Gavin Baker, an investor and analyst, shared a post on his X account (formerly Twitter) offering an optimistic view of artificial intelligence (AI). In the post, he focuses on the key trends that he believes are shaping the entire sector. His ideas are based on current events in the technology world and emphasize why we should "trust the process"—in other words, have confidence in ongoing developments without unnecessary concern.
Baker begins by discussing foundation models in AI—large systems that serve as the basis for various applications. According to him, if these models do not become commoditized (that is, widely available and affordable to everyone), the balance of power will tilt even further toward the United States. Why? Because leading U.S. labs have internal capabilities that likely surpass what is publicly available. This means American companies have an advantage in quality and innovation.
Baker further predicts that the performance gap between proprietary models (those owned by specific companies and not freely available) and open-source models will widen even more. This trend is expected to emerge with the deployment of Nvidia's new Blackwell chips. These chips are designed specifically for AI workloads and are expected to deliver significant improvements in speed and efficiency. Baker also notes that U.S. chip export bans will increasingly constrain Chinese open-source AI efforts starting next year. This gives the U.S. another advantage, as China will have more difficulty accessing the necessary technology.
— Gavin Baker (@GavinSBaker) November 19, 2025
An Oligopoly in Closed Models
The next part of Baker's post addresses the market structure. He describes competition in closed foundation models as a "four-horse oligopoly"—a situation in which the market is controlled by only a small number of players. Among them, Google ($GOOG) is the only major pure-play publicly traded company offering direct exposure to this segment. This means that investors who want to bet on AI through stocks have a solid option in Google.
Baker goes even further, arguing that Google is the most attractive long-term investment among the so-called "Mag 7"—a group of seven major technology giants consisting of Apple, Amazon, Meta, Microsoft, Nvidia, Tesla, and Google itself. The reason? Enormous infrastructure investments. Google is spending billions on data centers, servers, and other equipment needed for AI. This gives it an advantage in efficiency.
Cost per Token: A Key Metric Where Google Leads
One of the most important metrics Baker highlights is cost per token. In the AI world, a token is a basic unit of data—something like a word or part of a sentence that a model processes. According to Baker, Google leads in this area precisely because of its massive infrastructure spending. A lower cost per token means Google can offer AI services more cheaply and efficiently than its competitors. This is particularly important at a time when AI is becoming part of everyday applications such as search engines and assistants.
Baker also criticizes the prevailing view that Google is an AI "loser." According to him, this notion is becoming increasingly unfounded. As the market develops, it is becoming clear that Google holds a strong position, thanks to its long-standing presence in the technology world. Its infrastructure investments are not merely short-term but are building a foundation for long-term success. And with its new Gemini 3 model, it has once again returned to the proverbial top.
There Is No AI Bubble; We Are Only at the Beginning
Baker ends his post on a positive note: according to him, there is no impending AI bubble. Instead, we are only in the early stages of a major transformation. The sector is still evolving, and trends such as the growing performance gap, U.S. technological advantages, and the dominance of companies like Google suggest that the opportunities are enormous. His advice to "trust the process" means that we should have confidence in these developments and not panic over short-term fluctuations.
