The Investors Who Saw the AI Boom Early — and Bet Everything on It

Some investors don’t wait for consensus. They don’t require a chorus of analysts giving upgrades or Bloomberg terminals flashing green. They write the check despite seeing something early on, which is frequently messy, speculative, and widely disregarded. For a considerable amount of time, those who provided funding for the revolution in artificial intelligence were precisely that type of person. Quietly, often without fanfare, they placed bets that the rest of the financial world would spend years trying to understand.

It’s important to keep in mind how bizarre the early OpenAI narrative appeared. In 2015, a group of technologists gathered not to build a product or flip a company, but to create artificial general intelligence “for the benefit of humanity.” The names attached to that founding moment — Elon Musk, Reid Hoffman, Peter Thiel, Jessica Livingston — were already well-known in tech circles, but their involvement in something this abstract raised eyebrows. The revenue model was unclear. There was hardly any product. The majority of what existed was belief. And, as it happened, faith was sufficient.

Hoffman, who made his name building LinkedIn and then investing early at Greylock, had a reputation for thinking in long arcs. Thiel had already demonstrated an appetite for contrarian bets — Facebook, Palantir, early crypto. Naturally, Musk had accepted existential risk long before it became popular. The approval of a market cycle was not necessary for these individuals. In a way, they had already practiced ignoring one.

The companies that came after had similar instincts but distinct personalities. Instead of engaging in philosophical debates about intelligence, Andreessen Horowitz and Y Combinator were conducting large-scale pattern recognition. What they saw was that AI infrastructure was evolving into a foundational layer, much like the internet had been a generation before. They wagered not only on concepts but also on the individuals who developed them. This distinction is more important than it might seem.

Microsoft comes next. Given the company’s dominance in this market, it’s easy to forget that its shift to AI was a real strategic risk. Investing almost $35 billion in OpenAI’s infrastructure in a single quarter was a statement rather than a standard corporate investment. Observing from a company that had been lagging behind in consumer technology for years, Satya Nadella saw something in OpenAI that had the potential to completely transform Microsoft. Although there are still many unanswered questions regarding the return on that amount of capital, the initial results of integrating AI into Office, Azure, and developer tools have largely confirmed that intuition.

Nvidia’s stance merits a separate paragraph. Although the company is technically a chipmaker rather than a venture firm, its function in this narrative is more akin to that of the railroad owners during the Gold Rush. Nvidia was creating the tracks while everyone else argued over which AI startups would make it. Years before the term “generative AI” became widely used, Jensen Huang made an aggressive wager on GPU infrastructure for machine learning, which put the company at the center of every significant AI workload. Nvidia’s market value increased to $4.3 trillion by the beginning of 2026. It’s nearly impossible to comprehend that figure. It implies that investors were not only correct, but they were still underestimating the magnitude of what they anticipated.

Observing all of this, it seems as though the early investors profited from both proximity and intelligence. They were in the email chains, on the boards, and in the rooms. Reid Hoffman actually served on the board of OpenAI. Before ChatGPT altered the discourse, SoftBank’s Masayoshi Son had been investing billions in technology infrastructure for years. The information advantage was relational as well as monetary. In a market that prefers to think of itself as a level playing field, that may be difficult to admit, but it’s most likely the case.

But where this story has ended up is what makes it truly complex. In early January 2026, Ray Dalio, the hedge fund manager who turned Bridgewater Associates into one of the biggest macro funds globally, declared that the AI boom had entered a “early bubble phase.” Gold, which has historically been used as a hedge against speculative excess, had increased by more than 60% in 2025, while global stocks had recorded double-digit gains for the third year in a row, largely due to AI-linked names. Dalio doesn’t cry wolf. It is usually noteworthy when he uses such language.

In March 2026, BlackRock’s Larry Fink expanded on the warning, stating that the AI boom could lead to a widening of inequality on a scale that would surpass earlier technological cycles. His argument was that the benefits of AI are dangerously concentrated, not that it is overpriced. A small number of businesses that possess the data, infrastructure, and funding necessary to implement AI on a large scale are separating themselves from the rest. This was intuitively understood by the early investors. They wager on precisely that focus. Fink is now stating that the wager was successful and that the repercussions are still being considered.

Additionally, there are voices that completely reject the bubble narrative. Coatue Management’s Philippe Laffont, whose fund oversees about $70 billion in assets, argued at CNBC’s The “hyper-scaler advantage”—the structural capacity of corporations like Microsoft, Alphabet, and Amazon to invest hundreds of billions in AI—makes this era fundamentally different from the dotcom era, according to a late 2025 Alpha conference. A similar argument was made by Bill Ford of General Atlantic, who noted that the businesses driving change in AI are well-capitalized, profitable incumbents rather than venture capital-burning startups. That distinction is significant, though it’s still unclear if it will hold up under duress.

Looking back, it’s still amazing how few people who correctly answered this did so by looking at a spreadsheet. They were correct because they were keeping an eye on something more subdued than quarterly profits: a change in what was technically feasible and an indication of the boundaries of human ambition. The most astute individuals in the room weren’t always the investors who witnessed the early AI boom. They were the ones who were prepared to spend years sitting in that room before it gained notoriety, which may have been more beneficial.