Isomorphic Labs has secured $5 billion in a Series C financing round, marking the largest private capital raise in the history of biotechnology. The London-based company, founded as an Alphabet spinout by Google DeepMind co-founder and Nobel laureate Sir Demis Hassabis, completed the oversubscribed round just six months after closing a $2.1 billion Series B. The new capital infuses the computational drug designer with unprecedented financial firepower as it prepares to advance its first internally developed, computationally generated drug candidates into human clinical trials.
The financing was co-led by Greenoaks and Sequoia Capital, alongside Andreessen Horowitz (a16z) and several new institutional participants. Alphabet reinvested heavily alongside returning backers including Thrive Capital, MGX, GV, CapitalG, Temasek, and the UK Sovereign AI Fund. While the company did not formally publish its post-money valuation, market reporting values the enterprise at approximately $50 billion. That figure places a five-year-old venture that has not yet completed a Phase 1 human trial on par with century-old pharmaceutical giants such as Bayer and Biogen.
The capital will primarily underwrite the expansion of IsoDDE (Isomorphic Drug Design Engine)—the company’s proprietary generative biology and physics-simulation architecture—and fund the capital-intensive clinical trials needed to prove whether software-designed molecules can safely treat real human illnesses.
"Our drug design engine is generating molecules with unprecedented precision and speed," Hassabis stated upon the round's closing. "This round enables us to accelerate our progress: building foundational biology models, deepening our systemic disease understanding, and driving our first therapeutic programs towards clinical trials to deliver tangible impact for patients."
The injection of capital moves artificial intelligence in medicine out of academic theory and into human pharmacology. To understand why institutional investors committed $5 billion to an enterprise without commercialized drugs, one must examine how the company's underlying computational biology works, how it attempts to dismantle the historical cost curve of pharmaceutical research, and what must happen when algorithmically generated chemistry encounters human biology.
Anatomy of the $5 Billion Bet
Biotechnology venture funding follows well-defined conventions. Traditional Series C rounds typically raise between $100 million and $250 million, capital intended to steer one or two specific drug candidates through mid-stage human testing. Isomorphic Labs bypassed those conventions entirely.
By raising $5 billion on top of May’s $2.1 billion tranche and an earlier $600 million round, the company has accumulated over $7.7 billion in private capital within two years. Alphabet maintains overall corporate control, retaining more than 75% of shares and voting rights, while syndicating risk and validating the platform through global venture capital firms.
| Metric | Detail | Strategic Context |
|---|---|---|
| Series C Size | $5.0 Billion | Largest single venture round in biotech history |
| Total Capital Raised | ~$7.7 Billion | Raised within a 24-month operational window |
| Reported Valuation | ~$50 Billion | Comparable to established global pharmaceutical companies |
| Lead Investors | Greenoaks, Sequoia Capital | Traditional top-tier Silicon Valley tech growth investors |
| Corporate Controller | Alphabet Inc. (75%+ voting equity) | DeepMind lineage coupled with external venture autonomy |
| Core Capital Allocation | IsoDDE expansion & Phase 1/2 clinical build-out | Building translational labs and clinical operations |
Silicon Valley tech firms, rather than traditional life-science venture capitalists, drove the round. Firms like Sequoia and Greenoaks evaluate Isomorphic Labs not as a single-asset biotech company subject to binary regulatory failure, but as a foundational computing platform. In software economics, once a core engine is trained, the marginal cost of applying that model to an incremental problem drops significantly. If Isomorphic can demonstrate that IsoDDE systematically designs effective therapeutic molecules across multiple disease families, it will function as an operating system for pharmacology.
The massive scale of this Isomorphic Labs funding also reflects the operational cost of moving computational leads into physical patients. Running clinical trials requires far more cash than leasing high-performance compute clusters. Clinical development demands international trial site management, regulatory filings with the US Food and Drug Administration (FDA) and the European Medicines Agency (EMA), specialized contract manufacturing, and extensive patient monitoring networks.
By amassing a multi-billion-dollar war chest, the company eliminates the need to license out its most promising internal oncology and immunology drug programs early to survive, allowing it to retain full downstream commercial rights.
From Protein Sequences to the Nobel Prize
The scientific foundation of Isomorphic Labs traces directly to the long-standing computational challenge known as the protein-folding problem.
Proteins are the workhorses of biological life. They make up cell receptors, catalyze metabolic reactions, form muscle fibers, and dispatch immune responses. Every protein begins its existence as an unbranched linear chain of chemical building blocks known as amino acids, linked together like letters forming an intricate sentence. The sequence of these amino acids is directly dictated by an organism's genetic code.
Linear chains do not remain flat ribbons. Within microseconds of synthesis inside a cell, electrostatic charges, hydrogen bonds, hydrophobic interactions, and van der Waals forces compel the amino acid chain to contort into a compact, three-dimensional geometry.
Linear Amino Acid Sequence (1D Code)
│
▼ [Physical Folding Dynamics: Hydrophobic collapse, H-bonds]
Complex Native Tertiary Structure (3D Architecture)
│
▼ [Quaternary Assemblies & Ligand Interactions]
Functional Biological Action / Disease Mechanism
This final 3D architecture determines biological function. If an enzyme possesses a deep hydrophobic crevice of a specific shape, a matching metabolic molecule can settle inside it and undergo a chemical reaction. If an aberrant cancer protein develops an atypical surface structure, it can hook onto an adjacent receptor and broadcast an uncontrollable signal for cell proliferation.
Determining the precise 3D configuration of a single protein previously required months or years of painstaking experimental work. Scientists relied on methods like X-ray crystallography—which requires coaxing fragile proteins into solid crystals—or cryogenic electron microscopy (cryo-EM). Thousands of crucial human proteins resisted crystallization entirely, leaving vast swaths of structural biology opaque.
In 2020, Google DeepMind altered this dynamic during the Critical Assessment of Structure Prediction (CASP14) competition. Its neural network, AlphaFold 2, predicted protein structures from primary amino acid sequences with accuracy rivaling physical experiments. The system operated on a deep learning architecture known as an Evoformer, which combined evolutionary sequence alignment data with spatial geometric transformers to predict the exact physical coordinates of individual atoms.
By 2024, Demis Hassabis and his DeepMind colleague John Jumper received the Nobel Prize in Chemistry for their work on AlphaFold. Yet, while AlphaFold mapped the structural universe of nearly every known protein, static structure prediction alone does not generate new medicines.
A static snapshot of a protein is analogous to a photograph of an automobile engine: it reveals where the parts sit when idle, but it does not explain how the pistons operate under internal combustion, nor does it automatically reveal the blueprint for a custom part that can halt the engine without damaging the chassis. That operational transition is why Alphabet spun out Isomorphic Labs in late 2021.
Inside IsoDDE: How Simulation AI Designs Therapeutics
Predicting a native protein's static shape is step zero. Designing a drug requires solving a far more complex challenge: predicting dynamic molecular recognition and binding mechanics between entirely new synthetic chemical entities and biological targets.
Isomorphic Labs approaches this challenge through its proprietary platform, IsoDDE (Isomorphic Drug Design Engine). Built as a unified suite of simulation AI models, IsoDDE expands upon the structural foundations laid down by AlphaFold 3.
AlphaFold 3 moved beyond single proteins to predict complex molecular assemblies, including:
- Proteins bound to nucleic acids (DNA and RNA).
- Post-translational modifications (such as phosphorylated amino acids).
- Small-molecule chemical compounds (ligands) bound to protein receptors.
- Co-factor ions (zinc, iron, magnesium) essential for biological function.
IsoDDE pushes these capabilities into continuous simulation and targeted molecular generation. The platform operates across three integrated computational phases.
┌──────────────────────────────────────────────────────────┐
│ PHASE 1: TARGET MAPPING │
│ • High-resolution dynamic pocket prediction │
│ • Discovery of transient / allosteric binding sites │
│ • Modeling structural changes induced by ligands │
└────────────────────────────┬─────────────────────────────┘
│
▼
┌──────────────────────────────────────────────────────────┐
│ PHASE 2: DE NOVO GENERATIVE DESIGN │
│ • Multi-parameter spatial molecular assembly │
│ • Direct optimization of free energy (ΔG binding) │
│ • Synthetic accessibility filtering (retro-synthesis) │
└────────────────────────────┬─────────────────────────────┘
│
▼
┌──────────────────────────────────────────────────────────┐
│ PHASE 3: ADMET PHARMACOKINETICS │
│ • In silico modeling of metabolic liver clearance │
│ • Cytochrome P450 interaction & hERG cardiac safety │
│ • Membrane permeability & blood-brain barrier transport │
└──────────────────────────────────────────────────────────┘
1. Identifying the Hidden Target
Traditional drug discovery typically targets active catalytic sites—the obvious physical pockets where an enzyme performs its primary chemical task. Many critical disease drivers, particularly in oncology, lack these defined pockets. They present smooth, featureless molecular surfaces that medicinal chemists classify as "undruggable."
Proteins are not rigid plastic sculptures; they are flexible, pulsating physical objects governed by thermodynamic fluctuations. IsoDDE models these dynamics, identifying cryptic or transient binding pockets. These are functional openings that appear only when the protein shifts into specific physiological configurations.
The software can also identify allosteric sites—secondary regulatory pockets located away from the active center. Squeezing a synthetic molecule into an allosteric pocket can warp the shape of the entire protein, shutting down an oncogenic signal without needing to attack the primary active site.
2. Generative Small-Molecule Design
Once a structural vulnerability is mapped, IsoDDE does not simply search a static database of pre-existing chemical compounds. It uses generative neural networks to build new small molecules atom by atom.
The algorithmic engine must solve an optimization problem balancing multiple physical variables:
- Steric Fit: The synthetic compound’s physical volume must slot precisely into the target pocket without colliding with surrounding residues.
- Electrostatic Complementarity: Positively charged nitrogen groups on the drug must align with negatively charged aspartic or glutamic acids on the protein; hydrophobic rings must align with lipophilic pockets.
- Binding Affinity ($\Delta G$): The thermodynamic binding free energy must be sufficiently negative to ensure the drug molecule binds tightly to its target and remains attached long enough to alter disease signaling.
- Selectivity: The compound must target only the intended pathological protein. If a drug designed to inhibit the cancer kinase protein CDK20 inadvertently binds to dozens of related kinases needed for healthy heart or liver function, the resulting toxicity will halt development during clinical trials.
3. De Novo Synthesis Planning
A computer model can easily configure a molecular structure that binds tightly in simulation but violates the laws of chemical synthesis. If organic chemists cannot physically manufacture the compound in a laboratory, the computational design is worthless.
IsoDDE integrates synthetic accessibility filters directly into its generative loops. The engine pairs molecular generation with automated retro-synthesis algorithms. As it drafts a candidate molecule, the system checks whether the proposed structure can be assembled from readily available chemical building blocks using documented organic synthesis reactions.
This connection prevents the software from outputting chemical theoretical dead-ends, ensuring that designs can immediately move from computational files to automated physical chemistry benches.
Dismantling Eroom's Law: The Economic Imperative
To understand why investors committed $5 billion during this round of Isomorphic Labs funding, consider the structural economic distress plaguing modern drug development.
For half a century, the semiconductor industry followed Moore's Law: computing power doubled roughly every two years as costs plummeted. The pharmaceutical industry, conversely, has been defined by Eroom's Law ("Moore" spelled backward).
Coined by industry executive Jack Scannell in 2012, Eroom’s Law observes that the inflation-adjusted cost of developing a newly approved therapeutic doubles approximately every nine years.
MOORE'S LAW (Computing):
Transistors per chip double every ~2 years ──► Efficiency increases exponentially
EROOM'S LAW (Biotech):
R&D cost per approved drug doubles every ~9 years ──► Efficiency decreases exponentially
Today, bringing a single new molecular entity to market costs between $1.5 billion and $2.6 billion, taking 10 to 15 years from initial target discovery to commercial launch. Over 90% of candidate therapies that enter Phase 1 human trials fail.
The drivers of this financial drag sit in the early, highly speculative phases of discovery:
- High-Throughput Screening Inefficiencies: Traditional biopharma searches for new drugs by physically screening massive libraries containing millions of random chemical compounds against an isolated target protein. This process is expensive, takes months, and yields hits with marginal chemical potency that require years of iterative medicinal chemistry to optimize.
- High-Rate Preclinical Attrition: Over 80% of targets that appear promising in cellular cultures fail in animal disease models due to poor biological stability, unexpected off-target toxicity, or poor cell membrane penetration.
- Late-Stage Clinical Failure: The most catastrophic losses happen in Phase 2 and Phase 3 clinical trials, where pharmaceutical sponsors spend hundreds of millions of dollars testing a drug in hundreds of human patients, only to find the therapy lacks meaningful efficacy or triggers unanticipated side effects.
| Metric | Traditional Biotech Discovery | AI-First Engine (IsoDDE Target) |
|---|---|---|
| Hit Identification Timelines | 2 to 4 Years | 3 to 6 Months |
| Physical Compounds Synthesized | 5,000 to 10,000 molecules | 200 to 500 molecules |
| Average Cost to Preclinical Candidate | $30M – $70M | $5M – $15M |
| Target Space | Previously "drugged" protein classes | Expands to "undruggable" & novel pockets |
| Primary Failure Driver | Poor target selectivity & off-target toxicity | Addressed in silico before physical synthesis |
IsoDDE targets the front half of this discovery funnel. By transitioning from physical, blind high-throughput screening to predictive generative chemistry, early discovery timelines shrink from years to months. Instead of synthesizing thousands of experimental molecules in physical chemistry laboratories to find a lead, researchers can synthesize a few hundred computationally pre-filtered candidates.
If computational simulation raises human clinical success rates by even 10 to 15 percentage points, the economics of medicine change drastically. Drug development transitions from a speculative lottery reliant on empirical trial-and-error into a systematic engineering discipline.
Commercial Strategy: Dual Pipeline and Pharma Alliances
Isomorphic Labs is not structured as an enterprise software vendor selling software subscriptions to third-party labs.
"We're not selling access to the models, we're not selling tokens or API access," Max Jaderberg, President of Isomorphic Labs, clarified at Bloomberg Tech in London.
The company employs a hybrid corporate model. It operates a proprietary, internal drug discovery pipeline while concurrently signing high-value partnerships with elite global pharmaceutical developers.
ISOMORPHIC LABS (IsoDDE)
│
┌─────────────────────┴─────────────────────┐
▼ ▼
INTERNAL PIPELINE STRATEGIC ALLIANCES
• 100% Owned Assets • Eli Lilly ($1.7B+ milestones)
• Focus: Oncology & Immunology • Novartis ($1.2B+ milestones)
• High Risk / Maximum Return • Johnson & Johnson Collaboration
• Retains downstream global sales • Upfront cash + high royalties
The company's external partnerships already generate non-dilutive capital and provide access to deep disease-specific biological datasets:
The Eli Lilly Alliance
Signed in January 2024, this partnership focuses on discovering small-molecule therapeutics against multiple undisclosed targets. Lilly provided $45 million in immediate upfront cash, with Isomorphic eligible to earn up to $1.7 billion in milestone payments, alongside tiered royalties ranging up to low double digits on net commercial sales.
The Novartis Alliance
Also secured in January 2024 and expanded in 2025, Novartis committed $37.5 million upfront, funding select shared research costs, with up to $1.2 billion in performance milestones plus tiered royalties. The scope covers small-molecule development against three challenging oncology targets.
The Johnson & Johnson Expansion
In early 2026, Johnson & Johnson entered into an undisclosed collaboration agreement with Isomorphic to leverage IsoDDE for complex immune-system disorders.
These alliances demonstrate how established pharmaceutical manufacturers view Isomorphic's platform. The major drugmakers cover upfront and early developmental costs in exchange for access to compounds engineered against targets their internal discovery teams could not conquer. In return, Isomorphic receives immediate working cash and billion-dollar milestone ceilings.
Simultaneously, the internal pipeline remains Isomorphic’s primary driver of enterprise value. Focused on difficult targets in oncology and autoimmune disease, internal assets are designed, validated, and developed within the company.
The $5 billion Series C ensures Isomorphic does not need to sell off or out-license these internal assets when they enter human testing, preserving the multi-billion-dollar market value of any successful blockbuster drug for itself and Alphabet shareholders.
The Preclinical Gauntlet: The ADMET Filter
Before any computationally designed molecule can enter a human Phase 1 safety trial, it must survive the rigorous preclinical gauntlet known as ADMET.
A molecule may exhibit atomic binding affinity toward an isolated target inside a digital simulation. Yet when introduced into a living organism, complex biological defenses quickly mobilize to neutralize foreign chemistry. ADMET outlines the biological gauntlet every prospective drug must run:
[ INGESTION / INJECTION ]
│
▼
ABSORPTION ──► Can the drug traverse the intestinal epithelium or cell walls?
│
▼
DISTRIBUTION ──► Does the drug land in healthy tissues or stay concentrated at the disease site?
│
▼
METABOLISM ──► Does the liver rapidly degrade the molecule into inactive or toxic metabolites?
│
▼
EXCRETION ──► Are the drug and its breakdown products cleared safely by the kidneys or bile?
│
▼
TOXICITY ──► Does the molecule trigger liver toxicity or disrupt cardiac ion channels (hERG)?
Absorption
The molecule must reach systemic circulation. If designed as an oral pill, it must survive the stomach's hydrochloric acid, pass through the intestinal wall, and resist destruction by metabolic enzymes lining the digestive tract.
IsoDDE deploys specialized neural networks trained on historical pharmacokinetics data to forecast human intestinal permeability and oral bioavailability directly from a molecular design's topological structure.
Distribution
Once circulating in the bloodstream, the compound must navigate safely through physiological compartments. If a drug binds excessively to human serum albumin—a common blood protein—it becomes trapped in the bloodstream and cannot reach diseased tissue.
In oncology, the drug must penetrate deep into dense, hypovascular tumor microenvironments. For neurodegenerative indications, the candidate must navigate the blood-brain barrier: a tightly packed layer of endothelial cells that blocks nearly all foreign synthetic molecules from entering brain tissue.
Metabolism
The liver serves as the body’s detoxification plant. When a foreign small molecule enters the bloodstream, hepatic enzymes—predominantly the Cytochrome P450 (CYP) family—work to oxidize, hydrolyze, and neutralize it.
If liver enzymes process a compound too quickly, the drug vanishes before it can deliver a therapeutic effect. Conversely, if the drug inhibits these enzymes, it can cause co-administered medications to accumulate to toxic levels.
IsoDDE models prospective binding poses against common Cytochrome P450 variants to ensure designed candidates possess stable half-lives without triggering dangerous drug-drug interactions.
Excretion
The compound and its broken-down byproducts must be cleared safely from the patient's body, primarily filtered through the kidneys into urine or evacuated through biliary secretion. Compounds that linger indefinitely or crystallize within renal tubules trigger serious organ failure.
Toxicity
Toxicity is the primary executioner of clinical-stage drug candidates. The most notorious preclinical safety screen is the hERG (human Ether-à-go-go-Related Gene) assay. The hERG gene codes for a potassium ion channel critical for regulating cardiac electrical rhythm.
If a synthetic molecule binds to this channel, it can disrupt heart repolarization, causing a lethal ventricular arrhythmia known as Torsades de Pointes.
IsoDDE evaluates designs in silico against models of the hERG potassium channel and panels of human liver cells to discard toxic chemotypes long before raw chemicals are mixed in physical reactors.
The Great Clinical Crucible: What Happens Next
The multi-billion-dollar question facing this round of Isomorphic Labs funding is straightforward: will computational optimization translate into clinical success in living human patients?
Over the past decade, first-generation AI drug discovery companies raised billions of dollars on promises of transforming medicine, but experienced mixed results once they entered human clinical testing.
Companies such as Exscientia, BenevolentAI, and Relay Therapeutics proved that computational tools could speed up candidate discovery and design novel chemical structures. However, several of their early assets were halted in Phase 1 or Phase 2 trials after failing to demonstrate meaningful therapeutic efficacy or triggering unforeseen tolerability issues.
Biology is messy, non-linear, and inherently interconnected. Cellular systems are regulated by redundant signaling pathways; when a drug successfully shuts down one disease receptor, the target cell often activates a secondary bypass mechanism to survive.
Isomorphic Labs argues that its platform represents a generational leap over these early efforts. Early platforms typically used traditional machine learning to screen and tweak existing chemical libraries.
IsoDDE, by contrast, operates as a structural biology foundation model. It works from first-principles biophysics, predicting multi-scale atomic assemblies and protein movements across real physical space.
┌────────────────────────────────────────────────────────┐
│ PHASE 1 CLINICAL TRIAL │
│ • Cohort: Small group of healthy human volunteers │
│ • Primary Focus: Human safety, tolerance, and dosage │
│ • Validation: Does the drug clear without toxicity? │
└───────────────────────────┬────────────────────────────┘
│
▼
┌────────────────────────────────────────────────────────┐
│ PHASE 2 CLINICAL TRIAL │
│ • Cohort: 100 to 300 patients with target disease │
│ • Primary Focus: Biological efficacy & proof-of-concept│
│ • Validation: Does the candidate shrink human tumors? │
└───────────────────────────┬────────────────────────────┘
│
▼
┌────────────────────────────────────────────────────────┐
│ PHASE 3 CLINICAL TRIAL │
│ • Cohort: 1,000 to 3,000+ randomized global patients │
│ • Primary Focus: Statistical superiority vs. standard │
│ • Regulatory Goal: FDA & EMA formal market approval │
└────────────────────────────────────────────────────────┘
The clinical journey of Isomorphic's internal oncology and immunology programs will now proceed through three distinct regulatory checkpoints:
- IND Filings (Investigational New Drug): Isomorphic must file detailed preclinical safety packages with regulatory bodies like the FDA. Regulators evaluate non-clinical laboratory data, including toxicology tests in two animal species, stability measurements, and validated chemical manufacturing procedures. The FDA applies the same safety and manufacturing standards regardless of whether a molecule was sketched by a chemist or output by a neural network.
- Phase 1 Human Trials: The clinical trial process begins with testing in a small cohort of healthy human volunteers (or patients with refractory disease in oncology). These trials confirm whether the drug's human pharmacokinetics match IsoDDE's computational predictions, verifying safe systemic absorption without triggering liver, kidney, or cardiac toxicity.
- Phase 2 Proof of Concept: If an asset demonstrates safety, it advances to testing in patients diagnosed with the target pathology. Here, the platform faces its real test: proving that computationally designed molecules can shrink human tumors or suppress autoimmune inflammation better than existing treatments.
By building its own clinical and translational operations, Isomorphic can run tightly integrated feedback loops. Clinical trial data, patient biopsy results, and pharmacodynamic measurements can be fed directly back into IsoDDE to refine its simulation models for future drug iterations.
Compute Infrastructure and Closed-Loop Automation
Software models cannot scale in a biological vacuum. Machine learning models require steady access to high-performance computing infrastructure, balanced by rapid, empirical laboratory validation.
Isomorphic Labs maintains a distinct advantage through its relationship with Alphabet. Building foundational simulation models across multi-molecular complexes requires immense compute. The company trains its models across clusters of Google Cloud’s Tensor Processing Units (TPUs) and high-end Nvidia GPU clusters, scaling training runs that would bankrupt independent biotechs.
At the same time, Hassabis has structured Isomorphic around closed-loop robotic wet labs. High-performance computing creates hypotheses; automated physical chemistry verifies or refutes them.
┌──────────────────────────────────────────────┐
│ IN SILICO (Digital) │
│ IsoDDE generative design, dynamic pocket │
│ simulation, and ADMET property filtering │
└──────────────────────┬───────────────────────┘
│
Compound Coordinates
│
▼
┌──────────────────────────────────────────────┐
│ IN VITRO (Physical) │
│ Automated robotic synthesis, cryo-EM, │
│ binding affinity assays, and cell screens │
└──────────────────────┬───────────────────────┘
│
Real-World Biophysics
│
▼
┌──────────────────────────────────────────────┐
│ ACTIVE RE-TRAINING │
│ Discrepancies between prediction and reality│
│ update model weights and refining IsoDDE │
└──────────────────────────────────────────────┘
When IsoDDE designs a set of promising molecular variations, those coordinates are sent to automated synthesis platforms. Robotic liquid-handling systems synthesize the physical small molecules and run them through high-throughput binding assays, X-ray crystallography, and cryo-EM imaging.
If an experimental molecule binds with slightly less affinity than IsoDDE predicted, the structural and kinetic discrepancies are routed back to the machine learning platform. The model updates its predictive weights based on physical reality.
This closed-loop learning engine ensures that the software continuously incorporates physical biological data, refining its accuracy with every generation of molecules.
Demis Hassabis and Alphabet's Broader Strategy
The timing and scale of this round of Isomorphic Labs funding coincide with a deliberate change in leadership structure within Alphabet.
In August 2026, Sir Demis Hassabis stepped down from day-to-day operational leadership as CEO of Google DeepMind to assume the role of DeepMind Chairman and Alphabet Chief Scientist. This shift allowed Hassabis to redirect his operational focus toward Isomorphic Labs, where he remains full CEO.
For Hassabis, applying artificial intelligence to structural biology has never been a sideline hobby; it has been a lifelong scientific ambition. In public addresses, he has described drug design and human biology as the ultimate domain for artificial intelligence—a complex, multi-scale network problem governed by physical rules that can be decoded using the right computational tools.
From Alphabet’s perspective, Isomorphic Labs represents a major vehicle for long-term enterprise diversification. The search and advertising giant has long backed ambitious life-science ventures through subsidiaries such as Verily and Calico.
Yet Isomorphic holds distinct financial potential. While Verily has concentrated on healthcare data infrastructure, clinical software, and medical devices, Isomorphic targets the pharmaceutical drug market directly.
Global spending on prescription pharmaceuticals surpasses $1.6 trillion annually, with oncology alone accounting for hundreds of billions of dollars. If Isomorphic’s computational engine succeeds in bringing novel oncology and immunology assets through clinical trials, it will transform Alphabet from an enterprise software and digital media company into an essential owner of intellectual property in clinical medicine.
Key Milestones to Watch (2027–2028)
With $5 billion in fresh capital, the conversation surrounding Isomorphic Labs shifts from architectural design to clinical execution.
Over the next 18 to 24 months, the biotechnology and financial sectors will monitor five key milestones to determine whether computational generative medicine can deliver on its promise:
- Initial IND Submissions and Target Unveiling: To date, Isomorphic has kept its internal oncology and immunology pipeline targets confidential. When the company submits its first Investigational New Drug (IND) applications to clinical regulators, it must formally reveal its biological targets, chemical structures, and preclinical safety data.
- Phase 1 Trial Initiation and Safety Profiles: Dosing the first human patients will signal the start of safety and tolerability profiling. The critical question will be whether IsoDDE’s ADMET and toxicity models successfully prevented the off-target complications that historically derail early clinical trials.
- Milestone Triggers in Lilly and Novartis Collaborations: Watch for disclosures indicating that Isomorphic has achieved designated research and lead-selection milestones in its partnered programs. These payouts will serve as third-party validation that pharmaceutical partners are pleased with the quality of molecules coming out of the engine.
- Peer-Reviewed Validation of IsoDDE: DeepMind built trust by publishing AlphaFold's architecture openly in Nature. As IsoDDE continues to develop, researchers will watch for peer-reviewed papers disclosing benchmarks on how accurately the platform models protein-ligand interactions, allosteric sites, and binding free energies compared to physical experiments.
- Expansion of Wet-Lab Infrastructure: The rate at which Isomorphic builds its translational infrastructure, robotics-driven synthesis centers, and clinical operations will reveal how fast the company expects to manufacture and test drug candidates at scale.
The $5 billion Series C provides Isomorphic Labs with the runway needed to bridge the gap between digital simulation and clinical efficacy. If the company succeeds, it will establish a new paradigm for how human therapeutics are conceived, designed, and manufactured—replacing decades of costly trial-and-error with computational models that can tackle previously intractable diseases.
Reference:
- https://cellcog.ai/blog/isomorphic-labs-5-billion-series-c/
- https://www.rdworldonline.com/isomorphic-labs-raises-5-billion-in-biotechs-biggest-venture-round/
- https://techfundingnews.com/deepmind-spinout-isomorphic-labs-lands-5b-series-c-to-take-ai-designed-medicines-into-the-clinic/
- https://allsci.com/news/venture-financing/isomorphic-labs-series-c-raises-usd-5b-to-scale/
- https://www.prnewswire.com/news-releases/isomorphic-labs-announces-5-billion-series-c-funding-302942062.html
- https://pharmaphorum.com/news/isomorphic-labs-raises-5bn-massive-third-round
- https://www.fiercebiotech.com/biotech/isomorphic-labs-raises-5b-series-c-months-after-21b-round
- https://techcrunch.com/2026/11/17/isomorphic-labs-raises-5b-as-it-looks-to-take-drug-candidates-to-clinical-trials/
- https://endpoints.news/alphabets-ai-unit-isomorphic-inks-drug-discovery-deals-with-eli-lilly-novartis-for-up-to-3b/
- https://www.trendingtopics.eu/isomorphic-labs-hassabis-pharma-ai-set-for-5-billion-dollar-mega-round-without-a-single-drug-tested-in-humans/
- https://www.fiercebiotech.com/biotech/alphabets-isomorphic-stacks-two-new-deals-lilly-novartis-worth-nearly-3b-ahead-buzzy-jpm
- https://firstwordpharma.com/story/5816082
- https://fourweekmba.com/ai-isomorphics-pharma-deals-82-5m-upfront-2-9b-milestones/