What the 10 Largest VC Rounds in History Are Really Telling Us
Over the last three years we have seen something that had never happened before in the history of venture capital: a single company, OpenAI, raising a mega round of $122B in 2026, another of $40B in 2025, and a “non‑traditional” $10B round led by Microsoft in 2023; Anthropic showing up three times in the global top 10 with checks of $13B, $30B and $65B; xAI, Waymo, Ant Group and even an enigmatic “Project Prometheus” rounding out the list.
It is a ranking that looks like a scoreboard from another galaxy, but behind those figures there is something much more important than the shock of the number itself. Each round reveals how the relationship between capital, infrastructure, and technological power is being rewritten on a global scale, and it is precisely there where the following insights converge—insights that almost no one is seeing—surfacing from recent research by Cuantico VP.
SoftBank Is Building an AI Infrastructure Monopoly, Not Just Making Bets
There is a fundamental difference between betting on who is going to win a race and building the highway that everyone will have to drive on. Masayoshi Son, the legendary and controversial founder of SoftBank, chose the second option.
SoftBank led or co‑led the two largest rounds of OpenAI: the $122B one in March 2026 and the $40B one in March 2025. In total, it committed around $60B to that company alone. SoftBank controls ARM Holdings, the chip designer whose architectures are in practically every mobile device and increasingly in AI accelerators. And it also co‑founded Stargate, the AI infrastructure project in the United States that has committed up to $500B to data centers and advanced computing.
What emerges is an unprecedented position: SoftBank has exposure to the most widely used language model in the world (OpenAI), to the chips that process that intelligence (ARM), and to the infrastructure that hosts it (Stargate). If AI becomes the new electricity, SoftBank is building the distribution grid. The question is no longer whether OpenAI will win or lose. The question is: can anyone operate in AI without touching some SoftBank asset along the way?
Anthropic: The Fastest‑Growing Company in the History of Private Capital
Historical comparisons are hard to make, but this one is impossible to ignore. OpenAI, founded in 2015, took seven years to reach a $300B valuation. Anthropic, founded in 2021, closed its Series H in May 2026 at a $965B valuation, meaning it came close to a trillion dollars in less than five years from its founding.
Beyond the speed of valuation, it is important to look at the speed of revenue. At the time of the Series H announcement, Anthropic’s CFO revealed that the revenue run‑rate had surpassed $47B annualized. For context: in January 2025 that number was approximately $1B. In sixteen months, Anthropic multiplied its revenue by 47X. No tech company has done anything similar in that time frame.
With its Series H closed, Anthropic surpassed OpenAI in valuation for the first time, with an additional detail that many overlooked: the company confidentially filed its S‑1 with the SEC just days after the round closed. The largest round in its history may also be the last before its debut in public markets. The clock is already ticking.
Capital Is Turning into a Commodity
In January 2023, headlines said that Microsoft had invested $10B in OpenAI. What those headlines did not explain is that this investment was structurally different from any other on the list. The value did not arrive in cash, but in compute credit: access to Azure servers, supercomputing at scale, the ability to train models that would otherwise have cost billions of dollars in proprietary infrastructure.
This was not an accounting whim. It was one of the most massive and visible examples of an investment model that is now becoming a global trend: Equity in Exchange for Services. One example of this model is Media For Equity or M4E. In this structure, investors do not contribute cash but strategic assets such as visibility, infrastructure or audiences, in exchange for equity. What Microsoft did with Azure and OpenAI has a direct conceptual parallel to what specialized infrastructure or media investors like MediaForGrowth, Nascent VC or ILB Media Ventures do with their own platforms and assets.
In a world where building technology is increasingly accessible, the scarce assets are distribution, audience, and infrastructure. Whoever controls them has a moat of competitiveness that cannot simply be “financed.” Microsoft understood this years ago. The rest of the market is understanding it now.
The Most Strategic Investor in the Top 10 Is Not from Silicon Valley
When people talk about the big investors in AI, they mention Sequoia, a16z, SoftBank, Tiger Global. But there is an actor that appears at least three times in the top 10 VC rounds in history, without almost any media outlet explicitly pointing it out: GIC, Singapore’s sovereign wealth fund.
GIC co‑led Anthropic’s $30B Series G in February 2026 together with Coatue. It participated in Anthropic’s $13B Series F in September 2025. And it showed up again, once more as co‑lead, in the $65B Series H in May 2026. Added to its historical participation in Ant Group and other institutional rounds, GIC has quietly become one of the major builders of the global AI ecosystem.
What matters here is not just frequency but depth. GIC is not diversifying across dozens of small bets. It is concentrating capital in the same companies, round after round, increasing its position as the valuation grows. It is the strategy of an investor who is not seeking quick liquidity, but structural, long‑term influence in the companies that are likely to define the digital economy of the next decades.
Project Prometheus: A Mysterious and Important Entry
There is a name on the list that has no recognizable logo, no website, no public product and barely any media presence: Project Prometheus. With a $10B round in April 2026, it is the only company in the top 10 that is not publicly known and is not listed on any market.
It was founded by Jeff Bezos in November 2025, together with Vik Bajaj, a former Google X executive. Its mission, according to Bezos himself, is to create an “artificial general engineer,” a radically new tool for designing physical objects, which he has described as a “very modern version of CAD” applied to industries such as aerospace, semiconductors, automotive, and pharmaceuticals. It has nothing to do with robotics, he has been careful to clarify, although the confusion was understandable given the profile of its first 120 employees, recruited from OpenAI, DeepMind, Meta and xAI.
One of the most disruptive insights is not what the company does, but who is financing its growth: BlackRock and JPMorgan appear as anchor investors in the round. Two traditional financial institutions that have historically operated as asset managers or lenders, not as direct investors in growth‑stage VC rounds for early‑stage startups. The fact that large banks are joining AI startup rounds as direct co‑investors, without going through a VC fund as intermediary, is a structural break with the traditional venture capital model. And it suggests that Prometheus could become one of the most valuable private companies in the world over the next few years.
A Diverse Cap Table Is Also a Political Signal
Anthropic appears three times in the top 10. That is striking. But there is something even more revealing: in each of those three rounds, the lead investor was completely different.
Series F (September 2025, $13B): led by ICONIQ. Series G (February 2026, $30B): led by GIC and Coatue. Series H (May 2026, $65B): led by Altimeter, Dragoneer, Greenoaks and Sequoia. Three rounds. Three different coalitions. Zero repetition in the lead role.
It is a strategy that is completely different from OpenAI, where Microsoft has for years been the dominant strategic anchor and SoftBank has consolidated itself as the most relevant institutional player in the last two rounds. Anthropic, by contrast, seems to have made a conscious decision not to have a single “godfather.” That gives it strategic independence and operational flexibility, but it also means its governance is significantly more complex.
When critical decisions arrive—IPO structure, commercial alliances, acquisitions—there is no single actor who can approve or veto them. There is a table with many seats and voices. For now, that diversity looks like a strength. The question is whether it will remain so when the really hard decisions come.
The Only Data Point That Proves AI Is Not a Classic Bubble: Revenue
Every time headlines show multibillion‑dollar valuations in companies that have not posted profits, someone mentions the year 2000. And the comparison makes sense. But there is a data point about the big AI rounds of the last three years that completely breaks that analogy: the revenue is there, and it is real.
The dot‑com bubble was built on promises of future revenue that never materialized, at a time when only 6% of the world’s population had access to the internet. Today, Anthropic revealed in May 2026 a revenue run‑rate of $47B annualized. OpenAI reported approximately $10B in actual revenue during 2025 and projects reaching $100B or more in 2026. xAI closed a $20B Series E at a roughly $230B valuation supported by computing infrastructure that is already deployed.
These are not projections of a hypothetical future. They are current revenues growing at rates no tech company has ever recorded before in its history. The difference between 2000 and 2026 is not one of degree, but of nature. So if this is not a bubble, what is it? It is an accelerated concentration of economic power in a very small number of players. Which, depending on how you look at it, may be more worrying than a bubble.
Geographic Concentration Is Almost Total: 9 out of 10 Rounds Are US Companies
Of the 10 largest VC rounds in history, 9 correspond to US companies. The only exception is Ant Group from China in 2018. There is not a single European company. There is no Latin American one. There is no Indian, Israeli or Southeast Asian company.
This is not a market accident. It is the result of decades of ecosystem‑building, capital policies and talent strategies, and it is now generating a concentration of technological power that has not been seen since the early days of the internet. One particular aspect of this concentration is where the capital has come from: Middle Eastern sovereign funds such as MGX and Qatar Investment Authority, Asian funds like GIC and Temasek, European institutionals like Baillie Gifford. Global capital is flowing into the United States instead of building competitive local ecosystems.
For emerging ecosystems such as Latin America, the message is clear and urgent: the available regional capital is not being deployed to build its own AI infrastructure. Every dollar that an Asian or Gulf sovereign fund invests in OpenAI or Anthropic is a dollar that is not building an equivalent in Bogotá, Riyadh or São Paulo. The concentration is not only technological. It is geopolitical. And by the time local ecosystems want to compete, the gap may already be too large to close quickly.
