It’s been quite a week for followers of the news. We had elections in Brazil and Israel, a Federal Reserve interest rate decision, an Autumn Budget in the UK, Anthropic’s S-1 filing ahead of its initial public offering, and key earnings results from Microsoft, Amazon, Alphabet, Visa and Eli Lilly, to name but a few. We also have the not-so-insignificant mid-term elections in America this week.
There is no shortage of material to discuss. But while markets naturally focus on the immediate implications of elections, budgets and company results, some of the most important changes take much longer to reveal their true significance. So, rather than add to the commentary on all the latest news, I thought I’d look to medical history for some insight into how major innovations affect the real world, and what that might tell us as investors.
I spent part of last week looking into the socio-economic impact of the discovery of penicillin, as part of wider research into potential investment opportunities in the medical market.
While studies have quantified its effects on mortality and health inequalities in particular countries and periods, the total worldwide impact remains difficult to measure, such is the scale of its impact. Its widespread introduction has undoubtedly saved countless lives, improved life expectancy and transformed healthcare outcomes around the world. For investors, it is a useful reminder that the full economic value of a transformative innovation is not always obvious, or easily measurable, when it first emerges.
Penicillin is, of course, only one example of an innovation whose eventual impact was extraordinarily difficult to foresee. Gene sequencing offers another example. The first major innovation during my career was the Human Genome Project, which began in 1990. It took 13 years and cost roughly $2.7 billion to sequence the first high-quality human reference genome, giving scientists a detailed map of human DNA. Today, thanks to massive increases in computing power, the process takes only a few hours and costs a few hundred dollars. For investors, this is an extraordinary example of how rapidly falling costs can turn previously unthinkable possibilities into real-world applications.
Increasingly, advances in medicine and science are being accelerated by developments in computing and AI. DeepMind, where the Blue Chip Fund has exposure through Google’s holding company Alphabet, was founded in 2010 with ambitious goals to develop useful AI and ultimately artificial general intelligence. It began working on AI by teaching it how to play basic video games from the 1980s. By 2016, however, DeepMind, now under Google’s ownership, had turned its attention to protein folding, a biological challenge that scientists had been trying to solve for over 50 years.
In 2020, DeepMind delivered AlphaFold 2, significantly advancing our understanding of proteins and how they fold into complex three-dimensional structures. Without getting too technical, proteins are essential molecules that perform many of the jobs that keep our bodies working. Their three-dimensional shape helps determine what they do, so being able to predict that shape can help scientists understand how proteins function and interact with other molecules.
Today there are more than three million users of AlphaFold across 190 countries and it has successfully predicted the structures of nearly all catalogued proteins, which Google estimates will save hundreds of millions of years of research time. For investors questioning where the returns from today’s enormous AI spending might ultimately emerge, developments such as AlphaFold offer an interesting perspective.
This brings us back to today’s AI investment debate. With spending on AI rarely out of the headlines, parallels are being drawn with the dot-com bust at the turn of the millennium. At the same time, questions remain over how much of this investment is yet producing clearly measurable returns outside the technology sector.
So, are we witnessing excess investment, or laying the foundations for innovations whose economic impact will only become apparent over time? What if medical breakthroughs are on the horizon with an impact comparable to the widespread use of penicillin? This time, however, they could be enabled and accelerated by sustained investment in research, computing and AI.
Gene sequencing has enabled scientists to read the human genome and AlphaFold has helped us predict and investigate the structures of the proteins encoded within it. Each innovation builds on what came before it and together they have the potential to rank among the most consequential medical technologies in history. AI is increasingly helping to connect and accelerate these fields.
Perhaps it is in this context that we should view the AI capital-expenditure ‘arms race’ now under way. For investors, the challenge is distinguishing between hype, excess and the investment required to enable genuinely transformative change. Much like penicillin, the eventual impact could prove extremely difficult to quantify.
