The Baillie Gifford US Growth Trust trimmed its exposure to Nvidia this summer and portfolio manager Gary Robinson is now looking for the next superstar. He believes SpaceX and Stripe – the trust’s two largest unquoted holdings – could both enjoy Nvidia-like success, while other areas poised for exponential growth include autonomous cars, robotics, quantum computing and inferencing chips.
The trust has owned Nvidia for many years and it remains a top-10 holding but now that it has a $3trn market capitalisation, supranormal growth may be harder to achieve.
Two types of chips are needed for artificial intelligence (AI): graphics processing units (GPUs), which are required to train AI models; and inferencing chips, which allow AI models to consider different answers and draw conclusions. Nvidia has a 90% market share in GPUs but Robinson thinks inferencing will become a much larger market.
“Over the very long term, the inferencing compute market is probably going to be many multiples the size of the training compute market, because training is a function of the number of researchers, but inferencing is a function of the number of users of AI, which is going to be absolutely massive,” he explained.
Nvidia will participate in the inferencing chip market but many other players are building their own inferencing chips, including Google, Amazon’s AWS Inferentia and private semiconductor company Groq.
High-speed internet provision is another “gargantuan” market, in which SpaceX has a dominant position. SpaceX’s Starlink satellites can provide low latency, high bandwidth internet anywhere in the world and it has developed an aviation service to provide high-speed in-flight internet. “SpaceX is just so far head, there’s no one even close in terms of launch capability,” Robinson observed.
He is not deterred by Elon Musk’s support of Donald Trump or by the prospect of him taking on a political role. “Musk has surprised us over time by his unique ability to juggle multiple balls,” he said. “I've been very happy with the execution at Tesla and at SpaceX.”
SpaceX is currently the US Growth Trust’s largest holding (worth 7.6% of the portfolio as of 30 September), followed by The Trade Desk, Amazon, Stripe and Nvidia. Rounding out the top 10 are Meta Platforms, Shopify, Netflix, Tesla and DoorDash.
Quantum computing is another area with huge potential. Californian start-up PsiQuantum has received grants from the US and Australian governments and this high-level endorsement should help it to raise capital from private investors, Robinson noted.
He also believes investors outside the US have been slow to appreciate the breakthroughs in driverless cars. “The world is asleep at the wheel on autonomous driving,” he stated.
Robinson has just returned from five weeks in Silicon Valley. Hailing an autonomous Waymo taxi in San Francisco and stepping into an empty car was strange at first but “within two weeks, I became completely desensitized to it and it was just normal”.
A colleague also gave him a demonstration of Tesla’s self-driving mode which is only available in the US.
Baillie Gifford has invested in Aurora Innovation, which is developing software for self-driving trucks that it hopes will be on the road by the second quarter of next year. There have been concerns about AI taking people’s jobs but trucking companies are struggling to hire new drivers so automation would help solve the shortage of workers, Robinson said.
Robotics is another area of rapid innovation. Physical Intelligence, a start-up that received funding from OpenAI and Jeff Bezos, is building general-purpose software to communicate with robots in natural language. It’s early days but this is “the next frontier”, he noted.
Robinson expects the discussion about investing in AI to broaden out from AI-focussed companies to encompass any business using AI to improve its products or the efficiency of its processes.
Companies that are best placed to make the most of AI tend to have large customer bases, proprietary data sets and product-oriented leaders who have the moral authority to make the necessary changes to integrate AI tools. Cloud-native companies have an innate advantage because they are not encumbered with ‘tech debt’ (outdated legacy technology), he added.
Shopify’s Tobias Lütke is an example of a technical, product-savvy leader and the company has masses of data from the two million merchants using its services, Robinson said.
Meta is another obvious beneficiary of AI, he continued. It has an enormous, centralised data set and a product-oriented founder. AI is already having an impact on the levels of engagement on its platforms and conversion rates for advertisers.