Takeaways
#- Lee believed earnings expectations underestimated how long AI infrastructure spending could remain strong.
- He distinguished the current infrastructure buildout from a future phase in which software captured more of the opportunity.
- He described hedge-fund risk limits as one reason reduced exposure could persist through a sharp market rebound.
- Pearlman urged investors to decide their risk rules during calm conditions instead of improvising under drawdown pressure.
Mike Lee's bullish market view on July 15, 2025 rested on a disagreement with the earnings expectations he saw around artificial intelligence. He thought investors were assuming the infrastructure spending cycle would slow much sooner than the demand required.
Speaking with JC Parets on The Morning Show, Lee singled out Nvidia, AMD and Marvell. His argument was that demand for AI computing, including future business use of agents, could keep the buildout running for several more years. He believed forecasts that implied a sharp growth slowdown within roughly fifteen to eighteen months were too cautious.
That was the central investment disagreement. A company could already be benefiting from AI and still be undervalued in his framework if investors underestimated the duration of its growth.
Infrastructure and software occupied different stages
Lee described an AI market in which infrastructure companies and chipmakers were capturing much of the immediate spending. He expected software to account for a larger share of the opportunity as the technology spread into business applications.
Palantir was his existing software example. He acknowledged its demanding valuation while arguing that a much larger AI software market could eventually support substantially more revenue if the company maintained its position. Execution mattered to that scenario; the expected market expansion alone would not deliver the company's share of it.
He said he still owned Palantir after buying it at considerably lower prices. That history distinguished an appreciated holding from a fresh purchase at the valuation discussed on the show.
CoreWeave was another company he favored as a more concentrated exposure to AI cloud infrastructure. Lee contrasted that business with buying a diversified technology company partly for its cloud operation. He was enthusiastic about its possible revenue growth.
Positioning could help explain a recovery investors missed
Lee also offered an institutional explanation for why some investors had struggled to participate in the rebound from the spring selloff.
He described how a hedge fund's risk controls could reduce capital available to multiple teams after losses elsewhere in the firm. A team might then have to cut exposure even if its own positions were performing reasonably well.
After a drawdown, tighter limits could make it difficult to add risk fast enough to catch a sharp recovery. In Lee's account, that mechanism helped explain underexposure without assuming every investor had simply chosen to remain bearish.
He coupled that view with a forecast for substantially lower short-term interest rates over the following year. The AI earnings thesis, anticipated policy support and price momentum reinforced one another in his outlook.
Pearlman separated riding the trend from preparing for its end
Later in the episode, Phil Pearlman returned to how an investor might behave in a strong market. He recalled the persistence of the late-1990s technology advance: a trend that looked extended could continue much longer than expected.
His response was to retain a method for managing risk while allowing the trend to run. He gave a rising 200-day moving average as one possible reference, without presenting a single rule as suitable for every investor.
More pointedly, Pearlman argued that a calm period was the time to make the plan. Waiting until a portfolio was falling and emotions were elevated made decisions harder. The plan needed to fit the investor's circumstances and remain usable when conditions changed.
Parets supplied a personal example: his family regularly contributed to the S&P 500 and had already decided how to increase contributions after a sufficiently large monthly decline. Pearlman focused on the advance decision. Parets would know what he intended to do before the stressful day arrived.
Questions and answers
#Questions answered in the original 2025-07-15 episode.
What was Lee’s main disagreement with AI earnings expectations?
He thought the market anticipated a sharp slowdown too soon. He expected computing demand and business adoption to support infrastructure spending for several more years.
How did Lee distinguish AI infrastructure from software?
He described chipmakers and infrastructure providers as capturing much of the immediate spending, while expecting software to gain importance later. Palantir was an existing holding he used to illustrate the software opportunity.
Why did Pearlman prefer making a plan while markets were strong?
He thought investors could make clearer decisions before fear and losses became dominant. The point was to choose rules suited to their circumstances in advance, then use them when conditions became difficult.
Adapted by Stock Market TV Editorial from the original video. Source date: . Article published: .