AI Investing in 2026: The Next Phase of Artificial Intelligence
In a Nutshell
AI is shifting from a tech-sector story to a real-economy force, with enterprise adoption, robotics, and healthcare applications now driving investment returns. Physical bottlenecks—chips, refining capacity for copper/lithium/nickel, and electricity demand—are creating new supply constraints and opportunities in materials, power, and infrastructure. The decisive winners will be the companies that convert AI into measurable economic gains rather than those merely spending on the technology.
These notes were generated by AI and may contain inaccuracies.
When you hear AI, what do you picture? Chips, chatbots, Big Tech? That was part of the first chapter. But the next chapter of AI could look very different in our daily lives, from autonomous vehicles to how doctors screen for disease, maybe even data centers in space. Welcome to the AI economy. The AI is moving beyond the technology sector and deeper into the real economy. We think that that changes where investors should be looking.
For the last few years, much of the AI story has been about building the technology. Now we're starting to see another question matter. Who can actually put AI to work? Enterprise AI is gaining traction and not only where you might expect. Legal and medical administration, industries that have historically been slower to adopt new technology, are among the larger enterprise use cases outside of coding. That matters, because the next phase of AI may not reward every company equally. Simply spending on AI is one thing, turning it into better business outcomes is another.
But there's another side to this story. AI may feel digital, building it is extremely physical. Starting with chips, AI demand can accelerate quickly. Semiconductor supply cannot. Adding new memory chip capacity can take three to four years or more. And the broader buildout requires more than semiconductors. Data centers, power systems, energy storage, and the industrial infrastructure around them draw on a wide range of critical materials.
Across six materials tracked in the update, copper, lithium, nickel, cobalt, graphite, and manganese, the leading refining country accounted for an average of approximately 72% of global refined production in 2025. So the challenge is not simply mining more material, it's also building and diversifying the refining capacity needed to turn those materials into usable outputs.
And then, there's electricity. A historic amount of electricity is needed as demand for AI has resulted in the largest expected step-up in US electricity demand growth in a century. AI may live in the cloud, but the cloud still plugs into the grid, and it's creating demand for the physical infrastructure underneath them. And when existing infrastructure hits constraints, something interesting can happen. We start looking for new solutions, sometimes in unexpected places, like space.
Launch costs are roughly 95% lower than they were 65 years ago, helping make experimentation economically possible. One highly experimental idea, orbital computing. In theory, orbital computing could benefit from the more consistent solar generation than here on Earth, and a data center in orbit may not need the same land or grid connection as ones on Earth. Are we about to put all our data centers in space? No. But, the fact that we're even exploring ideas like this tells you how far the AI economy could reach.
And AI isn't only changing infrastructure, it's increasingly entering everyday life. One of the clearest examples is robotics. Robo-taxis are already operating commercially in more than 30 cities. More than 50 companies are developing humanoid robots, and the estimated cost of building one has fallen more than 30-fold in a decade. The shift from AI that can generate an answer to AI that can sense, decide, and act could open up a much broader set of applications.
Healthcare is another example. By 2030, one in six people globally is expected to be 60 or older. That means potentially more patients and more healthcare demand without necessarily having more medical resources. In one randomized mammography screening study, AI support increased cancer detection by as much as 50% for certain age groups. It's an example of where AI could potentially help scarce clinical resources go farther.
Put it all together and AI no longer looks like one isolated technology trend. It looks like a connected economy, an economy built on chips, materials, power, and infrastructure and expressed through increasingly real-world applications like robotics, healthcare, and potentially new frontiers such as space. But, the evidence, economics, and timing are not the same across every part of that economy. For investors, the next phase of AI may be less about asking, "Who is building AI?" and more about asking, "Where is AI beginning to change the economics?" Because that's where the next chapter gets interesting. Explore iShares 2026 Thematic Midyear Update, "Welcome to the AI Economy" at ishares.com.
Keep iShares by BlackRock in your library
Save the videos and channels worth coming back to, and find them again in one place.
