scopes the AI in drug discovery market as the full value chain of artificial intelligence applied to pharmaceutical research and development — AI platforms for target identification, hit discovery, lead optimisation, preclinical development, plus AI-enhanced clinical trial design and patient stratification; AI biotech companies with proprietary therapeutic pipelines (Insilico Medicine, Recursion, Schrödinger, BenevolentAI, Relay Therapeutics, Exscientia post-merger, Isomorphic Labs, Generate Biomedicines, Cradle.bio, plus emerging entrants); AI software-as-a-service platforms (Schrödinger LiveDesign, NVIDIA BioNeMo, Microsoft Azure AI Health, Google Cloud Life Sciences, OpenAI GPT-Rosalind); AI-enabled discovery services to big pharma (custom platform development, consulting, computational chemistry services); plus the AI-driven portion of big pharma R&D infrastructure (NVIDIA-anchored supercomputers, proprietary AI model development, integrated AI laboratories).

The scope excludes general healthcare AI applications outside drug discovery (medical imaging diagnostics, electronic health record analytics, hospital operations — covered in separate digital health markets), AI in manufacturing (covered in pharmaceutical manufacturing analytics), plus AI-driven medical device development (covered in medical devices market). The scope captures both AI biotech companies with proprietary pipelines and AI platform/services revenue from big pharma — recognising that the value pool is bifurcating between platform monetisation and discovered-drug monetisation.

Evolution and Genesis

The AI drug discovery market evolved through four structurally distinct phases since 2010. The 2010–2017 quantitative chemistry plus early ML phase was anchored by computational chemistry expertise at Schrödinger (founded 1990, IPO 2020), Numerate, and emerging machine learning applications at Atomwise (founded 2012) and BenevolentAI (founded 2013). Cumulative venture funding into AI drug discovery startups reached approximately US$500 million by end-2017 — modest scale, characterised by platform-development focus rather than commercial drug pipelines.

The 2018–2021 platform-and-deep-learning phase was triggered by AlphaFold 1 (December 2018, Google DeepMind) demonstrating that deep learning could predict protein structures with competitive accuracy. The 2020 AlphaFold 2 release (CASP14 dominant performance) plus the 2021 protein structure database open access (over 200 million proteins predicted) created the foundational infrastructure. Insilico Medicine (founded 2014) demonstrated generative chemistry capabilities (the GENTRL paper in Nature Biotechnology 2019 generated novel kinase inhibitors). Recursion Pharmaceuticals (founded 2013) IPO'd April 2021 at US$2.7 billion valuation. Exscientia IPO'd October 2021 at US$2.5 billion. Cumulative VC funding into AI drug discovery startups exceeded US$5 billion by end-2021.

The 2022–2024 partnership-and-pipeline phase saw the structural shift from platform-development to clinical-stage assets and pharma partnerships. Sanofi-Exscientia collaboration ($120 million for 10 programmes, January 2022). Schrödinger-BMS deal ($55 million upfront, $2.7 billion in milestones, 2024) for HIF-2α plus SOS1 and KRAS targets. Eli Lilly-Genesis Therapeutics (now part of broader Lilly AI strategy). Roche-Recursion partnership. Microsoft-Moderna deal. Insilico-Sanofi collaboration. The DEEP Genomics, Atomwise, Cradle.bio plus emerging AI biotech raised approximately US$8–12 billion in cumulative VC plus IPO funding through 2024. AlphaFold 3 release (April 2024, Google DeepMind + Isomorphic Labs) extended structure prediction to DNA, RNA, ligands, and protein-ligand interactions — the structural inflection from protein-only prediction to multi-modal molecular interaction modelling.

The 2025-onward clinical-proof and commercial-scaling phase is the current era. Three structural events define the new phase: (a) Insilico Medicine's Rentosertib Phase IIa GENESIS-IPF Nature Medicine publication on June 3, 2025 — the industry's first proof-of-concept clinical validation of AI-driven drug discovery, with patients receiving 60 mg QD Rentosertib experiencing mean +98.4 mL FVC improvement versus -20.3 mL placebo decline, (b) Isomorphic Labs raising US$600 million in April 2025 (Thrive Capital-led) plus preparing first human oncology trials in 2025–2026 — bringing AlphaFold 3-derived drug design to clinical testing, and (c) Eli Lilly-NVIDIA announcement (October 2025) of "the most powerful supercomputer owned and operated by a pharmaceutical company" plus parallel Roche-NVIDIA AI factory partnership plus Sanofi-OpenAI-Formation Bio collaboration plus AstraZeneca reporting over 50 percent faster target drug design via AI. The clinical-proof phase is structurally consequential because it converts AI drug discovery from speculative-technology to operationally-deployed pharmaceutical R&D infrastructure — but the path from clinical proof to FDA-approved drug remains uncertain.

Key Market Drivers

• Insilico Rentosertib Phase IIa Nature Medicine publication June 3, 2025. The Phase IIa GENESIS-IPF trial (71 patients, 12 weeks, 22 sites in China) — the industry's first proof-of-concept clinical validation of AI-driven drug discovery. Patients receiving 60 mg QD Rentosertib experienced mean +98.4 mL FVC improvement versus -20.3 mL placebo decline. Insilico is advancing to Phase IIb in 2025–2026.
• Isomorphic Labs US$600 million Thrive Capital-led funding April 2025. First external funding round positions Isomorphic for first human oncology trials in 2025–2026. Combined with existing Novartis (US$1.2B) and Eli Lilly (US$1.7B) partnerships, total Isomorphic value exceeds US$3.6 billion. AlphaFold 3 (2024) extended structure prediction to multi-modal molecular interactions.
• Big pharma plus NVIDIA supercomputing partnership wave. Eli Lilly-NVIDIA pharma's "most powerful supercomputer" announcement October 2025 using NVIDIA BioNeMo platform. Roche-NVIDIA large-scale AI factory partnership 2025. AstraZeneca reports over 50 percent faster target drug design and validation in early discovery where AI has been applied. Sanofi-OpenAI-Formation Bio collaboration 2024 plus emerging Pfizer, GSK, BMS, Moderna AI programmes.
• AI drug discovery investment and pipeline scaling. Cumulative VC plus IPO plus strategic funding into AI drug discovery exceeded US$15 billion through 2024–2025. Active clinical-stage assets discovered or designed with AI exceed 75 distinct compounds across over 30 therapeutic areas. The forward pipeline (Phase 2 and beyond) is forecast to grow from approximately 30 active programmes in 2025 to approximately 150–200 by 2030.

Macroeconomic and Regulatory Context

The market is operating against an enabling but uncertain regulatory framework: FDA "Predetermined Change Control Plan" (PCCP) guidance plus broader AI/ML medical device framework provides regulatory pathway clarity for AI-enabled clinical decision support but does not yet directly address AI-discovered drugs (the FDA approval pathway for AI-discovered drugs remains the standard pharmaceutical pathway: IND, Phase 1, Phase 2, Phase 3, BLA/NDA, post-marketing surveillance); EU AI Act (in force August 2024) classifies AI applications in healthcare as "high-risk" requiring conformity assessment but does not specifically address drug discovery applications; UK MHRA emerging AI guidance plus EMA's Q&A on AI/ML applications to medicinal product development; China NMPA AI medical device review guidance plus emerging AI drug discovery framework.

The macroeconomic backdrop is structurally supportive but increasingly bifurcated. NVIDIA's commercial success (BioNeMo platform plus DGX SuperPOD plus integrated AI factory partnerships) has accelerated big pharma adoption of AI infrastructure — Eli Lilly's announced supercomputer plus Roche's AI factory plus emerging Pfizer, GSK, AstraZeneca, Sanofi NVIDIA-anchored programmes collectively represent approximately US$4–6 billion in cumulative AI infrastructure investment 2024–2025. Hyperscaler cloud platforms (AWS, Google Cloud, Microsoft Azure) plus dedicated AI cloud services (Anthropic, OpenAI, plus emerging Replicate, Modal, Anthropic-Lilly) provide infrastructure abstraction. The structural challenge: AI drug discovery has not yet translated to FDA-approved drugs (zero approvals as of mid-2026), creating ROI uncertainty for big pharma plus skepticism from established pharmaceutical research leaders. The forward implication is that the 2026–2028 period will be the structural validation moment — either AI-discovered drugs reach FDA approval and ROI inflection materialises, or the category undergoes consolidation similar to the 2010s genomics boom-and-correction cycle.

Market Size & Growth Outlook

The market grew from approximately US$0.9 billion in 2020 to approximately US$3.8 billion in 2024 — a 4.2× expansion in four years driven by venture capital plus IPO funding (Recursion April 2021 IPO at US$2.7 billion, Exscientia October 2021 IPO at US$2.5 billion, Schrödinger February 2020 IPO at US$2.7 billion) plus pharma partnership transaction value (Schrödinger-BMS US$2.7 billion deal, Insilico-Sanofi, Recursion-Roche, Sanofi-Exscientia US$120 million, plus 30+ smaller transactions). The 2024 step-change (46.2 percent year-on-year value growth) reflects the convergence of three forces: (a) AlphaFold 3 release (April 2024) extending structure prediction to multi-modal molecular interactions, (b) NVIDIA's BioNeMo platform reaching commercial scale with Eli Lilly, Roche, and emerging pharma adoption, and (c) the broader AI infrastructure investment wave that lifted all AI-adjacent categories.

The 2025 expansion to US$5.5 billion (44.7 percent year-on-year value growth) is anchored by three structural catalysts. First, Insilico Medicine's Rentosertib Phase IIa Nature Medicine publication (June 3, 2025) — the industry's first proof-of-concept clinical validation of AI-driven drug discovery — generated material demand for AI biotech investment, partnership transactions, and pharma adoption. Second, Isomorphic Labs raising US$600 million in April 2025 (Thrive Capital-led) plus preparing first human oncology trials in 2025–2026 validated the AlphaFold 3 commercial pathway. Third, the big pharma-NVIDIA supercomputing partnership wave reached commercial scale — Eli Lilly-NVIDIA pharma's "most powerful supercomputer" announcement October 2025, Roche-NVIDIA AI factory partnership, AstraZeneca-Genentech-NVIDIA, plus emerging Pfizer, GSK, BMS programmes.

The forecast CAGR of 25–27 percent through 2032 anchors on three drivers. The first is continued pharma partnership transaction acceleration: from approximately 40 named AI biotech-pharma partnerships in 2025 to approximately 120–150 by 2030, with cumulative announced deal value growing from approximately US$8 billion in 2025 to approximately US$45–60 billion by 2030. Partnership economics include upfront payments (typically US$10–80 million per programme), milestone payments (typically US$200–2,000 million in cumulative milestones across discovery, preclinical, clinical, and commercial milestones), plus royalties (typically 5–15 percent of net commercial sales). The second driver is the structural scaling of AI infrastructure investment by big pharma: from approximately US$0.5 billion in 2024 to approximately US$3–5 billion per year by 2030 across NVIDIA-anchored supercomputers, dedicated AI laboratories, plus emerging hyperscaler-pharma partnerships. Big pharma AI spend is forecast to grow from approximately 3 percent of total pharmaceutical R&D in 2024 to approximately 10–14 percent by 2030. The third driver is the first AI-discovered drug FDA approval (forecast 2027–2029 base case): once an AI-discovered drug reaches commercial approval, the validation effect materially accelerates pipeline development plus pharma adoption across multiple therapeutic areas.

Cumulative investment over the 2025–2032 window (AI biotech VC plus IPO plus follow-on funding plus big pharma AI infrastructure plus partnership transactions) is forecast at approximately US$165–195 billion across the AI drug discovery value chain — anchored by approximately US$80–100 billion in cumulative AI biotech funding (VC, IPO, follow-on, plus emerging acquisition activity by big pharma), plus approximately US$45–55 billion in big pharma direct AI infrastructure investment, plus approximately US$40 billion in cumulative partnership deal value transferred to AI biotechs. This investment magnitude reconciles to approximately 8–10× the average annual market size in the forecast window — reflecting the speculative investment intensity inherent in early-stage technology-led pharmaceutical R&D. The implication for stakeholders is that AI drug discovery is currently in a pre-revenue (clinical-stage) phase where capital deployment materially exceeds revenue realisation — making the 2027–2029 first-FDA-approval inflection the principal value-realisation milestone.

Market Segmentation

By Drug Discovery Stage

Lead optimisation (small molecule design) dominates current AI application value share (approximately 24 percent in 2025) because it is the highest-value design challenge — moving from a hit compound to a clinical-stage drug candidate typically requires synthesis and characterisation of approximately 1,000–10,000 analogues, and AI-augmented design materially compresses this cycle. Insilico Medicine's Pharma.AI Chemistry42 platform plus Exscientia (now part of Recursion post-2024 merger) plus Schrödinger LiveDesign plus Genesis Therapeutics plus emerging Cradle.bio plus Atomwise collectively address this stage. The share is forecast to grow modestly to 26 percent by 2032 as lead optimisation remains the principal value-creation stage for AI-discovered drugs.

The fastest-growing application categories are clinical trial design and patient stratification (forecast to grow from 10 percent share in 2025 to 12 percent by 2032) and preclinical development (12 percent to 13 percent). The structural reason: as AI-discovered drugs progress to clinical-stage trials, the value pool shifts from chemistry-only applications to clinical operations applications (patient selection, biomarker discovery, trial design optimisation, dose finding, plus emerging adaptive trial design). Target identification and validation remains structurally important (22 percent in 2025, 20 percent in 2032) — Insilico Medicine's PandaOmics platform plus emerging AI-driven multi-omics integration plus OpenAI GPT-Rosalind for Amgen plus the broader Microsoft/Google Cloud life sciences offerings address this stage.

By Therapeutic Area

Oncology dominates the AI-discovered clinical pipeline (approximately 38 percent share in 2025) because of three structural reasons: (a) oncology has the largest revenue opportunity per approved drug (top oncology drugs reach US$25+ billion annual revenue versus US$5–10 billion in most non-oncology categories), (b) the FDA's accelerated approval pathway for oncology drugs plus PDUFA expedited review provides faster commercial validation, and (c) the structural understanding of cancer biology supports AI applications in target identification and lead optimisation. Isomorphic Labs explicitly stated its first AI-designed clinical trials will be in oncology. Schrödinger's HIF-2α (kidney cancer) plus SOS1- and KRAS-driven tumour collaboration with BMS plus emerging Genesis Therapeutics oncology pipeline plus Recursion's continuing cancer focus all anchor the oncology AI pipeline.

The fastest-growing therapeutic area in percentage terms is fibrosis (IPF, MASH, kidney) — forecast to grow from 7 active AI-discovered clinical assets in 2025 to approximately 20 by 2030 (a 2.9× expansion). Insilico Medicine's Rentosertib for IPF is the leading proof-of-concept, with parallel fibrosis programmes at Recursion (REC-3964 discontinued, but other fibrosis programmes continue), BenevolentAI plus emerging fibrosis-specific AI biotech entrants. The strategic implication: fibrosis represents a high-unmet-need indication (IPF has only two approved drugs — pirfenidone and nintedanib, both with limited efficacy — and emerging MASH plus kidney fibrosis indications have multi-billion-dollar revenue potential).

By Technology Approach

Generative AI (transformer-based plus diffusion models) is the largest current AI approach in drug discovery (approximately 32 percent share in 2025) and is forecast to grow to approximately 37 percent by 2032 — reflecting both the broader generative AI infrastructure investment plus the specific applicability of generative chemistry to small molecule design. Insilico Medicine's Chemistry42 platform plus Cradle.bio plus Genesis Therapeutics plus Iktos plus emerging pharma-internal generative chemistry programmes collectively dominate this approach.

Physics-based plus machine learning hybrid (approximately 21 percent share in 2025) is the structural defensible approach — Schrödinger's LiveDesign platform plus OpenEye (acquired by Cadence Molecular Sciences) plus Cresset combine first-principles physics simulation with machine learning for property prediction. The structural advantage: physics-based methods provide better extrapolation outside training data distribution versus pure ML methods, supporting drug candidate optimisation at the structural-chemistry frontier. Schrödinger's commercial success (BMS US$2.7 billion deal, Sanofi US$120 million, Eli Lilly integration into LiveDesign January 2026) demonstrates the structural commercial pathway for physics-plus-ML platforms.

Foundation models (AlphaFold 3, Meta ESM, NVIDIA BioNeMo, Baker Lab RFdiffusion) are forecast to grow from 18 percent share in 2025 to approximately 20 percent by 2032 — reflecting the structural scaling of large-foundation-model infrastructure across all drug discovery applications. The forward implication: foundation models become the infrastructure layer that other AI approaches build on, similar to how GPT-style models became infrastructure for general AI applications.

By Business Model

AI biotech companies with proprietary therapeutic pipelines collectively dominate value share (approximately 35 percent in 2025) — the structural reason being that proprietary pipeline ownership captures both the upside of successful drug development plus the partnership transaction value when assets are licensed to pharma. Insilico Medicine (proprietary IPF, fibrosis, oncology, immunology, and rare disease pipeline plus pharma partnerships with Sanofi and others), Recursion (post-Exscientia merger pipeline across cancer, rare disease, and emerging neurology), BenevolentAI (neurology plus oncology), Relay Therapeutics (oncology, plus emerging rare disease), Isomorphic Labs (oncology, plus emerging cardiovascular), Generate Biomedicines (protein therapeutics), Genesis Therapeutics, Verge Genomics, Insitro — collectively position this archetype as the principal value creator.

AI platform/SaaS for pharma (approximately 22 percent share in 2025) is the structural service-revenue layer. Schrödinger (NASDAQ-listed since 2020, FY2024 revenue approximately US$190 million combining software, plus drug discovery collaborations, plus partial pipeline economics) is the leading public platform company. NVIDIA BioNeMo (commercial scaling 2024–2025) plus Atomwise plus Cresset plus emerging clinical-trial AI specialists provide additional platform layers. Schrödinger's strategic positioning combines physics-based simulation, machine learning, plus emerging quantum-mechanical methods integration into the LiveDesign software platform — providing structural defensibility versus pure-ML platforms.

Big pharma internal AI R&D (approximately 18 percent of 2025 value share, forecast 20 percent by 2032) represents the structural scaling of in-house AI capability at major pharmaceutical companies. Eli Lilly (NVIDIA-anchored supercomputer announced October 2025, plus emerging AI laboratory infrastructure), Roche/Genentech (NVIDIA AI factory partnership 2025), AstraZeneca (over 50 percent faster target drug design where AI applied), Sanofi (OpenAI-Formation Bio collaboration plus internal AI), Pfizer (internal AI plus emerging partnerships), GSK, BMS, Novartis, Merck, Bayer — collectively represent approximately US$3–5 billion in cumulative AI infrastructure investment 2024–2025 with structural acceleration through 2030.

By Geography

The United States dominates AI drug discovery (approximately 56 percent of 2025 market value) because of three structural factors: largest pharmaceutical R&D market globally (US pharma R&D approximately US$120 billion annually), deepest AI talent pool plus US-based hyperscaler infrastructure (NVIDIA, OpenAI, Anthropic, Google, Microsoft, Meta), plus the most active VC plus capital markets for AI biotech. The share is forecast to compress modestly to 52 percent by 2032 as Chinese plus UK plus broader European deployment grows — but US dominance remains structurally durable.

The United Kingdom is the second-largest geography (approximately 14 percent share in 2025) anchored by Isomorphic Labs (London, Google DeepMind spin-out plus US$600 million April 2025 funding), BenevolentAI (London), Exscientia legacy (Oxford-founded, acquired by Recursion 2024), plus emerging Cradle.bio UK presence. The UK Government's AI Sector Deal plus Genomics England plus emerging UK biotech VC funding (Octopus Ventures, Atomico, Index Ventures UK office plus emerging Hadean) provide supportive ecosystem.

China is the third-largest geography (approximately 9 percent share in 2025) and is forecast to grow to approximately 12 percent by 2032 — driven by Insilico Medicine's Hong Kong + Mainland presence (where the Phase IIa GENESIS-IPF trial of Rentosertib was conducted), XtalPi (Shenzhen-based AI drug discovery), plus emerging Tencent and Alibaba healthcare AI investment. The Chinese 14th Five-Year Plan plus AI healthcare strategic plan plus emerging dedicated state funding for AI drug discovery position China as the principal non-Western AI drug discovery growth opportunity.

By Partnership Structure

Multi-target discovery collaborations dominate the partnership structure (approximately 38 percent of 2025 active deals) because of the structural alignment between pharma's need to fill pipeline gaps and AI biotech's platform-validation imperative. Schrödinger-BMS (US$2.7 billion multi-target including HIF-2α, SOS1, KRAS, plus undisclosed neurology programme) is the largest single multi-target deal. Sanofi-Exscientia (now part of Recursion) US$120 million for 10 programmes provides the structural template. Insilico-Sanofi plus emerging Roche-Recursion plus AstraZeneca-Verge Genomics plus emerging multi-target deals collectively dominate the partnership flow.

Single-asset license structures (22 percent of 2025 deals) represent specific clinical-stage compound licensing. The Insilico Rentosertib pathway (post Phase IIa positive readout) provides the forward template — Insilico is positioned to either advance Rentosertib through Phase III independently or out-license to a pharma partner for global development. The structural inflection: as AI-discovered drugs progress to mid-stage clinical trials with positive readouts, single-asset license deal values are forecast to materially increase (from approximately US$100–500 million per asset in 2024 to approximately US$500 million–2 billion per asset by 2030 for Phase II positive readout compounds).

Acquisition structures (6 percent of 2025 deals) represent the emerging consolidation wave. Recursion's 2024 acquisition of Exscientia (creating the combined Recursion-Exscientia entity) was the largest AI biotech-AI biotech merger to date. Forward consolidation pathways include big pharma acquiring AI biotech for platform-and-pipeline (rumored Roche-BenevolentAI discussions plus emerging AstraZeneca-Verge Genomics speculation) plus emerging mid-tier AI biotech consolidation. The forecast: 5–10 major AI biotech acquisitions by big pharma through 2030, with cumulative deal value of approximately US$15–25 billion.

Trends & Developments

Insilico Rentosertib Phase IIa Nature Medicine Publication (June 3, 2025)

The publication of Insilico Medicine's Rentosertib (ISM001-055, INS018_055) Phase IIa GENESIS-IPF trial results in Nature Medicine on June 3, 2025 represented the industry's first proof-of-concept clinical validation of AI-driven drug discovery. The trial enrolled 71 patients with IPF across 22 sites in China; patients receiving 60 mg QD Rentosertib experienced the greatest mean improvement in lung function as measured by forced vital capacity (FVC), with a mean change of +98.4 mL — compared to a mean decline of -20.3 mL in the placebo group. Insilico has begun discussions with regulatory authorities to facilitate prospective evaluation of Rentosertib in larger cohorts. The Phase IIb proof-of-concept study is targeted for initiation in 2025–2026. The structural implication: the Nature Medicine publication provides the scientific credibility benchmark for the AI drug discovery category — converting AI drug discovery from speculative-technology to clinically-validated R&D pathway.

Isomorphic Labs US$600 Million Funding and First Oncology Trials

Isomorphic Labs raised US$600 million in April 2025 in its first-ever external funding round (Thrive Capital-led). The funding plus existing Novartis (US$1.2 billion deal value) and Eli Lilly (US$1.7 billion deal value) partnerships brings total Isomorphic value to approximately US$3.6 billion. Isomorphic prepared to launch first human trials of AI-designed drugs in 2025–2026, with oncology candidates advancing first. The structural significance: Isomorphic is the most consequential test of the AlphaFold 3 commercial pathway — if Isomorphic's AI-designed oncology compounds achieve favorable Phase 1 readouts in 2026–2028, it validates the foundation-model approach to drug discovery; if they fail, it raises questions about the structure-prediction-only approach versus the broader generative chemistry plus phenotypic screening approaches of other AI biotechs.

Eli Lilly-NVIDIA Pharma's Most Powerful Supercomputer (October 2025)

Eli Lilly announced in October 2025 a partnership with NVIDIA to build what the company claims will be "the most powerful supercomputer owned and operated by a pharmaceutical company." The lab co-locates Lilly domain experts in biology, science, and medicine with AI model builders and engineers from NVIDIA — generating large-scale data and building AI models that accelerate medicine development using NVIDIA's BioNeMo platform. Parallel partnerships: Roche-NVIDIA AI factory (2025), Genentech-NVIDIA, plus AstraZeneca reporting over 50 percent faster target drug design where AI has been applied. The structural implication: big pharma is investing structurally in AI infrastructure at the magnitude of US$0.5–1.5 billion per major pharmaceutical company through 2028 — converting AI from external partnership category into integrated internal R&D capability.

Recursion Clinical Failures and the Cautionary Signal

Recursion discontinued three clinical-stage programmes through 2024–2025: REC-2282 (neurofibromatosis type 2, Phase 2 failed futility threshold primarily driven by the 40mg cohort — 60mg and combined dose arms did not pass futility criteria, limited overall tumor shrinkage and clinical activity), REC-994 (cerebral cavernous malformation, Phase II SYCAMORE trial validated safety but failed efficacy — long-term extension results showed no promising trends in MRI or functional outcome in patients crossing over from placebo to 400 mg dose), and REC-3964 (C. difficile, weighing out-licensing options after limited internal pursuit). The structural significance: Recursion's failures illustrate that AI-discovered drugs face the same dreaded fate as traditionally developed drugs — many fail in human clinical trials. While AI has demonstrated discovery-stage acceleration, none of the AI-discovered drugs have yet reached FDA approval. Recursion's post-failure strategic response (Exscientia merger, pipeline rationalisation, increased focus on oncology and rare disease) reflects the industry-wide pattern of AI biotech consolidation and refocus.

Schrödinger Physics-Plus-ML Hybrid as the Defensible Platform Model

Schrödinger's commercial success (US$2.7 billion BMS deal, US$120 million Sanofi collaboration, Eli Lilly LiveDesign integration January 2026, plus emerging AstraZeneca and Bayer collaborations) demonstrates the structural commercial pathway for physics-based plus machine learning hybrid platforms. Schrödinger combines first-principles physics simulation with machine learning for property prediction — providing structural defensibility versus pure-ML approaches that depend on training data distribution. The company's FY2024 revenue of approximately US$190 million (software licensing plus drug discovery collaborations plus partial pipeline economics) plus emerging revenue from BMS milestone payments plus the HIF-2α plus SOS1 plus KRAS proprietary pipeline economics position Schrödinger as the most commercially-mature AI drug discovery company. The strategic implication: physics-plus-ML hybrid approaches may prove more structurally durable than pure-ML approaches as the AI drug discovery category matures.

Big Tech Entry into Life Sciences (OpenAI, Anthropic, Google, Microsoft)

OpenAI is working with Amgen, Moderna, the Allen Institute, Thermo Fisher Scientific, and others to apply GPT-Rosalind across workflows that accelerate research and discovery. Microsoft-Moderna partnership (announced 2023, scaling through 2024–2025) integrates Azure AI plus Microsoft Copilot for Moderna's mRNA design workflows. Google Cloud Life Sciences plus emerging Anthropic partnerships with Eli Lilly, Pfizer plus other pharma collectively position big tech as both infrastructure provider and emerging drug discovery participant. The structural implication: AI drug discovery is converging with the broader AI infrastructure layer — meaning that pure-play AI biotechs face structural competition from big tech-pharma direct partnerships that bypass the AI biotech intermediary. The forward question: does the AI biotech archetype (proprietary pipeline plus platform model) sustain through 2030, or does it consolidate into big-pharma-led plus big-tech-anchored deployment?

Competitive Landscape

The competitive landscape is structurally organised into five archetypes. First, the leading AI biotech with proprietary clinical pipeline — Insilico Medicine (10 active clinical assets, Rentosertib lead), Recursion (8 active assets post-Exscientia merger), Schrödinger (7 active assets including HIF-2α plus SOS1/KRAS proprietary plus BMS partnership pipeline), Relay Therapeutics (6 active assets, RLY-2608 in Phase 3 oncology), Genesis Therapeutics (5 active assets) — collectively control approximately 50 percent of active clinical-stage AI-discovered assets. These players combine integrated discovery platforms plus proprietary pipeline development plus selective pharma partnership transactions.

Second, the AI biotech preparing first trials — Isomorphic Labs is the most consequential — preparing to launch first human oncology trials in 2025–2026 backed by US$600 million April 2025 funding plus US$3 billion in Novartis-Eli Lilly partnerships. Isomorphic Labs represents the first commercial test of the AlphaFold 3 protein-DNA-RNA-ligand prediction platform. The strategic significance: if Isomorphic's AI-designed oncology compounds achieve favorable Phase 1 readouts in 2026–2028, the AlphaFold-based approach validates structurally; if they fail, the AI drug discovery category faces the question of whether structure-prediction-only AI is sufficient versus the broader generative chemistry plus phenotypic screening approaches.

Third, the specialized AI biotech with focused therapeutic areas — BenevolentAI (4 active assets, neurology focus), Generate Biomedicines (4 active assets, protein therapeutics), Verge Genomics (3 active assets, ALS plus other neurology), Insitro (3 active assets, broad therapeutic), Atomwise (3 active assets, virtual screening), Iktos (2 active assets, generative chemistry), Cradle.bio (protein engineering), plus emerging Chinese players (Variational AI, XtalPi) — collectively represent the broader AI biotech ecosystem. Fourth, the AI platform-only players — NVIDIA BioNeMo, Schrödinger LiveDesign platform layer, Atomwise platform, OpenEye/Cadence Molecular Sciences — focus on platform monetisation through pharma licensing without proprietary pipeline development. Fifth, the big pharma internal AI R&D — Eli Lilly, Roche, Sanofi, Novartis, Pfizer, AstraZeneca, GSK, BMS, Merck, Amgen, Moderna — represents the structural scaling of internal AI capability through NVIDIA partnerships plus internal laboratories.

Insilico Medicine's strategic posture is the most consequential in 2025 — the Rentosertib Phase IIa Nature Medicine publication (June 3, 2025) established the company as the first AI biotech to deliver clinical proof-of-concept. Insilico's broader pipeline includes 10 active clinical-stage assets across IPF, oncology (USP1, USP30, multiple kinase targets), immunology, plus emerging neurology programmes. The forward Phase IIb GENESIS-IPF trial (2025–2026 initiation) plus emerging Phase 1 oncology trials position Insilico for cumulative pipeline value materially above the company's current US$2.5–3 billion private valuation. The cautionary signal: even with positive Phase IIa data, Phase III success rate for IPF compounds historically is approximately 30–40 percent — meaning Rentosertib may face additional clinical risk before FDA approval.

Recursion's strategic posture (post-Exscientia merger 2024) represents the structural challenge for AI biotechs that have not yet delivered clinical success. The discontinuation of REC-994 (cerebral cavernous malformation, Phase 2 failed efficacy), REC-2282 (neurofibromatosis type 2, Phase 2 failed futility), and REC-3964 (C. difficile, weighing out-licensing) plus the broader pipeline rationalisation reflects the post-failure strategic reset. Recursion's combined Recursion-Exscientia portfolio focuses on oncology and rare disease post-2024 — but the company's market capitalisation (approximately US$1.4 billion at end-2025) has compressed significantly from the post-IPO peak of approximately US$7 billion in April 2021. The strategic question: does Recursion's combined Recursion-Exscientia platform deliver subsequent clinical success that validates the AI biotech model, or does the company face additional pipeline failures?

Schrödinger's strategic posture combines the most mature commercial model with the broadest pharma collaboration footprint. The company's FY2024 revenue of approximately US$190 million plus the BMS US$2.7 billion deal plus Sanofi US$120 million plus Eli Lilly LiveDesign integration January 2026 plus emerging additional collaborations position Schrödinger as the structurally most-defensible AI drug discovery platform. The proprietary HIF-2α plus SOS1 plus KRAS pipeline economics (royalties plus milestone payments) plus the LiveDesign software licensing revenue plus the multi-target BMS collaboration provide diversified revenue streams that pure-play AI biotechs lack.

Isomorphic Labs represents the most contested 2025–2028 strategic outcome. The Google DeepMind spin-out (2021) is the structural test of the AlphaFold 3 commercial pathway — whether protein structure prediction at scale translates to drug design at the molecule level. The US$600 million April 2025 Thrive Capital funding plus US$3 billion Novartis-Eli Lilly partnerships provide approximately 4–6 years of operating runway. First oncology trials in 2025–2026 plus expected Phase 1 readouts in 2027–2028 will determine whether Isomorphic emerges as the structural leader of the AI drug discovery category or faces the challenge that other AI biotechs have not yet overcome — the gap between AI-driven discovery efficiency and AI-driven clinical success.

The cautionary cases that anchor the AI drug discovery execution risk include: (a) Recursion REC-994 and REC-2282 clinical failures, (b) Exscientia clinical-stage compound failures prior to Recursion merger, (c) earlier AI-discovered compound failures including DSP-1181 (Exscientia-Sumitomo Dainippon, abandoned 2021 after Phase 1), (d) BenevolentAI BEN-2293 (atopic dermatitis, Phase 2a missed primary endpoint), (e) AbCellera Biologics' bavibalimumab (autoimmune disease, terminated 2024), plus (f) the broader industry pattern that AI-discovered drugs to date have not reached FDA approval. The cumulative cautionary signal: AI-driven discovery has reduced discovery-stage cycle time (Insilico Rentosertib went from target identification to Phase 1 in approximately 30 months versus 4-5 years traditional) and has improved discovery-to-clinical-candidate hit rates (some AI biotechs report 2-3× higher hit rates), but the AI-discovered compounds still face the structural clinical attrition risk inherent to pharmaceutical R&D.

Patent and Lifecycle Management

(Not directly applicable to traditional drug pipeline patent cliff analysis; the parallel consideration is AI platform IP and clinical-stage compound IP protection).

AI platform IP includes: AlphaFold (Google DeepMind, with Apache 2.0 license for AlphaFold 2 academic use plus proprietary AlphaFold 3 commercial use through Isomorphic Labs partnerships), Meta ESM (open-source protein language model), NVIDIA BioNeMo (proprietary platform with commercial licensing), Schrödinger LiveDesign (proprietary software platform), Insilico Pharma.AI suite (proprietary), Recursion phenotypic screening platform (proprietary), Cradle.bio (proprietary), Genesis GEMS (proprietary), plus emerging open-source models (RFdiffusion from Baker Lab, plus other academic spin-outs).

AI-discovered compound IP protection follows standard pharmaceutical patent strategy — composition of matter patents, formulation patents, method-of-use patents. Insilico Medicine holds composition-of-matter patents on Rentosertib (ISM001-055) plus extensive related patent estate. Recursion-Exscientia combined patent portfolio extends across multiple therapeutic areas. The structural consideration for AI drug discovery: whether AI-discovered compounds face challenges to patent validity (USPTO has not yet ruled on AI inventorship in drug discovery contexts, though the broader USPTO ruling on AI inventorship for software has restricted AI-as-named-inventor; pharmaceutical patent practice continues to identify human inventors with AI-assistance).

Challenges & Opportunities

Key Challenges

Clinical attrition risk and Phase 3 success rate

AI-discovered drugs face the same structural clinical attrition risk as traditionally discovered drugs. Industry-wide Phase 1 to Phase 2 success rate approximately 40-50 percent; Phase 2 to Phase 3 success rate approximately 30-35 percent; Phase 3 to FDA approval success rate approximately 50-65 percent — implying cumulative clinical-to-approval success rate of 5-12 percent for compounds entering Phase 1. The structural signal: AI-driven discovery acceleration (reducing discovery-stage cycle time) does not translate to materially reduced clinical attrition risk. Recursion REC-994 and REC-2282 failures plus earlier Exscientia DSP-1181 failure plus BenevolentAI BEN-2293 failure illustrate the persistent clinical attrition pattern.

Zero FDA approvals for AI-discovered drugs as of mid-2026

No AI-discovered drug has yet reached FDA approval. While discovery-stage efficiency improvements are real (Insilico Rentosertib 30-month target-to-Phase-1 versus 4-5 years traditional), the lack of FDA approval validation creates structural ROI uncertainty for big pharma adoption plus AI biotech valuation. The first FDA approval (forecast 2027-2029 base case, anchored by Insilico Rentosertib Phase III progression plus emerging Isomorphic oncology trials plus Schrödinger BMS pipeline progression plus Relay Therapeutics RLY-2608) is the principal value-realisation milestone — but timing uncertainty creates risk for current pipeline investments.

Big tech-pharma partnerships potentially bypassing pure-play AI biotech

Eli Lilly-NVIDIA, Roche-NVIDIA, Sanofi-OpenAI-Formation Bio, Microsoft-Moderna, Amazon Web Services pharma deployments, plus emerging Anthropic-pharma partnerships collectively represent structural alternative to pure-play AI biotech intermediaries. The forward implication: AI biotechs face competition not only from each other and traditional pharma but from big-tech-direct-to-pharma offerings that bypass the AI biotech intermediary entirely. The strategic question: does the AI biotech model (proprietary pipeline plus platform) sustain through 2030, or does consolidation accelerate into big-pharma-led plus big-tech-anchored deployment?

Ethical, regulatory, and IP uncertainty around AI-generated compounds

The FDA, EMA, MHRA, and NMPA have not yet established specific frameworks for AI-discovered drug regulatory submission requirements. USPTO has restricted AI-as-named-inventor in patents (the Thaler v Vidal decision plus subsequent USPTO guidance) creating uncertainty for AI-discovered compound patent strategies. EU AI Act classification of healthcare AI as "high-risk" creates compliance complexity. The structural risk: regulatory or IP uncertainty constrains commercial deployment timelines or creates structural friction for AI drug discovery investment.

Key Opportunities

First FDA-approved AI-discovered drug as inflection-point opportunity

The first FDA-approved AI-discovered drug (forecast 2027-2029 base case) creates a category-validation inflection point. Pipeline candidates positioned for first approval include: Insilico Rentosertib (Phase IIa positive June 2025, Phase IIb 2025-2026, potential Phase III initiation 2027-2028, FDA approval forecast 2029-2031), Relay Therapeutics RLY-2608 (Phase 3 oncology, potential FDA approval 2027-2028 if Phase 3 succeeds), Schrödinger HIF-2α (BMS partnership, potential approval 2028-2030), plus emerging Isomorphic Labs Phase 1 oncology candidates entering clinical 2025-2026. The first approval generates structural commercial-validation effect — accelerating pharma adoption plus AI biotech valuation multiples plus partnership transaction value across the category.

Big pharma AI infrastructure investment scaling

Big pharma direct AI infrastructure investment growing from approximately US$0.5 billion in 2024 to approximately US$3-5 billion per year by 2030 — driven by NVIDIA-anchored supercomputers (Eli Lilly, Roche, plus emerging Pfizer, GSK, BMS, AstraZeneca, Novartis, Sanofi, Moderna), dedicated AI laboratories, plus emerging proprietary foundation model development. The cumulative big pharma AI infrastructure investment 2025-2030 is forecast at approximately US$15-22 billion — providing structural demand-side support for the AI platform and services categories.

Pharma partnership transaction value growth

Cumulative AI biotech-pharma partnership deal value growing from approximately US$8 billion in 2025 to approximately US$45-60 billion by 2030 — driven by multi-target discovery collaborations, single-asset licenses post Phase 1/2 positive readouts, platform licensing, plus emerging acquisitions. Schrödinger-BMS US$2.7 billion provides the multi-target deal template. The forward implication: as AI-discovered drugs progress to mid-stage clinical trials, single-asset license deal values are forecast to materially increase from approximately US$100-500 million per asset in 2024 to approximately US$500 million-2 billion per asset by 2030 for Phase II positive readout compounds.

Chinese AI drug discovery emerging as parallel deployment

China's AI drug discovery market grows from approximately 9 percent of 2025 global value to approximately 12 percent by 2032. Insilico Medicine's Hong Kong + Mainland presence (Rentosertib Phase IIa GENESIS-IPF conducted in China) plus XtalPi (Shenzhen) plus emerging Tencent and Alibaba healthcare AI plus state-funded AI drug discovery initiatives plus Chinese 14th Five-Year Plan strategic emphasis collectively position China as the principal non-Western AI drug discovery growth opportunity.

Key Policies & Regulatory Environment

FDA Predetermined Change Control Plan (PCCP) and AI/ML Framework

The FDA "Predetermined Change Control Plan" (PCCP) guidance provides regulatory pathway clarity for AI-enabled clinical decision support but does not yet directly address AI-discovered drugs. The FDA approval pathway for AI-discovered drugs remains the standard pharmaceutical pathway: IND, Phase 1, Phase 2, Phase 3, BLA/NDA, post-marketing surveillance. The structural implication: AI-discovered drugs face the same regulatory pathway as traditionally discovered drugs — meaning that AI-driven discovery efficiency gains do not translate to reduced regulatory timelines.

EU AI Act (In Force August 2024)

The EU AI Act classifies AI applications in healthcare as "high-risk" requiring conformity assessment, but does not specifically address drug discovery applications. The Act applies to AI-enabled medical devices (separate from drug discovery) and may create indirect regulatory friction for AI biotech operations in the EU. The structural implication: EU regulatory complexity may shift AI biotech operations toward US, UK, or Asia-Pacific deployment.

EMA and MHRA AI/ML Guidance for Medicinal Products

The EMA Q&A on AI/ML applications to medicinal product development (published 2024) provides regulatory framework clarity for AI-assisted clinical trial design, patient stratification, plus emerging AI-enhanced regulatory submissions. UK MHRA emerging AI guidance plus the broader MHRA AI Sandbox programme provide supportive regulatory environment for AI biotech operations. The structural implication: UK plus EU regulatory environment is supportive but unsettled.

USPTO AI Inventorship and Patent Strategy

The USPTO ruling on AI inventorship (Thaler v Vidal, plus subsequent USPTO guidance 2024) has restricted AI-as-named-inventor in patents — meaning AI-assisted discoveries must identify human inventors. The structural implication for AI drug discovery: composition-of-matter patents for AI-discovered compounds must identify human chemists or biologists as inventors, with AI as a research tool rather than co-inventor. Patent strategy for AI-discovered compounds remains within standard pharmaceutical patent practice.

China NMPA AI Medical Device and Emerging AI Drug Discovery Framework

The Chinese NMPA AI medical device review guidance (published 2022, updated 2024) provides regulatory framework for AI-enabled medical devices. Emerging AI drug discovery framework under development through 2025-2026. The 14th Five-Year Plan strategic emphasis on AI healthcare plus emerging dedicated state funding for AI drug discovery position China as the principal non-Western AI drug discovery growth opportunity. Insilico Medicine's Hong Kong + Mainland presence (Rentosertib Phase IIa GENESIS-IPF in China) demonstrates the structural pathway.

US AI Executive Orders and AI Safety Framework

US Executive Orders on AI (Biden October 2023 EO 14110, plus emerging Trump administration AI policy through 2025-2026) plus the NIST AI Risk Management Framework provide broader AI policy context. The structural implication for AI drug discovery: federal AI policy does not specifically address drug discovery but creates broader regulatory environment expectations around AI safety, bias, transparency, plus emerging AI-system accountability. The December 2025 EO on AI federal preemption of state AI laws may simplify AI biotech operations.

Future Outlook

The global AI drug discovery market is positioned for sustained 25–27 percent CAGR through 2032, reaching approximately US$28 billion in value with approximately 270 active clinical-stage AI-discovered assets. The market has crossed a structural inflection from technology promise to clinical proof in 2025 — anchored by Insilico Medicine's Rentosertib Phase IIa Nature Medicine publication, Isomorphic Labs preparing first oncology trials backed by US$600 million in Thrive Capital-led funding plus US$3 billion in Novartis-Eli Lilly partnerships, and Eli Lilly-NVIDIA building pharma's most powerful supercomputer — but Recursion's REC-994 and REC-2282 clinical failures plus the multi-year ROI uncertainty illustrate that AI-driven drug discovery remains a high-risk frontier rather than a proven commercial pathway. The forecast structure is three-phased: a 2025–2027 clinical-validation phase (33–46 percent annual value growth) anchored by Insilico Phase IIb progression plus Isomorphic first trials plus continued partnership transactions, a 2028–2030 first-approval phase (21–30 percent annual value growth) where the first AI-discovered drug FDA approval validates or challenges the category, and a 2031–2032 scaling phase (15–20 percent annual value growth) as approved AI-discovered drugs scale commercially.

The competitive structure is forecast to evolve from the current AI biotech-led pattern (35 percent of 2025 value share) toward a more diversified structure with big pharma internal AI plus big tech-pharma partnerships gaining share. AI biotechs with proprietary pipelines maintain approximately 38 percent share by 2032 — driven by Insilico Medicine's clinical-stage leadership, Isomorphic Labs' AlphaFold 3 commercialisation, Schrödinger's physics-plus-ML platform maturity, plus emerging next-generation AI biotechs. Big pharma internal AI R&D grows from 18 percent to approximately 20 percent share by 2032. AI platform/SaaS for pharma grows from 22 percent to 24 percent — anchored by Schrödinger plus NVIDIA BioNeMo plus emerging clinical-trial AI specialists. Big tech entry into life sciences (OpenAI, Anthropic, Google, Microsoft, AWS) collectively represents approximately 3 percent of 2032 value share — small absolute but structurally important for platform-and-infrastructure layer.

The geographic structure remains US-dominated through 2032 — approximately 52 percent of 2032 market value (versus 56 percent in 2025) anchored by US-headquartered big pharma plus the deepest AI biotech ecosystem. UK grows to approximately 15 percent share (versus 14 percent in 2025) driven by Isomorphic Labs commercialisation plus BenevolentAI plus emerging Cradle.bio UK presence. China grows to approximately 12 percent share (versus 9 percent in 2025) driven by Insilico Medicine's Hong Kong + Mainland presence plus XtalPi plus emerging Tencent and Alibaba healthcare AI plus state-funded AI drug discovery initiatives.

The therapeutic area mix shifts toward oncology (forecast 85 active clinical-stage AI-discovered assets by 2030, up from 29 in 2025) plus neurology and neurodegeneration (40 vs 13) plus fibrosis (20 vs 7). The expansion is driven by both the structural unmet need (oncology with US$25+ billion top-drug revenue potential, neurology with limited current AI biotech competition, fibrosis with high unmet need post-Insilico Rentosertib proof-of-concept) plus the comparative AI applicability (well-characterised cancer biology plus emerging neurology data resources plus growing fibrosis biomarker understanding). The combined oncology plus neurology plus fibrosis represent approximately 60 percent of 2025 active assets and approximately 67 percent of forecast 2030 assets.

The technology approach mix shifts toward generative AI (forecast 37 percent of 2032 value share, up from 32 percent in 2025) plus foundation models (forecast 20 percent, up from 18 percent). Physics-based plus machine learning hybrid maintains approximately 20 percent share — sustaining Schrödinger's structural commercial position. The strategic implication: generative AI plus foundation models converge into the infrastructure layer that other AI approaches build on — similar to how transformer-style architectures became infrastructure for general AI applications.

The first FDA approval of an AI-discovered drug (forecast 2027–2029 base case) is the principal value-realisation inflection. Candidate compounds positioned for first approval include: Insilico Rentosertib (Phase IIa positive June 2025, Phase IIb 2025-2026, potential Phase III initiation 2027-2028, FDA approval forecast 2029-2031), Relay Therapeutics RLY-2608 (Phase 3 oncology, potential FDA approval 2027-2028 if Phase 3 succeeds), Schrödinger HIF-2α (BMS partnership, potential approval 2028-2030), plus emerging Isomorphic Labs Phase 1 oncology candidates entering clinical 2025-2026 with potential approval 2029-2031. The first approval generates structural commercial-validation effect — accelerating pharma adoption plus AI biotech valuation multiples plus partnership transaction value across the category.

The principal risk to the outlook is sustained clinical attrition without FDA approval through 2028–2030 that materially compresses AI drug discovery valuations and pharma adoption. If Insilico Rentosertib Phase III fails or stalls, plus Isomorphic Labs Phase 1 readouts disappoint, plus Recursion remaining pipeline faces additional discontinuations, plus Schrödinger BMS pipeline fails to progress — the cumulative cautionary signal could materially compress AI biotech valuations (similar to the 2010s genomics boom-and-correction cycle), accelerate consolidation, and slow big pharma AI infrastructure investment. The mitigation pathway depends on sustained clinical success across the leading AI biotech assets — making the 2026–2028 period the structural validation moment for the category.

The secondary risk is big tech entry that materially disintermediates pure-play AI biotechs. Eli Lilly-NVIDIA pharma's most powerful supercomputer plus Roche-NVIDIA AI factory plus Sanofi-OpenAI-Formation Bio plus emerging partnerships collectively represent structural alternative to pure-play AI biotech intermediaries. The forward implication: AI biotechs that combine proprietary clinical-stage pipeline (Insilico, Recursion, Relay, Isomorphic) plus differentiated platform (Schrödinger physics-plus-ML, Genesis geometric deep learning, Generate Biomedicines protein therapeutics) sustain through 2030 — but pure-play platform companies face competitive pressure from big tech-pharma direct partnerships.

For tailored support and detailed market analysis, see our offerings on Services or Contact Us.