An influential operative in the life sciences area – and this is relevant because we are expanding holdings in drug companies Eli Lilly and Moderna in the portfolios we manage – last week noted the mismatch between the rising number of drug discoveries over the past decade and the unchanged approvals rate from the US Food and Drug Administration (FDA) for those drugs. The chart below reveals the picture.
NB: The left hand scale is in thousands, but not the right.
Drug discovery is easy, the paper posited, because of the huge leap in processing power (yes, AI, but also not AI, since even without it the silicon chip makes possible in seconds what used to take months). But approval for use in the affected patient cohort is still a laborious process given that the move-fast-and-break-things mantra of almost all technology and disruption companies does not play well when it results in the shortening of human lives.
Science is advancing on cures for a number of cancers
Notwithstanding, something profound is happening in life sciences. In the past ten years there have been very significant new treatments for blood and solid cancers (from companies including Merck, Roche and Bristol Myers Squibb) to obesity-related heart, liver and kidney diseases (because of the weight-loss drugs of Eli Lilly and Novo Nordisk, now possibly in decline over a medium term timeframe).
In the past few weeks, there has been a breakthrough in the treatment of pancreatic cancer (previously considered undruggable) with the drug Daraxonrasib.
Source: Stat News; An unheard-of standing ovation for Revolution Medicines for its success in pancreatic cancer trials at the recent American Society for Clinical Oncologists address
Reading between the lines, smart scientists, probably with AI’s help, are bringing together computer-aided drug designs, protein folding algorithms, gene and genome sequencing all of which are implemented through known chains of command in the cellular cancer switching mechanisms.
As the saying goes, “every overnight success is 20 years in the making” and this seems to be no exception. 25 years ago this would have been impossible without the human genome being sequenced and DeepMind’s AlphaFold being released in 2020. Our own Rob Forage, a PhD with a multi-decade career in the field and a member of our Expert Investment Advisory Panel for the past five years, confirms the emergence of these trends. Investors and friends should expect to see Loftus Peak become more active in this area, subject to the universally applied risk metrics used in Loftus Peak’s portfolio construction.
OpenAI and Anthropic await their IPO moment
Meanwhile, we expect shortly the Initial Public Offering documentation of AI leaders Anthropic and OpenAI. Loftus Peak did not invest in SpaceX because it did not meet our investment valuation criteria. Anthropic and OpenAI have demonstrable revenue growth metrics and business models rooted in more prosaic endeavours (relative to occupying Mars using a SpaceX rocket) such as code-writing and business process optimisation.
Anthropic is pushing frontier model development with a strong focus on safety and reasoning. OpenAI sits at the centre of global AI adoption, enterprise integration, and consumer interface transformation. This is a usage boom, as well as a roll-out boom – a crucial difference from the 1999 internet crash, which in its execution was deeply flawed in so far as it operated with globally low broadband speeds. Who can forget believing the hype, then waiting 15 minutes for a still picture to load?
Source: Alamy
The market is no longer debating whether AI is real – it is instead debating how large the monetisation curve ultimately becomes and to what extent it is priced into company valuations.
Currently, we are seeing an increase in the number of clear, consistent datapoints, with burgeoning revenue realisation across the ecosystem. Enterprise demand for AI tools is accelerating, inference workloads are scaling rapidly and infrastructure providers are reporting sustained demand signals rather than one-off surges.
Market moving from theoretical enthusiasm to observable cash flow impact as AI bubble narrative is muted.
In our view, this is reflected in a combination of accelerating structural growth in AI, especially agentic AI, and a rapid repricing of what constitutes the “terminal” value, essentially what the business is worth in a slow-or-no-growth environment.
The big picture is that while markets have always produced extraordinary winners, small companies compounding 10, 20, even 50 times over long periods, we have never seen the world’s largest companies double and redouble in value within a 5 year period. So it isn’t just speed, but it is also scale: the re-rating is occurring at the very top of global capitalisation tables – the biggest companies – where incremental capital flows are measured in trillions rather than millions.
If and when Anthropic and OpenAI enter public markets, it will represent a structural expansion of listed technology exposure. More importantly, it will further blur the boundary between private and public market value creation, as some of the most important AI infrastructure in the world transitions into investable form.
As we look ahead, the key question for investors is not whether these trends continue, but how to maintain exposure to them through cycles of volatility, valuation compression/expansion and macro uncertainty.
Our approach remains focused on identifying businesses with durable structural growth, scalable economics, and exposure to long-term technological disruption.
Share this Post