Alejandro Betancourt López Made a 20x AI Bet Five Years Before the Boom

By the third quarter of 2025, artificial intelligence had consumed more than half of all global venture capital. That’s not hyperbole. According to the WIPO Global Innovation Index, AI captured 53% of total VC deal value worldwide, with Northern America alone absorbing roughly $162 billion. Single funding rounds had ballooned to figures that would’ve seemed absurd a decade earlier: Anthropic closed $13 billion in one raise; xAI secured $10 billion shortly after. The asset class had become, by any measure, the defining bet of the decade. Every major pension fund, sovereign wealth vehicle, and endowment was scrambling to build AI exposure, often at valuations that bore little resemblance to current revenues.

The investors who profited most from this wave were, for the most part, the earlier movers. They committed capital before large language models entered mainstream conversation, before ChatGPT made AI a household acronym. Among them is Alejandro Betancourt López, the entrepreneur and investor whose family office, O’Hara Administration, placed what he has described as a “big ticket” position in AI around 2019 to 2020. By early 2025, that position had returned approximately 20 times its original cost. The return speaks less to luck and more to a particular approach: patient capital, concentrated conviction, and the structural freedom to hold through years of market indifference.

The Bet Before the Boom

Betancourt López hasn’t disclosed the name of the company in which he invested. He has cited confidentiality obligations. What he has confirmed, in multiple interviews, is the scale and the timing. “I have a big investment I made about five years ago in AI, and now it’s 20 times its investment,” he told one outlet in early 2025. The phrasing is matter-of-fact; the maths is not. A 20x multiple on a pre-boom AI position places the return among the strongest individual bets of the current cycle, comparable in percentage terms to what early backers of companies such as OpenAI or DeepMind achieved before those firms reached institutional valuations.

The timing matters. Between 2019 and 2020, AI venture funding existed in a fundamentally different climate. Total global VC into AI-related companies was a fraction of what it would become by 2024. Generative AI hadn’t yet demonstrated consumer applications, and the prevailing institutional view treated machine learning as a niche within enterprise software rather than a platform shift. Transformer architectures had been published but remained the concern of research labs, not capital allocators. Committing substantial capital at that point required either deep technical conviction or a willingness to bet on long-term structural trends without short-term validation.

Betancourt López appears to have relied on the latter. “What is AI? It’s a machine that thinks faster and finds solutions faster. So AI just makes everything more efficient. It’s not only in energy. In anything,” he has said. That framing, efficiency as the core thesis, is notable because it sidesteps the hype cycles that have characterised much AI discourse. Rather than betting on a specific model architecture or a particular application layer, the position was anchored to a broad productivity thesis: machines that accelerate decision-making will find buyers across every sector. It’s the kind of reasoning that ages well precisely because it doesn’t depend on any single product roadmap.

The Family Office Advantage

O’Hara Administration, the investment vehicle through which Betancourt López operates, was founded in 2014 as an international investment group and family office. Its structure matters here. Family offices occupy a peculiar position in the capital markets: they manage private wealth, answer to no external limited partners, and face none of the quarterly reporting pressures that constrain institutional funds. This freedom has made them increasingly prominent in venture capital. According to data compiled by VCStack, family offices accounted for 10% of all VC deals in 2025 and represented approximately 31% of all startup funding.

The structural advantages are concrete. A traditional venture fund operates on a 10-year lifecycle with defined deployment periods; a family office can hold a position indefinitely. A fund manager who places a bet in 2019 faces pressure to show markups by 2022; a family office principal can wait for the thesis to compound without triggering governance reviews or LP anxiety. That’s a real difference. Alejandro Betancourt López held his AI position through the pre-boom period, through the initial GPT-3 excitement of 2020, through the quieter stretch of 2021 and 2022, and into the explosive revaluation that began in late 2023. That holding period, roughly five years before the 20x materialised, is one that few fund structures would have accommodated without pressure to exit or distribute.

A JPMorgan survey found that 65% of family offices cited AI as their primary investment priority, a figure that reflects both the returns already generated and the conviction that the cycle has further to run. O’Hara’s early entry positions it differently from the majority of these allocators, most of whom didn’t begin committing capital until after the generative AI breakout of late 2022. The difference between a 2019 entry and a 2023 entry, in terms of valuation multiples, is often the difference between a 20x return and a 2x return. That gap is enormous. Family offices that arrived late still benefited from the sector’s momentum, but the magnitude of their gains was compressed by the valuations at which they entered.

A Pattern of Asymmetric Bets

The AI position didn’t emerge from a vacuum. Betancourt López has described an investment philosophy that explicitly favours high-conviction, high-asymmetry positions. “I don’t swing for first base. I always swing for a home run,” he told Analytics Insight. The baseball metaphor carries a mathematical logic he has articulated elsewhere: out of 10 speculative positions, “if two of them go well, they pay for the eight and make you a good profit for everything else.”

This is recognisable as a variant of the venture capital returns distribution, the well-documented skew in which a small number of winners generate the vast majority of portfolio returns. What distinguishes the O’Hara approach is the willingness to apply this framework at scale with concentrated positions rather than diversifying across dozens of small cheques. A “big ticket” bet, by definition, concentrates risk; when it works, it also concentrates returns. The discipline lies not in avoiding losses entirely but in sizing winners large enough that the portfolio arithmetic works even when most positions fail.

Earlier bets from Betancourt López follow the same pattern. His €50 million commitment to Hawkers, the Spanish sunglasses brand, in 2016 was a large position on a direct-to-consumer model at a time when the DTC thesis was still unproven in southern Europe. The bet required conviction that digital distribution could bypass traditional retail economics in a market where e-commerce penetration lagged behind northern European averages. His accumulation of VTC (vehicle for tourism with chauffeur) licences in Spain through Auro Travel represented a different kind of anticipation: regulatory foresight. By acquiring licences before the ride-hailing market matured in Spain, he built an asset base that became valuable precisely because supply was constrained. Uber acquired a 30% stake in Auro for €220 million in Feb. 2025. That acquisition validated a thesis that had required years of patience to bear fruit.

Each of these positions shared a common structure: capital deployed ahead of institutional consensus, held through periods of uncertainty, and exited or revalued after the market caught up to the underlying logic. The Hawkers bet, the Auro bet, and the AI bet all required the investor to be early and to remain committed when external validation wasn’t there. The returns, when they arrived, compensated for that period of ambiguity.

Portfolio Parallels and the Efficiency Thesis

O’Hara Administration’s current portfolio focus includes robotics and technology manufacturing alongside its AI holdings. These sectors share a thread that connects back to the efficiency thesis Betancourt López articulated about artificial intelligence. Robotics, at its core, applies computational intelligence to physical tasks; technology manufacturing scales production of the hardware that enables software-driven efficiency gains. The portfolio reads less like a collection of separate bets and more like a series of positions along the same value chain.

This coherence matters. It suggests the AI return wasn’t an isolated lucky trade. The same reasoning that led to a pre-boom AI position, that machines accelerating problem-solving will create value across industries, also led to adjacent positions in robotics and manufacturing. Whether those positions will generate comparable returns remains to be seen, but the analytical consistency is clear. The investor isn’t chasing sectors; he is following a thesis about where efficiency gains will compound over the coming decade.

The robotics allocation, in particular, mirrors the timing logic of the AI bet. Industrial robotics is currently where AI was around 2018 to 2019: well understood in theory, increasingly capable in practice, but not yet the subject of a broad capital rush. Humanoid robots, warehouse automation, and autonomous systems are attracting growing attention from both governments and corporations, but the venture capital frenzy hasn’t yet reached the intensity that AI funding hit in 2024. For an investor who has already profited from arriving before the crowd, the parallels are presumably difficult to ignore.

Timing, Conviction, and the Capital Structure

Venture capital’s rush into AI during 2024 and 2025 has created a paradox. The sector’s potential is arguably clearer than ever, with real revenue from enterprise deployments, measurable productivity gains, and a growing installed base of consumer-facing tools. Yet the capital flooding in has compressed future returns. Anthropic’s $13 billion round and xAI’s $10 billion raise imply valuations at which even a successful company may struggle to deliver venture-scale multiples to its latest investors. The best AI returns, measured as multiples on invested capital, are likely already behind us, locked in by those who committed before 2022.

Betancourt López’s 20x return illustrates a broader principle about private markets: the greatest rewards accrue to capital that arrives before consensus forms. That’s how it works. Institutional investors, by their nature, move after consensus. They require data, diligence, committee approvals, and mandate alignment, all of which takes time. Family offices, particularly those run by principals with high risk tolerance and long time horizons, can act on conviction alone. The gap between those two speeds, measured in years, is often the gap between a generational return and an average one.

O’Hara Administration’s AI bet is a case study in that pattern. The capital was private, the timeline was open-ended, the principal was willing to take a concentrated position, and the thesis turned out to be correct. Five years and a 20x multiple later, the rest of the market has arrived at the same conclusion, only at valuations that make replicating that outcome extraordinarily difficult. What matters now for Alejandro Betancourt López is whether the same instincts, applied to robotics, manufacturing, and the next layer of the efficiency thesis, will prove similarly prescient when the next wave of institutional capital arrives.

Author Profile

Adam Regan: Deputy Editor

Features and account management. 7 years media experience. Previously covered features for online and print editions.

Email Adam@MarkMeets.com