Suggestions
Marcelino Pantoja
Vice President at Tribe Capital
Marcelino Pantoja is a venture capital professional and entrepreneur based in the San Francisco Bay Area. He is currently the Founder-in-Residence at Platform Venture Studio, a venture studio supported by PivotNorth Capital, where he helps build and scale startups. He also serves as the CEO and Founder of Measurement, a company he established in 2021, and is a Vice President at Amity Ventures, a role he has held since 2023.123
Pantoja has over 15 years of experience in venture capital and investment management. He previously worked as Vice President at Tribe Capital, Director at Costanoa Ventures, and Senior Investment Associate at Stanford Management Company, where he managed a $500 million direct investment portfolio spanning over 500 startups.12 Additionally, he has contributed to building and advising emerging venture funds that collectively raised and deployed over $2.4 billion into startups.2
He is a first-generation college graduate from Yale University, where he earned his BA and was a Varsity Oarsman on the Heavyweight Crew Team.2
Highlights
It took twenty years to build enterprise software and cloud infrastructure into a $1.6 trillion annual market. Enterprise software is the larger piece, about $1.2 trillion: the CRM platforms, the ERP systems, the collaboration tools that successfully replaced the meetings we used to have with meetings about the meetings we are going to have. Cloud infrastructure sits underneath at $419 billion, doing the unglamorous work of keeping all of it running. Every dollar was sold to a specific buyer against a specific budget, through a procurement process that someone was accountable for.
It was slow, boring, and real.
AI needs to surpass that combined total within four years. The $4 to $5 trillion going into new infrastructure requires roughly $2 trillion in annual revenue by 2030 to justify itself. This number is difficult to reconcile against technology budgets. The industry has responded by switching the comparison. If you measure AI against the $58 trillion the world pays people to work, $2 trillion appears as a modest 3 percent. The chart makes this move visually. It places AI revenue next to technology markets where it towers over everything, then next to global labor income where it practically vanishes. Both frames are true. The investment thesis is a bet on which one matters more.
The trouble with 3 percent of global labor income is that it describes a possibility, not a pipeline. Twenty years of enterprise software proved that technology markets grow by attaching themselves to budgets that already exist. AI is trying to attach itself to one that has to be created, company by company, role by role, within organizations that are still working out what the technology does for them.
AI does make people faster at their jobs. The evidence there is solid. Some companies are using that as justification for significant layoffs. Others are keeping the same workforce and capturing the extra output as margin. In both cases, the value stays inside those companies. Much of it will not flow back to the people financing the new data centers, which is good for corporate earnings but not so good for the return assumptions behind $5 trillion in concrete and silicon.
The sequencing is backwards from everything that came before. AWS added capacity as demand converted into contracts. Salesforce grew one seat at a time. Capital followed demand. AI has reversed the order, deploying over $650 billion this year into infrastructure for workloads that largely remain theoretical. And unlike fiber laid during the dot-com bubble, which sat dark for years after the bust but was still perfectly usable when demand finally caught up, GPUs depreciate in two to three years. The industry is building at the scale of a public utility using components with the lifespan of a smartphone.
The $2 trillion has to come from somewhere, and the somewhere is payroll. So far, payroll has not volunteered.

Here is what happened over the last century of U.S. public markets, now fully measured: $91 trillion in total wealth creation from 1926 to 2025. Of that, eight venture-backed companies that currently sit above a trillion dollars in market cap created $24 trillion, or 26% of the whole thing. Apple, Nvidia, Microsoft, Alphabet, Amazon, Broadcom, Meta, Tesla. Eight companies. A hundred years. A quarter of everything.
It gets stranger when you split the timeline. Through the end of 2016, total wealth creation was $43 trillion. Then, in just the nine years from 2017 to 2025, another $48 trillion. More wealth was created in the last nine years than in the preceding ninety-one. And those same eight venture-backed firms accounted for $21 trillion of that recent stretch, which is 43% of the period's total. Eight firms, nearly half of the wealth.
These companies now make up more than a third of the S&P 500. Eight of the top ten by market cap are venture-backed. This is the kind of concentration that makes index fund managers uncomfortable and venture capitalists insufferable.
But here is the thing: none of this required venture access. Every single one of these companies has been available in the public markets, for years, at a fraction of its current value. You did not need a fund. No lockups. No capital calls. Only a brokerage account and patience.
So why does anyone still invest in venture? Because the most important companies in the public markets were produced by an industry where most bets fail. The S&P 500 absorbed these companies. And now they account for more than a third of it.
Now those eight companies are collectively planning to spend somewhere around $650 billion on AI infrastructure this year alone. Some of them are issuing bonds to do it. The companies that defined the last decade of wealth creation are making the largest forward capital bets in corporate history, funded partly by debt, on a technology whose returns are still mostly theoretical. The concentration that built $24 trillion in wealth is now concentrating risk, too. Same companies, same story, different chapter.

