The Managers Behind the Biggest AI Winners

A look at investors in public and private AI companies + Situational Awareness profile

March 31, 2026

Private Holdings – Now in OWL

 

AI has driven an influx of capital across markets as it continues to permeate nearly every facet of daily life. Some of the most well-known AI companies are still private, which is why OWL is excited to share that we have added private holdings to our manager profiles.

 

OWL users can now see a given manager’s disclosed private holdings by navigating to its “Positions” tab on the left-hand menu under “Manager Views” and clicking “Private Positions” at the top.

 

 

On that tab, users will find a manager’s private holdings, the current company stage, and the date and stage of the round that manager last participated in. In addition, users will see a table of the manager’s private co-investors and how many deals/which companies they have invested in alongside.

 

 

OpenAI and Anthropic both received backing from blue-chip venture firms early on in their fundraising history. Users can easily see which managers backed a private company in each of its rounds by clicking on the company name.

 

In addition, users will be able to quickly reference which rounds each manager participated in, notated by a check mark in the fourth column from the left. As an example, below we include OpenAI’s fundraising rounds, including Y Combinator’s Pre Seed investment:

 

 

Below is a handful of the earliest backers of OpenAI and Anthropic.

 

Private Market Winners

 

Khosla Ventures 

  • Location: Menlo Park 

  • Founded: 2004 

  • Regulatory AUM: $14.2B 

  • Key Person: Vinod Khosla 

  • Earliest OpenAI investment: Seed (2019) 

  • Disclosed LPs: NYS Common Retirement Fund, University of California, The Heinz Endowments 

Menlo Ventures 

  • Location: Menlo Park 

  • Founded: 1976 

  • Regulatory AUM: $5.8B 

  • Key Person: Matt Murphy 

  • Earliest Anthropic investment: Series C (2023) 

  • Disclosed LPs: Mass PRIM, Alaska Permanent Fund, LACERS 

Spark Capital 

  • Location: Boston 

  • Founded: 2005 

  • Regulatory AUM: $13.4B 

  • Key Person: Alex Finkelstein, Santo Politi 

  • Earliest Anthropic investment: Series C (2023) 

  • Disclosed LPs: UTIMCO, California Teachers, Hillman Family Foundations 

Y Combinator 

  • Location: Mountain View, CA 

  • Founded: 2005 

  • Regulatory AUM: $16.2B 

  • Key Person: Garry Tan 

  • Earliest OpenAI investment: Pre-seed (2016) 

  • Disclosed LPs: MIT, University of Pennsylvania, UTIMCO

 

While the most well-known AI companies remain private, a broad ecosystem of public companies is already capturing significant value from the AI boom. In fact, many hedge funds and institutional investors have generated strong returns not by backing the model builders themselves, but by investing in the infrastructure that makes AI possible. This week, we highlight the managers capitalizing on those upstream opportunities.

 

It's easiest to classify these companies by how they interact with the tier one companies like OpenAI and Anthropic.

 

Tier two typically includes the “engines” that the models run on - semiconductors. The semiconductors are designed by blue-chip companies like Nvidia, Advanced Micro Devices (AMD), and Broadcom and manufactured primarily by Taiwan Semiconductor Manufacturing Company (TSMC), but also historical technology manufacturing leaders like Samsung.

 

Tier three of the AI landscape includes the “fuel tanks” or the storage devices used to house the massive amounts of data that AI models generate. This tier includes companies like Sandisk, which was acquired by Western Digital in 2016 for $19 billion before being spun out again in early 2025, SK Hynix, Micron and Western Digital itself.

 

For the purpose of this analysis, we opted to exclude the large public companies with AI models (i.e. Alphabet, Amazon, Meta, Microsoft) given the breadth and diversification of their businesses.

 

Outside of those tiers, a number of energy companies have benefitted from the AI boom as the increasing number of data centers require more and more power to run the models’ processes as AI adoption rises. A combination of legacy energy providers like EQT and newer, more AI-focused energy companies like Bloom Energy can be found in the portfolios of many of the managers below.

 

Public Market Winners

 

In last week’s newsletter users received a detailed list of public equity and long/short managers that have benefited in the public markets from the recent AI boom, below is an abridged version:

 

Analog Century Management 

  • Location: New York 

  • Founded: 2018 

  • Regulatory AUM: $1.7B 

  • Key Person: Val Zlatev 

  • Top AI-related Public Holdings: Lumentum Holdings, Advanced Energy Industries, Applied Material

Castleknight 

  • Location: New York 

  • Founded: 2020 

  • Regulatory AUM: $2.7B 

  • Key Person: Aaron Weitman 

  • Top AI-related Public Holdings: Micron Technology, Modine Manufacturing, Vertiv Holdings 

  • Disclosed LPs: The Macmillan Family Foundation, Peter and Carmen Lucia Buck Foundation, Discovery Capital Holdings 

Rivermont 

  • Location: New York 

  • Founded: 2020 

  • Regulatory AUM: $413m 

  • Key Person: Ross Gough 

  • Top AI-related Public Holdings: Seagate, Western Digital, TTM Technologies 

Situational Awareness 

  • Location: San Francisco 

  • Founded: 2024 

  • Regulatory AUM: $383m 

  • Key Person: Leopold Aschenbrenner 

  • Top AI-related Public Holdings: Bloom Energy, Lumentum Holdings, Sandisk 

Value Aligned Research Advisors 

  • Location: Princeton, NJ 

  • Founded: 2022 

  • Regulatory AUM: $542m 

  • Key Person: Ben Hoskin and David Field 

  • Top AI-related Public Holdings: Lumentum Holdings, Vertiv Holdings, CoreWeave 

 

Altogether, these managers have the highest concentration in storage-related companies. Semiconductors and components that go into them (optical components and printed circuit boards) come in second, followed by digital infrastructure (the hardware and software systems that enable data centers to function), energy, and AI Infrastructure (cloud services and other software infrastructure specifically designed for AI use).

 

 

Situational Awareness 

Situational Awareness has made headlines in the past few months for its founder’s impressive, albeit short and somewhat controversial, history.

 

Leopold Aschenbrenner founded Situational Awareness in June 2024 to capitalize on the attention that his self-published manifesto on AI gained a few weeks before. He wrote the paper, titled Situational Awareness: The Decade Ahead, just two months after being fired from his role on OpenAI’s “Superalignment” research team over an alleged information leak, which he contests the validity of.

 

Aschenbrenner joined OpenAI from the FTX Future Fund, which was the philanthropic grant-making initiative of the now-bankrupt crypto exchange. He graduated from Columbia University as valedictorian at 19 years old with dual bachelor's degrees in mathematics-statistics and economics.

 

While at Columbia, Aschenbrenner won a grant from Emergent Ventures, the impact-focused fellowship program managed by economist Tyler Cowen, who referred to Aschenbrenner as an “economics prodigy.”

 

Aschenbrenner also co-founded Columbia’s effective altruism (EA) chapter, a movement based on using research and evidence to produce outcomes that benefit the most people possible, interned at the Global Priorities Institute, an EA-minded research center at the University of Oxford, and wrote essays for Works in Progress, a newsletter and publication on scientific, technological, and economic progress backed by Stripe.

 

Now 24, Aschenbrenner’s thesis of effectively trying to build an AGI-focused “think tank inside of a hedge fund,” as he said on a podcast at the time of his fund’s launch, seems to be resonating with investors.

 

By the time his firm was a year old, Aschenbrenner reportedly managed over $1.5 billion of capital including anchor commitments from Stripe co-founders Patrick and John Collison, tech founder and investor and current head of Meta’s Superintelligence Lab Nat Friedman, and tech entrepreneur and former head of Apple’s AI division Daniel Gross.

 

Situational Awareness has so far raised approximately $2 billion as of March 10th , which can be seen on OWL’s manager-specific activity feed.

 

Situational Awareness filed its first 13F in Q4 2024 showing roughly $250 million invested across six long positions. Today, it has built a portfolio of at least 23 disclosed positions across every sector related to AI and its development, representing approximately $3.9 billion (excluding puts and calls).

 

 

Aschenbrenner was joined by chief operating officer and limited partner Nicholas Gross-Whitaker and director of research and limited partner Carl Shulman at inception. The team was expanded a few months later in early 2025 with the addition of chief compliance officer Chris Zellner, former co-founder of L/S equity healthcare specialist Asymmetry Capital and previously a longtime executive at Merrill Lynch, where he most recently served as director of global markets financing and futures. 

 

 

 

Other News & Events

About Old Well Labs

OWL is an intelligence platform built for allocators, by allocators. Leading endowments, foundations, and family offices use the system to find, monitor, and connect with thousands of fund managers globally. OWL's analytics engine has collected over one billion data points from 65 countries. We make it easy for allocators to find and track information about the managers they care about – not just positions but also performance analytics, people data, business information, and details about the manager investments of other allocators.

Disclaimers

Returns represent the return on invested capital of publicly disclosed long positions, as calculated by OWL. Actual returns may vary based on a number of factors, including (but not limited to) undisclosed positions, short exposure, non-equity holdings, cash holdings, and lagged disclosure of positions.

This newsletter and the material on the Old Well Labs platform are for informational purposes only and should not be considered investment advice or a recommendation of any particular security, manager, or strategy. Old Well Labs shall not be liable for any investment gain or loss that may occur from the use of this material. No part of this material may be reproduced in any form or used in any publication without express written permission from Old Well Labs.