Ajinomoto, a Japanese food company famous for developing MSG (monosodium glutamate), a popular food seasoning, is one of the most important players in the AI industry. Yes, you read that right - the AI industry.
Why? Because in complex processors (like Nvidia GPUs, AMD server chips, and custom AI accelerators), microchips are not soldered directly to motherboards. Instead, they sit on an intermediate substrate composed of multiple-layered, microscopic circuits. In the 1990s, Ajinomoto’s researchers realized that a resin formulation derived from MSG made a far cleaner, more stable, and easier-to-apply substrate insulator than the liquid inks chipmakers were using at the time. And that’s how a food seasoning is supporting the “AI bubble” as we know it.
While we often think of AI models when we think of investing in the industry, we do not realize that there are other products and services that form the foundation of Artificial Intelligence. And the mystery that’s got investors scratching their heads right now is: Where will the next layer of value actually accumulate?
That was one of the key questions explored during the Family Office Investment Outlook Panel, featuring investors including SooMan Wolffs, Amanda Reed, Bryant Hayward, Anurag Chandra, and Gordon Chan, as they discussed how family offices and institutional investors are thinking about AI, infrastructure, and the opportunities emerging beneath the most visible names in the market.
One conclusion stood out: AI investing may be moving beyond the companies everyone already knows.
The Opportunity Is in the Gap
One of the panel's most compelling observations was that the opportunity may lie in the gap between the pace of AI companies’ growth and the pace at which the infrastructure supporting them can be built.
During the discussion, the panelists identified two major bottlenecks: time to power and time to deploy. Hyperscalers may need infrastructure within two to three years, while the broader ecosystem can take four to seven years to provision it.
As Bryant Hayward explained, “In that gap lies the opportunity.” This situation has created a different way to think about AI investing.
Rather than asking only, “Which AI company will win?”, investors can ask: “What does AI need to keep growing — and who is solving that problem?”
AI's Growth Is Creating a Much Bigger Ecosystem
The scale of the AI buildout is already changing the investment landscape.
During the panel, SooMan Wolffs discussed the strength of the current AI cycle, pointing to estimates that AI investment and spending accounted for 50% of US’ GDP growth at the end of 2025, rising to 60% in Q1, with projections potentially reaching 70% by the end of the year. He used those figures to illustrate the extraordinary amount of capital and economic activity now being directed toward AI and its infrastructure.
The broader point is that AI, as an opportunity, is becoming too large to think of as simply a software story. It is increasingly a physical infrastructure story as well.
Look for the “Picks and Shovels”
Data centers need electricity. They need cooling. They need connectivity, chips, land, and increasingly sophisticated power infrastructure.
The panel discussion highlighted transformers, modular energy sources, and other “picks and shovels” as areas worth watching as the grid struggles to keep pace with new demand.
For investors, this creates an important distinction. A company doesn't necessarily need to build the next major AI model to benefit from AI. It could provide a critical component, technology, or service that multiple AI companies need. That can potentially create a more diversified way to participate in the AI growth cycle.
Power May Become One of AI's Biggest Constraints
Among the biggest challenges discussed by the panel was access to power.
As data-center requirements increase, securing electricity supply and connecting new facilities to the grid can become limiting factors. This is pushing companies to explore alternative energy sources and new approaches to infrastructure.
The panelists also pointed to nuclear energy — including both fusion and fission — as areas attracting significant interest. They emphasized, however, that investors should bring in outside expertise when entering highly technical sectors rather than becoming what one panelist called “venture tourists in nuclear.”
An important lesson from this circumstance is that when a rapidly growing industry creates a bottleneck, the bottleneck itself can become an investment opportunity.
Here’s What You Must Do If You’re An Entrepreneur
If you’re an entrepreneur trying to hop on the AI bandwagon, this shift creates opportunities well beyond building another AI application.
The most interesting companies may be those solving the problems that become more urgent as AI adoption accelerates:
- How do we deliver power faster?
- How do we build and deploy data centers more efficiently?
- How do we manage increasing cooling requirements?
- How do we move data faster?
- How do we build more efficient AI hardware?
- How do we apply AI to industries where specialized expertise creates a defensible advantage?
In other words, you don't necessarily have to build the next AI giant. You can build something the AI giants need.
Investors, Look One Layer Deeper
Suman Wolfs argued that investors should distinguish between having an AI thesis and actually deploying capital against it, pointing to a gap between the number of families identifying AI as a priority and those actually allocating to venture, private equity, or other growth opportunities where AI is developing.
That gap highlights an important challenge for investors: recognizing a trend is not the same as knowing where to invest within that trend.
The next wave of value creation could come from businesses providing the infrastructure, technology, and specialized solutions that enable AI to scale.
*This article is inspired by the discussions in the Family Office Investment Outlook Panel during the Summer Expo 2026.