From Nvidia to TSMC and Shopify, Baillie Gifford's Mike Taylor outlines where he sees AI valuation
The charge into artificial intelligence stocks has been so frenzied that even seasoned institutional investors are asking if they’re staring at the next big structural shift or just another expensive mirage. But according to Baillie Gifford’s Michael Taylor, they're asking the wrong question if they only focus on whether AI is a “bubble,” noting that it’s a contradiction in terms.
The investment manager and partner at the firm believes the potential of AI can be “absolutely transformative”, noting the technology has potential to reshape almost every sector, from finance to medicine and scientific research, especially where vast datasets and additional intelligence can be applied efficiently. That structural demand, in his view, is real and substantial, and not just a passing fad.
“The demand here is real. I think the industry has been constrained so far by access to chips, and I think the use cases for the technology are still emerging. We're very, very optimistic about the long-term picture,” said Taylor.
However, Taylor stresses that not every company caught up in the AI rally deserves its current valuation, and that the real task for investors is discrimination, not blanket exposure. After several years where equity markets were dominated by a small group of mega-cap tech names, he now sees a broader landscape where active stock pickers can be selective within AI and still find high-quality, underappreciated businesses.
For Baillie Gifford, the goal is to identify businesses that are extremely difficult to replicate and control genuine bottlenecks in the AI supply chain, where economic value is most likely to concentrate. Nvidia and TSMC, for example, sit at a unique chokepoint in advanced chip design and manufacturing.
“There is no other company that can design cutting-edge GPUs like they can. There are no other company that can make them like TSMC, which is the foundry,” noted Taylor.
But while supply can and will be built, that’s not where pricing power is likely to stick, Taylor underscored. Instead, Baillie Gifford is hunting in less obvious corners of the value chain. Notably, Taylor argues that the most compelling AI opportunities often sit in overlooked corners of the supply chain rather than in the headline names. To that end, Taylor sees two main areas where AI-related investments look most compelling.
First is in the semiconductor supply chain and the companies that either design chips or provide the most advanced tools and technology required to manufacture them. These businesses, including niche equipment makers, sit at the heart of AI’s infrastructure and benefit from deep, defensible competitive advantages.
While Nvidia and TSMC are well-known bottlenecks, he’s just as interested in specialist firms like Japan’s Disco, which dominates the niche market for grinding, polishing and cutting tools used on silicon and silicon carbide. Those ultra-precise processes are essential to manufacturing advanced chips, according to Taylor, and Disco’s control over this step gives it real pricing power.
“While the cutting of silicon wafers is really niche, having totally smooth silicon is going to be essential to get the sort of advanced semiconductor silicon chips that you require to enable the most important technology,” and Disco “dominates that market,” said Taylor.
He applies the same logic further down the chain, where AI’s infrastructure demands create unexpected winners. Comfort Systems, which installs heating, ventilation and air conditioning, benefits directly from the explosive growth of power-hungry, heat-intensive data centres.
After all, these facilities generate lots of heat and require advanced air conditioning to keep them running, he said, noting that specialist installation has become “a bit of a bottleneck” because of skilled labour shortages in the US. That’s why both Disco and Comfort Systems fit the profile of the AI enablers Baillie Gifford values.
Taylor also acknowledged the software and platform companies that either control powerful distribution channels or are already using AI and machine learning effectively inside their own operations. This is where he sees the real money being made: by the cloud giants and chip suppliers that power the entire ecosystem.
“Some of the big beneficiaries at the moment will be those that rent the chips to the startups that are experimenting,” he said, pointing to Microsoft Azure, Amazon Web Services and Google Cloud as the core platforms most new AI products are likely to run on.
Because of that, he argues investors can capture broad upside from AI innovation by owning these hyperscale cloud providers and the companies supplying the underlying silicon, rather than trying to guess the next breakout application.
Domestically, Taylor pointed to Shopify as a prime example, noting the firm is a globally scaled e‑commerce platform that has deliberately shifted investment away from logistics and is positioning itself to sit inside the future AI-enabled retail value chain.
Because AI relies on vast physical networks of power, grids and assets, Taylor views Brookfield’s ability to manage cycles and run infrastructure assets effectively as a key advantage in an environment where demand for power and related services may become more volatile as AI expands.
Additionally, Stella-Jones is a more indirect but compelling beneficiary as the company supplies wood products - notably utility poles - during a time when he believes North America’s electrical grid is outdated and in need of substantial investment. As data centres proliferate and the energy system shifts further toward renewables like solar and wind, he expects a strong multi-year demand cycle for grid reinforcement.
“What we're looking for are companies that have one of two things,” he said. “They have advantage distribution or can use it in their own internal operations… Part of the picture here is that we’re trying to pick stocks that are unique and hard to replicate, but at the moment they're enabling the transition. Where it's a little uncertain is how this technology is going to be used and who the winners from that will be.”
Taylor firmly believes long-term investors can build portfolios which pair carefully chosen AI beneficiaries with unrelated growth stories, rather than relying solely on the big, obvious technology winners. While he acknowledged the litany of macro concerns, he argues that a portfolio anchored in genuine structural growth themes, including AI and other secular trends, can offer resilience against those shocks while still delivering attractive long-term returns.
"We think it's possible today to put together a portfolio that can really deliver strong earnings growth over the coming five years and therefore deliver for your underlying clients,” noted Taylor.
