In 2008, Facebook astonished observers when it took just four and a half years to reach 100 million users. Last year, Deepseek achieved that in seven days. ChatGPT is on track for a billion users.
With astonishing progress in machine learning capabilities, adoption rates continue to accelerate. Over the next decade, business models will be turned on their heads in ways we have yet to imagine.
But index investors find themselves in a bind. Their portfolios are increasingly exposed to the most capital-intensive parts of the Artificial Intelligence (AI) arms race. With the large hyperscalers – Alphabet, Microsoft and Meta – competing to outspend each other, they face a dilemma: stick with the giants, hoping for further gains, or look elsewhere?
Index concentration brings risks and questions about whether there’s an AI bubble. The reality is that no one knows because, after three years of astonishing progress and staggering levels of spend, there is little clarity on the market’s future shape and where value will accrue, especially if AI models become commoditised and monetisation proves elusive.
Whilst index investors grapple, active investors can seek opportunities away from the crowded trades, focusing on businesses with less exposure to these competitive dynamics – especially if they take a long-term view
So, where can we explore what the market is missing in AI?
Bottlenecks and pinch points
History shows that the greatest value from new technologies often accrues to businesses that address critical bottlenecks in the system. In this regard, both TSMC and ASML occupy enviable supply chain positions, as the engine rooms of modern technology. Over 90% of the world’s most advanced chips are made in TSMC’s foundries, chips 20,000 times thinner than a human hair.
ASML’s high-end ultraviolet lithography equipment is also critical for etching those minuscule semiconductors. Without these two companies, AI would not exist.
Energy represents another strategic pinch point. For decades, computing energy efficiency doubled every year and a half or so. But recently, the rate has slowed.
AI data centre energy consumption has soared, with power spending up 15-fold in three years. Now, computing capacity is measured in watts, not bytes.
This has profound implications. In the US, investment in renewables and grid upgrades is lagging, while China is surging ahead, benefiting companies such as CATL, which is well-positioned through its battery and energy systems. CATL’s energy storage could become a much larger part of its revenue.
Business models built on AI foundations
Beyond the core infrastructure a new generation of companies is emerging by building business models on the foundations of AI.
Cloudflare optimises web performance and security with the potential to play a key role in AI governance. Its ‘Crawl Control’ software helps online publishers control access to their data and enables machine-to-machine micropayments at scale.
Samsara is harnessing AI to transform the operations of trucks and industrial equipment. Having them in the right place at the right time, thanks to predictive analytics, saves customers millions of dollars a year.
Duolingo uses generative AI to expand its education offerings and personalise learning – launching over 150 new courses this year. Without AI, it took over a decade to create its first one hundred courses. Duolingo’s machine-learning model ‘Birdbrain’ continuously assesses each learner’s knowledge and assigns exercises at a level of difficulty to maximise engagement.
Applovin uses AI-powered advertising targeting and analytics to help app developers acquire users and monetise apps, processing millions of ad-decision requests every second.
Intuitive Surgical deepens its competitive edge with AI-powered surgical robots and analytics, improving patient outcomes and returns on investment for hospitals.
Further East
China remains overlooked by many investors, yet it is registering hundreds of new generative AI tools each month. An improving business environment and the global mobility of Chinese technology talent mean the country will be a major force in shaping the future of AI.
Homegrown infrastructure and software companies such as Enflame, Moore Threads, Meta X, Biren, Kunlunxin, and Minimax have made major strides forward. Meanwhile, Chinese robotics players such as Unitree are widely considered to be a long way ahead of Western counterparts.
Chinese electric vehicle (EV) players such as BYD have transformed their brands and quality levels. It will be interesting to see how BYD develops autonomous driving capabilities and premium models from here.
Adventures and imagination
These examples show the breadth of opportunity created by AI. History suggests that the winners of the future are rarely the same as the winners of the past. Those enabling, supporting, and innovating around the core trends look best placed to benefit.
For those concerned about an AI bubble, the answer is not to retreat from global equities, but to look beyond the index. The greatest returns rarely materialise from following the crowd, but from imaginative, adventurous and patient exploration.
Tim Garratt is an investment specialist on the Baillie Gifford Long Term Global Growth fund. The views expressed above should not be taken as investment advice.