By Brett Winton, chief futurist at Ark Invest
Nvidia’s meteoric rise has herded many investors to concentrate their AI play into a limited universe, and the sell off of on Monday 24 June – which saw it drop 16% – reminds us there is more to AI than a few mega tech stocks.
Like many students, tech pioneer OpenAI took stock and fine-tuned its approach after it failed the US Bar last year. When it came back for a re-sit a few months later, its AI platform GPT not only passed, but beat 90% of lawyers who took the exam.
Legal representation from a robot is probably some time off, but OpenAI’s success still goes to show you may as well leave the rulebook at the door when it comes to the pace and scale of AI’s evolution.
It is making forecasting difficult for experts. But despite regular revisions, their conclusion is always the same – AI is revolutionising the global economy and it’s doing so very quickly.
As it becomes more sophisticated, the technology is automating a growing number of tasks for knowledge-based workers worldwide. It is increasing their productivity by multiples. By the decade’s end, we expect AI software to be addressing a market worth some $13trn.
Early adopters are already getting exposure by investing in the AI names dominating the US market. Nvidia recently became the most valuable company in the world. Yet days later (June 24th) it relinquished that spot as a sell-off resulted in a 16% drop in its share price, wiping as much $550bn off its market cap overnight.
Though some of the ground was clawed back the following day, this serves as a reminder that to capture the full scale of AI’s evolution, investors should be casting their net way beyond these highly concentrated hardware stocks and across the entire supply chain.
The rapid AI revolution
AI’s rapid growth kicked off two years ago with the public launch of Chat GPT. OpenAI’s flagship chatbot marked a watershed moment for general users, allowing them to explore AI’s true capabilities using natural language for the first time.
Fast forward to today, and ChatGPT users are growing beyond 200 million while more than 80% of companies are adopting AI.
The knock-on impact of this explosive adoption is that the cost of training an AI model is falling by a rate of around 50% every six months. That’s around four times faster than the rate set out by Moore’s Law – the benchmark relationship between advancing technology and falling prices.
It’s why things are advancing at a rate confounding even the experts.
When OpenAI first announced GPT-3, the consensus was it would be 80 years before the release of an AI system matching or surpassing human intelligence.
The timeframe shrank to 18 years when ChatGPT was launched. Today, experts believe the robots will start becoming smarter than us by 2030 or sooner.
Profiting from productivity
The area where AI’s rise is creating the most value is in the productivity of knowledge workers – those who jobs centre on critical thinking and solving problems.
We are already seeing this today. The productivity of software developers coding tasks on the Github Copilot AI platform is more than twice those not using it. But the uplift is becoming more pronounced and pervasive in line with AI’s sophistication. We expect a productivity uplift across all knowledge workers of up to five times by 2030.
That is a lot of additional value in the economy. And the market opportunity for investors lies in the software enabling this productivity to emerge.
These companies typically capture about 10% of the value they create for the end enterprises purchasing their solutions. On this basis, we forecast an addressable market in AI solutions increasing productivity of $13trn by 2030. This increases to $26trn if they can slice 20% off the back end.
Regardless, it represents a huge pool of capital for AI companies to profit from. The issue is investors right now are concentrating heavily on companies such as Nvidia and Taiwan Semiconductor on the chip design and manufacturing end of the spectrum.
Due to their absence from the over-concentrated, top-end of indices like the S&P500, companies further along the AI software supply chain are not enjoying the same attention. Yet it’ is these names we expect to generate the greatest returns of all as AI reaches terminal growth over the coming years.
Take Teradyne, for example, which commands 60% of the growing market for chip testing. Likewise, Palantir is consolidating the data integration and analytics market for large organisations and governments integrating AI.
Meanwhile, Kratos Solutions is at the forefront of the disruptive mobility technology market reshaping defence and security worldwide.
These are but a few examples and many more exist. The point is, AI is going to increase productivity and create disruption at an unimaginable scale over the coming years.
This is a once-in-a-generation opportunity that investors can benefit from if they diversify their exposure, rather than pile into the small handful of stocks dominating headlines today.
