11/19/2026 | Press release | Distributed by Public on 11/19/2026 02:04
By Christina Takke
In early 2000, as the unravelling of the first draft of the human genome neared completion, many investors were certain they knew what it meant. Genomics companies traded at valuations that seemed to assume medicines would follow the genomic data almost automatically.
Two years later, the picture looked very different.
The company Celera Cooperation was a famous example of the genomics hype. Founded in May 1998 by J. Craig Venter, it was known for its pioneering work in rapid DNA sequencing and its famous race against the publicly funded Human Genome Project. While the rivalry ended in a truce, the company lost about 80% market capitalization between September 2000 and March 2001.
The genome was real. The business model - as priced - was not.
Today, a different technology is attracting a similar sentiment. Companies developing or using Artificial Intelligence (AI) are currently presenting with some of the largest private valuations on record. Where AI is used in drug discovery, some of these valuations are hard to reconcile with clinical evidence.
For those of us who invest in therapeutics, the question is not 'whether AI matters', but to define where the value will be. The genomics cycle taught us how to back a transformative technology, but we need to remember to do that without paying for the whole future value up front.
The genomics and AI cycles both began with genuine breakthroughs. The first working drafts of the human genome, published in 2001, were scientific landmarks. Similarly, the AI tools developed to predict the 3D structure of proteins won Demis Hassabis a Nobel Prize in 2024. Neither development was a hype in and of itself - these were remarkable feats that opened the door to extraordinary potential scientific and medical progress.
Both movements also summarized their significant steps in simplistic slogans. In 2000, the message was that 'sequence leads to target, and target leads to drug'. Today the perception is that an 'AI model leads to a molecule, and a molecule to a medicine'.
In both cases, the slogan skips the clinical development, where most candidates still fail.
As with the genomics hype, valuations in AI currently seem to be adhering to this simplistic perception rather than the evidence at large. Isomorphic Labs Inc. is a contemporary illustration of this in action. The company was established in 2021 under the wings of Alphabet Inc., as a spin-off from its AI research lab DeepMind. In May 2026, Isomorphic announced a $2.1 billion raise at a reported valuation of $10 to $17 billion, without even having a clear product candidate in their pipeline.
Mere months later, a BioCentury commentary reflected on rumors that Isomorphic Labs is already looking to raise a new round - this time at a valuation of $40 to $50 billion. The article notes that "Isomorphic Labs would require near perfect clinical execution to justify the valuation" and juxtaposes the preclinical company with commercial-stage companies such as argenx, whose lead product Vyvgart is on track for more than $6 billion in sales this year.
There are many similarities between the early days of the genomics hype and what we're currently seeing in the AI space, with extreme valuations reflecting the huge expectation many investors have for the success of this new technology. Yet there are also differences between the two.
Many genomics companies sold sequence data, databases, and tools to a limited set of pharmaceutical and academic customers. Issues arose when that customer base eventually shrank or wasn't willing to pay the prices calculated in their business plans. By contrast, AI products are already being sold well beyond biotech, which gives AI platforms an economic base that the genomics pioneers never had.
Looking specifically at AI drug-discovery companies, the cost structure differs markedly. Genomics companies burned cash on sequencers and wet labs; AI companies like Isomorphic Labs burns more on computing capacity.
The investor base has also changed. In 2000, many of the leading genomics companies were listed, so pricing was public and the hype correction quick and visible. Today, many of the largest AI financings are private rounds, anchored by technology companies, growth funds, and sovereign wealth funds.
The result is a less transparent market, where valuations are not continuously tested by public investors - potentially allowing exuberance to persist longer before a correction becomes visible.
With over 25 years of hindsight, it is safe to say that the unravelling of the human genome did indeed change medicine for good. The development underpins targeted therapies and precision oncology now available to patients, and the cost of sequencing a genome fell from about $95 million in 2001 to roughly $500 by 2022.
But the payoff arrived on a longer timeline than the valuations of 2000 assumed. We often overestimate what can be achieved in the short term, but vastly underestimate what can be achieved in the long term.
It's important to remember that:
Biotech and deep tech: each side has what the other lacks.
Biotech investors know the attrition curve: roughly nine in ten drug candidates that enter clinical trials fail, and the distance between a compelling AI prediction and a safe, effective drug is a very wide gap. These investors bring critical judgements on target validation, trial design, and regulatory strategy.
Deep tech investors bring an understanding of computational economics, the capabilities and limits of AI models, and the talent market. They may also be more comfortable with infrastructure spending than many life sciences funds.
Left apart, each type of investor makes characteristic mistakes. Biotech investors may underestimate how fast AI models improve, while tech investors may benchmark a preclinical company against general AI valuations, rather than against that actual worth of a clinical program.
Collaboration between biotech and deep tech investors can be practical for both parties, providing joint diligence on both the AI model and the biology; syndicates that pair a technology lead with a therapeutics lead; and boards with both clinical and computational expertise.
This is our approach at V-Bio Ventures: the science, the team, and the development plan are assessed together. We believe that combining skills and expertise will pay off, particularly in a new field like AI drug-discovery. Diversity matters.
We are neither sceptics nor evangelists. AI will certainly change how medicines are discovered. The genomics cycle suggests that a technology can become very valuable in the long term, even where its early prices fall flat. The task for those of us who back biotech founders is to stay close to the evidence, pay for milestones, and keep patients at the center of the thesis.