In June 2026, Anthropic announced the launch of Claude Science, its dedicated platform for scientific research. The reaction from biotech AI investors and founders was swift and largely uniform, with the consensus being that software is now fully commoditised and defensibility must shift to proprietary datasets and wet labs. While this feels like the logical conclusion with capital already flowing in this direction, it is not fully accurate.
Anthropic as infrastructure
Anthropic is a platform builder, not a direct competitor of startups and it is not pursuing a traditional biotech company path. It is building the foundational operating system on which the next generation of biotech will run, rather than operating as a drug-discovery startup. That is a categorically different move, and it should not be evaluated on the same terms as a company bringing therapeutics to market.
Claude Science will accelerate consolidation among smaller AI biotech tools. Standalone point-solutions that rely on general AI capabilities without meaningful proprietary data are now structurally exposed. They must either evolve into specialised nodes, or risk being rendered obsolete. The ones that survive will be those with a clear, defensible role in a larger system.
We know that pharma startups are not retreating from Anthropic. Claude Science has already been validated by AstraZeneca, and when a pharmaceutical giant of that scale endorses a platform, the rest of the industry follows. Startups will not try to compete with it, instead they will build on top of it. The factor key to their survival will be if they remain defensible once they do.
The wet lab arms race
The most common response in investor circles has been to back a physical moat by funnelling capital into proprietary wet labs. These specialised research facilities designed for analysing drugs and biological material using liquids (hence, “wet labs”) create infrastructure that a pure-software layer cannot replicate. This logic is understandable, but is also likely to lure a generation of startups into an expensive trap.
There is a strong possibility that the physical lab layer will be abstracted into cloud APIs within the next few years, shifting the competitive advantage back to algorithmic architecture. Emerald Cloud Lab already offers remote-controlled experiments run through software, while automated “self-driving” foundries are making it possible to run thousands of biological experiments via API without needing to own a single centrifuge. Strateos, an early mover in the space, has since narrowed to a private, on-premises model, further proof that the category is still consolidating. Relatively soon, the physical lab could just be a commodity, rather than a competitive advantage.
If this happens, a small team operating a superior AI reasoning loop will be able to outperform startups that have spent tens of millions on lab infrastructure, simply by orchestrating outsourced robotic experiments at scale. The competitive advantage is shifting away from who owns the data repository and towards who has the better algorithmic search logic. The startups building brick-and-mortar moats today may find those walls dissolve faster than their depreciation schedules suggest.
The investment case has shifted
Over the last few years, acquisition logic in healthtech has inverted, and Claude Science is accelerating this trend. Historically, private equity, health systems, payers, and tech giants made headlines. They acquired healthtech and digital health platforms to optimise, expand to adjacent markets, protect existing infrastructure, and delay disruption from new entrants. In today’s market, a new cohort of acquirers will become increasingly relevant.
Venture-backed healthtech companies, fuelled by large funding rounds, have become the primary acquirers. OpenEvidence and Talkiatry, which collectively raised a little under $500 million in early 2026, are scaling aggressively and acquiring AI automation capabilities and medical knowledge assets. AI-native drug makers such as Earendil Labs, backed by $787 million, are expanding their platforms in parallel. Recently, OpenLoop, a privately held digital health infrastructure provider that partners with telehealth brands, acquired Hey Revia, and Innovaccer, a healthtech company operating an AI-powered data and analytics platform for healthcare organisations, picked up CaduceusHealth.
Beyond healthtech companies, a new generation of AI market leaders, supported so far by a growing share of venture capital’s total funding, is starting to look at healthcare to prove the utility of their models and chart paths towards sustainability. The $100 million to $1 billion acquisitions happening now will look small in the coming years as companies like Anthropic, OpenAI, Mistral, DeepSeek, and NVIDIA, through BioNeMo, begin acquiring at scale. Take NVIDIA as an example: it is expanding its healthcare strategy beyond medical imaging and into surgery with open models, datasets, and developer tools designed to help medtech companies accelerate innovation.
The acquisition landscape will shift in ways similar to how it shifted when Google, Optum, and IBM entered it in the early 2010s, replacing the former traditional players like McKesson, Siemens, and GE Healthcare, which had made the bulk of the largest acquisitions at the turn of the century.
The question for early-stage investors is: who will be worth acquiring when the shift comes? For startups building AI tools for scientific discovery, these moves change the investment case substantially. Technical capability alone is no longer sufficient for defensibility. The startups best positioned are those with proprietary data assets, a clear role in a larger orchestration architecture, and the commercial traction to prove they are delivering value rather than just venture-backed.
The last mile will not change
There is a broader question that the Claude Science debate has not yet addressed, and it is the one I find most interesting from an investment perspective. When AI solves the science of drug discovery, what will remain the same? In my opinion, the clinical trial frameworks will remain, the infrastructure of insurance payers, hospital networks, and government regulatory bodies may slowly shift, if at all. Even with all the incremental advancements announced recently, new opportunities in Expanded Access, RWE and Adaptive Clinical Trial frameworks, regulatory pathways for bringing a drug from discovery to market will remain complex and heavily human-mediated.
If anything, progress at the discovery layer creates more pressure on this last mile. A world with significantly more drug candidates in the pipeline, discovered at a fraction of the current cost, will urgently need better tools for navigating clinical development, market access, and patient adoption. The bottleneck will move downstream, where I believe the next decade of durable healthtech value will be built.
The future of healthtech investment
The Claude Science announcement is best understood as a signal to the healthtech market, rather than a threat. It tells us that the orchestration layer of biotech is being claimed by a well-funded foundation model company with pharmaceutical validation. This means the infrastructure question is largely settled. The more productive investment question is what gets built on top of it, and what remains stubbornly human, or at least, human-mediated.
Looking from the vantage point of these few dominant platforms requiring massive capital to reach scale and that are now already consolidating, the startups worth backing in this environment are those operating in the last mile of healthcare, where regulation and human judgment remain unavoidable. And, those building specialised capabilities that make them indispensable nodes within a larger platform rather than standalone applications. The wet lab arms race is a distraction and the last mile of healthcare is more valuable today than it has ever been, precisely because the pressure upstream has never been higher.
Dr. Marta G. Zanchi
Dr Marta G. Zanchi is the Founder and Managing Partner of global healthtech VC Nina Capital. She was previously a member of the faculty at the Stanford School of Medicine and founding director of the digital health programs at Stanford Biodesign, for which she continues to be an ambassador. Marta was also the founding CEO of RenovoRX, a Nasdaq-listed medtech company, and leverages her founding and operational expertise to support Nina Capital’s companies.
