Explosive Adoption and Usage
Artificial intelligence is the single biggest technological force in motion right now, and it is growing at an extraordinarily rapid pace. The scale of adoption in 2026 is striking. In the year-to-date period alone, AI-related applications have generated over 2 billion downloads, accompanied by nearly 3 billion in in-app purchase (IAP) revenue. Users have collectively spent roughly 12 billion hours of time engaged with these tools. These figures demonstrate just how thoroughly AI has permeated every facet of daily life.
A comprehensive view of the AI landscape can be broken down into distinct segments: chatbots, agentic AI shopping assistants, advertising, and the consumer-facing ChatGPT product. Each of these areas illustrates both how the technology is expanding and how different businesses are attempting to monetize their underlying capabilities. The throughline across all of them is rapid growth coupled with experimentation in revenue models.
The Disconnect Between Broad Usage and a Narrow Stock Market
A central tension emerges when comparing how widely AI is used against how the stock market is rewarding AI companies. Despite the fact that people are spending enormous amounts of time using AI, and despite its reach into so many parts of life, the equity market remains highly concentrated and narrow — the gains are clustered in a small set of names.
The key question is when this narrowness will break and the market will broaden out. The answer lies in real-world demonstration of value. A recent visit to Taipei with Intel — a company skilled at bringing its customers on stage — offered a glimpse of this dynamic. Customers such as Ericsson and GE Healthcare presented how they are actually deploying AI. The market will begin to broaden when companies like these can return and articulate concrete benefits: new lines of business, higher revenues, or improved margins. Once those tangible results become visible and demonstrable, equity gains should extend beyond the core infrastructure players — the companies building chips, providing power, or handling the communications between the various chips — to a much wider universe of businesses.
Are Real-World Efficiencies Keeping Pace With Market Rewards?
Are our lives genuinely being improved by real efficiencies from AI to the same degree that we are rewarding these companies in the stock market? The honest answer is nuanced. AI tools are definitely being used and are definitely creating efficiencies in certain businesses, but the players have positioned themselves in very distinct ways, which shapes where and how those efficiencies materialize.
ChatGPT is fundamentally a business-to-consumer (B2C) play. Claude, by contrast — where data shows that 71% of users are web-only — operates much more as a business-to-business (B2B) play oriented toward enterprise customers. Because of this enterprise targeting, Claude appears poised to integrate itself somewhat more quickly into certain enterprises and to drive efficiency within their workflows, since that is precisely where it seems aimed.
ChatGPT, at this point, remains more of a consumer product. People are certainly using it, but it is uncertain how deeply it has embedded itself into everyday professional workflows compared to Claude. At least anecdotally, and according to the available data, Claude appears to be more woven into people's everyday work than ChatGPT is.
Consumer Spending and the Question of Price Wars
Consumer spending on AI is, frankly, eye-watering. Real-world anecdotes underscore the point: some individuals subscribe to the top-tier Claude platform, and others report spending as much as a thousand dollars a month on their AI applications.
This raises a pointed question: are we spending a lot now, and is that likely to come down in the future — particularly as companies like OpenAI talk about potentially and drastically cutting prices in order to compete with the likes of Anthropic? Notably, the industry is already discussing price wars before these companies have even gone public.
Such price cutting could indeed happen as competitors fight for market share. However, the critical consideration is how much it actually costs these companies to produce the service they deliver — especially given the energy and infrastructure costs involved. For that reason, aggressive price competition may function only as a near-term or short-term tactic to remain competitive; it is not a sustainable long-term business plan.
This cost dynamic also helps explain why ChatGPT is now moving into advertising, attempting to capture advertising revenue and secure a first-mover advantage in that arena. It would be very difficult to operate sustainably on bargain-basement prices alone. If these companies do want a competitive, low-cost entry point, the likely outcome is a tiered model — with or without ads — very similar to what the streaming services and cable companies have already done. Without that kind of structure, the model simply would not be sustainable given the cost of infrastructure and power.
The Semiconductor Backbone: Singapore, Manufacturing, and R&D
The AI revolution rests entirely on a physical foundation of semiconductors and the machines that build them. A recent trip to Singapore highlighted a crucial and underappreciated fact: roughly 20% of the machines that build semiconductors are manufactured there. The specific purpose of the visit was to examine a new half-billion-dollar investment by Applied Materials to expand its production capabilities in the region.
A vital point about AI is that everything we use today depends on machines that companies like Applied Materials, Lam Research, or ASML sold years ago. Consequently, if you want to understand the mid-to-distant future of AI, the right place to look is at the manufacturing and R&D investments being made now in semiconductor equipment. Across the semiconductor segment, there is a pressing need to accelerate how chips are designed and built across new generations, because AI has an endless appetite for ever greater speed and power in semiconductors — an appetite that is not going to go away.
Is It a Bubble? The Intertwined Market
A recurring question is whether all of this constitutes a bubble. The firm conclusion is that it is not a bubble. The applicability of these AI services is radically expanding, which underpins the durability of the trend rather than signaling a speculative excess.
Equally important is where chips are made and where they are packaged. There is phenomenal capability in Taiwan, and there is an emerging and growing capability within the United States. As reliance on AI extends into every part of life — exemplified by what Qualcomm is doing in automotive — the location of manufacturing and the reliability of the supply chain become genuinely crucial. When every part of your life is tied to AI, where it is built and how dependable that supply chain is matter enormously.
These threads are all intertwined and are being reflected directly in stock prices. The Intel–Apple story is a vivid example. Apple is reportedly looking to raise its prices because of memory costs, which in turn creates spillover effects for storage names — those companies gain more pricing power as a result. The Intel–Apple relationship is also a remarkable testament to how far Intel has come; such a partnership would have been unthinkable just two years ago. An endorsement from the U.S. government also helped advance that story. Taken together, these developments show an ecosystem in which usage, monetization, manufacturing, pricing, and supply-chain reliability are all deeply connected.
