Rockefeller Capital CEO Greg Fleming on managing new-wealth AI founders and navigating the private market bubble

Greg Fleming on AI wealth, private market froth, and the limits of one-cycle thinking

Greg Fleming built Rockefeller Capital Management from a thin shell into a full-service multi-family office over nine years, starting in 2018 with Viking Global Investors as his founding backer. The firm now operates across more than 30 wealth centers in over 50 cities. Its sweet spot is families with assets up to a couple of billion dollars — wealthy enough to need sophisticated advice, but not wealthy enough (Fleming puts the threshold at north of $10 billion) to justify a fully staffed internal family office.

The AI founder opportunity

One of the firm's fastest-growing client segments is new-wealth founders at AI companies. Fleming says even younger clients want comprehensive advice immediately — not just investment allocation, but philanthropy structures, generational wealth planning, and tax strategy. The US gives away roughly $600 billion annually in philanthropy, and he says almost every family the firm works with wants a giving plan from the start, including people who have only recently made significant money.

The sheer pace of company formation is helping. Fleming cites 6.2 million new businesses expected to be created in the US this year, which he describes as a record, and argues the geographic spread of wealth creation — beyond New York and Boston into Nashville, Charlotte, and Florida — is exactly why the firm built local offices rather than hub-and-spoke around a single city.

“We have 6,200,000 new companies being created in the United States this year, which will be a record. This entrepreneurial surge has been great for us in terms of creating clients all over the country... Not every dollar is gonna get the return that it's looking for when you have a time like this with this kind of massive investment.”

Private market froth

Fleming is candid that parts of the early-stage private market look like prior cycle excesses. Companies going from zero to $2 billion valuations in months, with thin revenue and investment theses built around talent density rather than product, rhyme with what he saw in 1999 when companies were priced on clicks. He stops short of calling it a bubble outright, but he's close: "A lot of these companies are not going to get the valuations that they're being awarded now."

His practical response is diversification and position sizing. Only a defined portion of any client's capital goes into speculative private positions, and the firm tries to tilt toward managers who have actually seen a cycle go wrong, not just one sustained upswing.

The psychological risk he flags is arguably more interesting than the valuation math. Managers who have put up strong returns for several consecutive years, and clients who got into SpaceX early and now apply that mental model to every private deal, stop being able to imagine failure. Fleming's advisers are explicitly coached to hold that line — "let's stick with the plan" — against clients whose reference point is 30-to-50x outcomes as a normal expectation.

The macro picture

On AI broadly, Fleming's view is that the technology is real and the productivity effects will justify the capital. Output per hour is rising in some industries without employment rising, which he takes as early confirmation. The risk isn't that AI fails — it's that returns will be highly differentiated, with clear winners and losers, rather than a rising tide lifting everyone. He draws the comparison to the internet era, where the bubble burst but the underlying companies that survived became transformative.

The one structural concern he flags is the US fiscal debt position, which he describes as historically unusual and a potential driver of elevated long-term rates. He doesn't see it as an AI-specific problem, but as a macro constraint that investors shouldn't ignore.

AI inside the firm

Rockefeller is applying AI internally as a productivity tool, not a client-facing one. The human adviser stays front and center — Fleming says that will hold for "decades and decades" — but the work around the client relationship is being automated. The most concrete example he gives is meeting workflow: agents handling prep, post-meeting follow-up, and ensuring investment instructions are executed without advisers manually chasing each step. The goal is to free adviser time for more clients, not to replace the relationship.

That's a conservative but defensible posture for a business where the product is trust.