AI stocks continue their upward momentum at the beginning of the new year, following a remarkable surge in 2025. Even though their performance last year far outpaced the broader market average, their outlook remains widely optimistic. The VanEck Semiconductor ETF, heavily invested in AI chip manufacturers, has more than tripled over the past five years. Meanwhile, as cryptocurrency mining companies pivot towards AI infrastructure, the CoinShares Bitcoin Mining ETF has risen approximately 30% year-to-date. These sustained gains not only reflect the continuation of investor enthusiasm but may also signal further expansion in the field.
One of the key drivers behind the AI boom is the continued increase in financial commitments from major tech companies. For instance, Mark Zuckerberg’s recent announcement of the Meta Compute initiative highlights Meta Platforms’ ambition to significantly expand its AI capabilities. The company plans to build tens of gigawatts of computing power to handle more AI workloads within this decade, with a long-term goal reaching hundreds of gigawatts or more. To achieve this, Meta Platforms expects to sustain substantial investments for many years to come. Microsoft has also unveiled its “Community-First AI Infrastructure” plan, involving heavy investment in AI data centers, even drawing an analogy between AI and the next chapter of American infrastructure following canals, railroads, power plants, and highways. Massive capital is flowing into the AI sector, and this trend is likely to persist for many years.
While much current discussion on AI focuses on software applications like ChatGPT and their role in enhancing enterprise efficiency, a wave of physical AI is emerging. Autonomous vehicles are already being tested on roads in some major cities, and Tesla’s humanoid robots are expected to begin rollout this year for internal users and major industrial clients. The development of physical AI is consequently increasing the demand for AI chips, energy, and other related infrastructure. This technology not only holds immense potential but has already generated substantial profits for some large enterprises—a feat not achieved by all companies during the internet bubble era. Physical AI has the potential to reshape nearly all industries. For example, Uber Technologies is investing in autonomous vehicle technology to avoid being overtaken by competitors heavily investing in AI. Its vast prospects make it difficult for long-term investors to ignore.
Although Nvidia has demonstrated that AI stocks can deliver returns transcending eras, and AI chip manufacturers remain in the spotlight, a growing number of investors are recognizing various bottlenecks within this boom. By shifting attention to often-overlooked bottleneck areas—such as gigawatt-scale power, storage, and raw materials—one can discover some rapidly growing smaller companies. These opportunities are not permanent, and some investors are already positioning themselves ahead of the curve. As trillion-dollar companies like Nvidia, Meta Platforms, and Microsoft continue to exhibit robust growth, it becomes easier to understand that smaller companies focused on solving AI bottlenecks may also offer substantial long-term returns.
