Logo
sparkle
Subscribe
Logo
Oliver Buchannon
Paulina Gozdzialska

Founder of Life Sciences Digital and Nuna Digital

Claude Science Arrives. Takeda Signs Another AI Platform. Europe Locks In Its AI Timeline for Medical Devices.

Jul 6, 2026

•

5 min read

Claude Science Arrives. Takeda Signs Another AI Platform. Europe Locks In Its AI Timeline for Medical Devices.

Last week we covered BIO 2026 in San Diego, where the industry admitted the real constraint on AI in life sciences is not model capability but the fragmentation of the data and tools researchers work with. This week, Anthropic launched a product built exactly for that problem. Claude Science, released Tuesday in San Francisco, connects more than 60 scientific databases and the tools researchers rely on daily (PubMed, Jupyter, R, cluster access) into a single working environment. Anthropic also announced it will run its own drug discovery programs on neglected diseases, using the same tool. Dario Amodei called it the Claude Code moment for life sciences.

Paulina Gozdzialska
Paulina Gozdzialska
BIO 2026 Opened With Billions in New AI Deals. Data Quality Is the Industry's Real Constraint.

Jun 29, 2026

•

5 min read

BIO 2026 Opened With Billions in New AI Deals. Data Quality Is the Industry's Real Constraint.

Previously we tracked AI-designed molecules entering the clinic and Big Pharma writing nine-figure checks for the data infrastructure that AI runs on, with the conclusion that the edge is no longer the model but who controls the data. Last week the industry gathered at BIO 2026 in San Diego and said the quiet part out loud. The deals kept getting bigger: Insilico opened the convention with a $2.5 billion CNS partnership, NVIDIA shipped an agentic toolkit for life sciences, and Boehringer Ingelheim bought into another AI discovery platform. At the same time, pharma leaders on stage were unusually frank that AI has not yet cracked the harder problems. The thread running through both the optimism and the caution is the same one we covered last week. The constraint that decides who actually gets value from AI is the quality and readiness of the data feeding it.

Paulina Gozdzialska
Paulina Gozdzialska
AI-Designed Molecules Are Entering the Clinic. Merck Commits $510M to AI Protein Design. The Competitive Edge Is the Data.

Jun 22, 2026

•

6 min read

AI-Designed Molecules Are Entering the Clinic. Merck Commits $510M to AI Protein Design. The Competitive Edge Is the Data.

Last week we tracked Big Pharma building AI it owns rather than rents - Astellas, Sanofi, Novartis writing nine-figure checks for AI-native pipelines, and Microsoft signing the largest generative AI deployment in healthcare history with NHS England. This week, the proof of concept moved into patients. Insilico Medicine completed first-in-human dosing of an AI-generated NLRP3 inhibitor - a molecule that originated entirely inside a computational system and is now being evaluated in a Phase I study for CNS inflammation. On the same week, Merck committed up to $510 million to Protillion Biosciences for megascale protein data generation purpose-built to train AI, and Charles River joined Lilly's TuneLab federated platform, adding standardized nonclinical testing to a model that improves with every new data contributor. The through-line across all three stories is the same: the race is no longer about which AI model is most capable — it is about who controls the data infrastructure and the validated workflows that AI runs on. OpenAI this week launched LifeSciBench, a 750-task evaluation framework built by PhD scientists to measure whether AI systems can actually do the work of biological research.

Paulina Gozdzialska
Paulina Gozdzialska
Big Pharma Builds Its Own AI Empire. The NHS Goes All-In on Copilot. Regulators Try to Keep Up.

Jun 15, 2026

•

5 min read

Big Pharma Builds Its Own AI Empire. The NHS Goes All-In on Copilot. Regulators Try to Keep Up.

Last week we tracked Sanofi licensing an AI scientist from Owkin and Collate crossing a near-billion dollar valuation on regulatory document automation. This week the story split into three clear threads. In R&D, Big Pharma kept building AI capability it owns rather than rents: Astellas credited part of a $406 million two-year cost transformation to insourcing clinical operations with AI-assisted tools, Sanofi and Owkin deepened their five-year collaboration to build AI drug discovery agents, and Novartis committed up to $1.4 billion in milestones to Orionis Biosciences' AI-driven molecular glue platform. In distribution, AI moved into the operational core of entire health systems: Microsoft signed the largest generative AI deployment in healthcare history with NHS England, and TCS became a Global Premier Partner in Anthropic's Claude network, rolling Claude out to 50,000 employees across regulated industries including life sciences.

Paulina Gozdzialska
Paulina Gozdzialska
Sanofi Licenses an AI Scientist. Collate Hits $1 Billion on Regulatory Docs. The Governance Gap Is Becoming a Product.

Jun 8, 2026

•

4 min read

Sanofi Licenses an AI Scientist. Collate Hits $1 Billion on Regulatory Docs. The Governance Gap Is Becoming a Product.

Last week we tracked Veeva shipping its agentic architecture and pharma building AI it owns. This week, the deal flow moved further down the stack. Sanofi signed a five-year agreement with Owkin to deploy K Pro, an AI scientist agent, across biopharma development workflows. Collate raised $95 million and crossed a near-billion dollar valuation by cutting regulatory document processing time by up to 90%. Agilent partnered with OpenAI and BCG to embed AI across its instruments, operations, and customer workflows. And Microsoft announced a partnership with Mayo Clinic to build a frontier AI model for healthcare — trained on Mayo's longitudinal clinical data and owned by the clinic, not by Microsoft. Underneath all of it, two stories signal where the industry is heading next: the FDA is now sending deficiency letters over AI explainability, not just performance, and a new analysis shows that AI governance in biotech CMC is fragmented and largely ungoverned.

Paulina Gozdzialska
Paulina Gozdzialska
Veeva Launches Its Agentic Platform. OpenAI Enters Biodefense. Pharma Is Building AI It Owns.

Jun 1, 2026

•

4 min read

Veeva Launches Its Agentic Platform. OpenAI Enters Biodefense. Pharma Is Building AI It Owns.

Last week we tracked foundation models racing to own the pharma vertical — BMS deploying Claude across drug development, Cohere acquiring biopharma-native AI, and Incyte paying $120 million upfront for a model trained exclusively on its own data. This week, the platforms shipped back. Veeva unveiled Falcon, an agentic AI architecture designed to automate clinical, regulatory, and safety workflows, with early adopter access expected by November. OpenAI launched a dedicated life sciences model for governments and vetted researchers working on biodefense and pandemic preparedness. Sanofi disclosed that it built its own internal AI ecosystem rather than adopting SaaS tools, with 80% of its workforce already on the proprietary platform. And Solstice raised $21 million to prove that pharma marketing — one of the last manual-review holdouts in the industry — can go from three months to ten days.

Paulina Gozdzialska
Paulina Gozdzialska
BMS Deploys Claude Across Drug Development. Cohere Buys Into Pharma. Foundation Models Are Racing to Own the Vertical.

May 26, 2026

•

4 min read

BMS Deploys Claude Across Drug Development. Cohere Buys Into Pharma. Foundation Models Are Racing to Own the Vertical.

Last week we tracked Isomorphic Labs closing a $2.1 billion Series B and the FDA holding the line on AI device oversight. This week, the model layer got more specific. Bristol Myers Squibb struck a strategic agreement with Anthropic to embed Claude across research, engineering, and manufacturing workflows, moving explicitly beyond chatbot-level use toward connected data infrastructure. On the same day, Incyte committed $120 million upfront to Genesis Molecular AI to train protein-ligand prediction models on its proprietary experimental data. Cohere acquired Reliant AI, a biopharma-focused AI company with clients at GSK and Ipsen, adding pharma-native capabilities to its North platform.

Paulina Gozdzialska
Paulina Gozdzialska
Isomorphic Labs Raises $2.1 Billion. FDA Holds the Line on AI Device Oversight. AI Drug Discovery Approaches Its First Clinical Test.

May 18, 2026

•

6 min read

Isomorphic Labs Raises $2.1 Billion. FDA Holds the Line on AI Device Oversight. AI Drug Discovery Approaches Its First Clinical Test.

Last week we tracked Roche absorbing the pathology AI layer through its PathAI acquisition, Tempus extending its diagnostic reach into Alzheimer's detection, and Sanofi committing $294 million to AI infrastructure in Toronto. The story was about exits and ownership: pharma deciding to own the AI capabilities it had spent years evaluating. This week, the capital kept moving. Isomorphic Labs closed a $2.1 billion Series B, the largest funding round yet for an AI-native drug discovery platform, with its first IND now targeted for late 2026. The FDA rejected a proposal to ease premarket review requirements for AI diagnostic tools, making clear that regulatory accountability scales with clinical risk regardless of developer track record. And upstream of the headline deals, the physical infrastructure is catching up: Merck KGaA's bioprocessing division grew 16% organically in Q1, tracking the surge in biologics and cell therapy manufacturing that the AI investment wave is driving.

Paulina Gozdzialska
Paulina Gozdzialska
Roche Buys the Pathology AI Layer. Tempus Pivots Beyond Oncology. Pharma's AI Infrastructure Bets Keep Compounding.

May 11, 2026

•

6 min read

Roche Buys the Pathology AI Layer. Tempus Pivots Beyond Oncology. Pharma's AI Infrastructure Bets Keep Compounding.

Last week we tracked Lilly committing $2.25 billion to AI-designed DNA editors and clinical AI raising another $227 million in a single week. The capital story was about scale. This week, the story is about exits. Roche acquired PathAI for up to $1.05 billion, formalising a five-year partnership and signalling that AI pathology is now a strategic asset to own, not a vendor to evaluate. Tempus AI raised $400 million and extended its diagnostic reach beyond oncology with a new Alzheimer's blood biomarker collaboration. And Sanofi committed $294 million to scale its Toronto AI Center of Excellence.

Paulina Gozdzialska
Paulina Gozdzialska
Eli Lilly Bets $2.25 Billion on AI-Designed DNA Editors. Clinical AI Raises $227 Million in One Week.

May 4, 2026

•

5 min read

Eli Lilly Bets $2.25 Billion on AI-Designed DNA Editors. Clinical AI Raises $227 Million in One Week.

Last week we tracked the compliance and infrastructure layer being built beneath the model stack, with Iridius, Boehringer Ingelheim's London AI centre, and J&J moving AI into neurovascular imaging. This week, the deals got larger and more structurally significant. Eli Lilly committed up to $2.25 billion to a 2022-founded AI startup designing recombinase enzymes capable of DNA insertions that CRISPR-based methods cannot achieve at scale. Aidoc and Iterative Health together raised $227 million for clinical AI infrastructure in a single week. Mayo Clinic published results showing an AI model flags pancreatic cancer on routine CT scans an average of 16 months before clinical diagnosis. And new data confirmed what many in clinical operations suspected: more than half of all registered clinical trials are now run entirely outside the United States.

Paulina Gozdzialska
Paulina Gozdzialska
Iridius Builds the Compliance Layer. Boehringer Opens a £150M AI Centre.

Apr 27, 2026

•

5 min read

Iridius Builds the Compliance Layer. Boehringer Opens a £150M AI Centre.

Last week we covered the model layer arriving — GPT-Rosalind from OpenAI, Amazon Bio Discovery. This week, the stack below the models is getting built. Iridius closed an $8.6 million seed round to embed compliance directly into AI workflows across regulated industries, with Accenture putting capital behind the thesis. Boehringer Ingelheim committed £150 million over a decade to a new AI and machine learning centre in London. Johnson & Johnson moved AI into neurovascular imaging through a partnership with Viz.ai. Meanwhile, Europe is catching up on two fronts at once: WHO data confirms AI diagnostics are already deployed across three quarters of EU member states, and a coalition of 40 industry organizations is pushing to build the legislative infrastructure that will let the continent scale what it has started.

Paulina Gozdzialska
Paulina Gozdzialska
OpenAI Releases Its Drug Discovery Model. Amazon Opens the Biology Platform.

Apr 20, 2026

•

6 min read

OpenAI Releases Its Drug Discovery Model. Amazon Opens the Biology Platform.

Last week we tracked capital concentrating fast — Rock Health's Q1 data confirmed $4 billion raised in digital health in a single quarter, sixty percent of it landing in just twelve transactions. The platforms capturing investment are the ones that have already crossed from evaluation to infrastructure. This week, two of the largest technology companies in the world decided they would build that infrastructure themselves.OpenAI launched GPT-Rosalind, its first domain-specific reasoning model, built for drug discovery and named after Rosalind Franklin. The same week, Amazon Web Services launched Amazon Bio Discovery — a platform of more than 40 biological foundation models connected directly to wet lab partners for physical validation of computationally designed candidates. Novo Nordisk anchored OpenAI's pharma ambitions with a company-wide integration spanning drug discovery, manufacturing, supply chain, and commercial operations.Seven weeks ago we covered Anthropic acquiring a small biology team. This week, the products arrived.

Paulina Gozdzialska
Paulina Gozdzialska
Digital Health Capital Is Concentrating Fast. Tempus Embeds in Gilead. AI Enters the Chemotherapy Decision.

Apr 13, 2026

•

5 min read

Digital Health Capital Is Concentrating Fast. Tempus Embeds in Gilead. AI Enters the Chemotherapy Decision.

Last week we tracked Anthropic's $400 million bet on biology and watched AI displace the pharmaceutical rep as the default front door to clinical evidence. This week the market answered with data. Rock Health's Q1 2026 report documented $4 billion raised in digital health in a single quarter — the highest average deal size since 2021, with sixty percent of all capital landing in just twelve transactions. AI companies now account for the majority of large rounds, and the gap between the platforms capturing investment and the rest of the field is widening each quarter. The concentration is not a cyclical pattern. It is a structural one. Tempus AI secured enterprise-wide adoption at Gilead Sciences, converting an existing data agreement into a platform embed across Gilead's full oncology portfolio. Caris Life Sciences moved AI into the chemotherapy decision for the most common form of lung cancer. And PathAI deployed its FDA-cleared digital pathology platform across MedStar Health's laboratory network. Across discovery, diagnostics, and clinical operations, the same dynamic is playing out: AI is moving from evaluation to infrastructure.

Paulina Gozdzialska
Paulina Gozdzialska
Anthropic Bets $400M on Biology. AI Becomes the Front Door to Pharma. Governed AI Is the Next Product Category.

Apr 7, 2026

•

5 min read

Anthropic Bets $400M on Biology. AI Becomes the Front Door to Pharma. Governed AI Is the Next Product Category.

Last week we covered the largest capital commitment to an external AI drug discovery platform in pharma history — Lilly paying $115 million upfront and $2.75 billion in potential value to Insilico Medicine. Caris brought molecular AI into first-line pancreatic cancer therapy selection. Tempus moved from data vendor to embedded development partner inside a Daiichi Sankyo drug program. And Merck confirmed 34,000 monthly users on its internal AI. This week, the story moved upstream. Anthropic paid $400 million for an eight-month-old team of fewer than ten people — all former Genentech researchers, all building biology-native AI. Frontier AI labs are no longer selling tools to life sciences. They are becoming part of it.

Paulina Gozdzialska
Paulina Gozdzialska
Insilico and Lilly Reset the Price of AI Drug Discovery. Oncology AI Enters the Clinical Standard.

Mar 30, 2026

•

4 min read

Insilico and Lilly Reset the Price of AI Drug Discovery. Oncology AI Enters the Clinical Standard.

Last week, the compute got installed. This week, the deals got signed. Insilico Medicine closed a $2.75 billion collaboration with Eli Lilly — $115 million upfront, exclusive worldwide license, multiple therapeutic areas — the largest capital commitment to an external AI drug discovery platform from a top-5 pharma to date. Caris Life Sciences launched an AI tool that guides first-line therapy selection for pancreatic cancer at the molecular level. The tool runs on tissue already collected for standard molecular profiling — no additional biopsy required. Tempus AI moved from data vendor to embedded development partner inside a Daiichi Sankyo drug program. And Merck confirmed 34,000 monthly users on its internal AI, with autonomous agents cutting commercial insight generation time in half. The question was never whether AI would reshape drug discovery. The question was who would pay for it, at what scale, and when. This week, Lilly answered all three.

Paulina Gozdzialska
Paulina Gozdzialska
Roche Locks In the Compute. IQVIA Deploys the Agents. Google Moves Into European Pharmacy.

Mar 23, 2026

•

4 min read

Roche Locks In the Compute. IQVIA Deploys the Agents. Google Moves Into European Pharmacy.

Last week we tracked Veeva absorbing the compliance layer and IQVIA declaring 150 agents in production. This week, the infrastructure war got physical. Roche activated over 3,500 NVIDIA Blackwell GPUs across the U.S. and Europe — the largest declared AI compute footprint in pharma — and called it what it is: a factory. IQVIA launched IQVIA.ai at GTC 2026, an agentic platform already embedded in 19 of the top 20 pharmaceutical companies. Earendil Labs raised $787 million on an AI biologics platform that has already produced over 40 programs. And Google partnered with DocMorris to build an AI-first digital pharmacy for 11 million patients in Europe. The compute is being installed. The agents are running. The capital is following platforms that can prove output, not just potential.

Paulina Gozdzialska
Paulina Gozdzialska
Veeva Buys the Compliance Layer. Agentic AI Goes to Work. Life Sciences Software Is Choosing Its Infrastructure.

Mar 16, 2026

•

5 min read

Veeva Buys the Compliance Layer. Agentic AI Goes to Work. Life Sciences Software Is Choosing Its Infrastructure.

Last week we tracked Google Cloud threading itself through the infrastructure layer of American healthcare. This week, the decisions got more concrete. Veeva acquired the leading compliant conversational AI platform for patients and HCPs, signalling that AI-powered brand engagement is now a commercial requirement, not an experiment. IQVIA confirmed it has 150 AI agents running in production across its operations, with 500 targeted by 2027. Medidata launched a commercial AI companion that compresses clinical trial build timelines from weeks to hours. And NVIDIA committed $2 billion to the compute infrastructure that will run life sciences AI for the next decade. The platforms are being chosen. The infrastructure is being locked in. The window for "we are evaluating AI" is closing.

Paulina Gozdzialska
Paulina Gozdzialska
Google Builds the Healthcare OS. Pharma Hands AI the Keys to Clinical Trials.

Mar 9, 2026

•

5 min read

Google Builds the Healthcare OS. Pharma Hands AI the Keys to Clinical Trials.

Last week we covered the moment AI stopped being a promise and started paying bills. This week, the bills got bigger. Google Cloud showed up at HIMSS 2026 with not one but five major healthcare partnerships announced in the span of days. AstraZeneca's digital health spinout Evinova signed AI deals with three top-20 pharma companies simultaneously. And RadNet dropped €230 million on radiology AI company Gleamer, extending a consolidation wave that is quietly reshaping how imaging gets delivered worldwide. The pattern is clear. AI in life sciences is no longer about individual companies experimenting. It is about platforms being chosen, infrastructure being locked in, and portfolios being assembled through M&A.

Paulina Gozdzialska
Paulina Gozdzialska
AI Stops Being a Promise and Starts Paying the Bills

Mar 2, 2026

•

4 min read

AI Stops Being a Promise and Starts Paying the Bills

Something shifted this week in how the industry is talking about AI. The word "pilot" is nearly gone from the conversation. What replaced it: IPOs, billion-dollar deals, and a landmark survey that puts hard numbers on what AI is actually returning. The era of proof-of-concept is over. This week, AI in life sciences officially entered the era of proof-of-results.

Paulina Gozdzialska
Paulina Gozdzialska
The Agentic R&D Shift: How Life Sciences Is Rebuilding the Lab Around AI

Feb 23, 2026

•

8 min read

The Agentic R&D Shift: How Life Sciences Is Rebuilding the Lab Around AI

In 2026, we are officially entering the "Agentic Era." As we look at the headlines from the past week, the theme is clear: AI has moved from an assistant to an active researcher that executes workflows. From Eli Lilly’s new "AI Factory" to Merck’s massive data play with Mayo Clinic, the industry is no longer just "using AI"—it is rebuilding the laboratory around it. In every issue, we pull together the most relevant developments at the intersection of life sciences and AI — from product launches and funding moves to events coming up.

Paulina Gozdzialska
Paulina Gozdzialska
© 2026 Life Sciences Digital.
beehiivPowered by beehiiv