Interview with Wall Street's Jay Hatfield: Energy Crisis Impacts and AI Infrastructure Prospects
Wall Street Veteran confirms market's forward movement
Every now and again, comparing notes with smarter people is in order. Jay Hatfield, CEO of Infrastructure Capital Advisors, spoke with me from an advisors conference in Florida, fitting in the interview between sessions. The conversation covered the Middle East oil crisis, midstream’s energy upside, and AI infrastructure investment. We also discuss some nuanced ways to invest in AI and tech to minimize volatility. We spoke ahead of the big tech earnings tape.
A few key areas are highlighted.
On oil prices.
Hatfield uses a rule of thumb: every one million barrel change in supply moves the price by five dollars. With sixteen million barrels per day offline due to the conflict, the math puts the worst-case ceiling at $140 and the best case — full resolution — at around $70. He pegs the equilibrium at $100. Markets won’t price in the worst case, he argued, because they assume the Strait of Hormuz won’t stay closed indefinitely. If negotiations drag on without a resolution, prices will drift higher; if the conflict ends, they fall toward $70, which is roughly what futures are already pricing in.
He did not believe that high oil prices would seriously damage the U.S. economy. Electricity is priced off natural gas, which is cheap domestically. Gasoline accounts for about 3% of consumer spending — meaningful but not decisive. The AI-driven economic boom, in his view, more than offsets the headwind.
Video Interview: Energy Crisis Hits US Less Hard: AI Infrastructure Plays On
(I have interviewed Hatfield on energy and markets periodically since 2014.)
On midstream MLPs.
The master limited partnership (MLP) space is where his firm recommends investors play the energy dislocation. The sector was out of favor for years, left behind when tech boomed. And it still hasn’t caught a real momentum bid despite the tailwinds from AI power demand and energy security concerns. He sees that as an advantage. Energy Transfer trades at roughly 13X earnings with a 7% yield. (His firm’s AMZA ETF fund yields over 9%.) If the war ends, MLPs might fall three or four percent. Upstream oil stocks could fall 10-15%. That’s the trade-off he forsees.
(Paywalls are to be removed some days/weeks after paid or founder subscribers, based on content. The occasional, organically-derived sponsored post, adhering to independent editorial standards, is possible in future and flagged as such to support the work efforts.)
On AI infrastructure.
Hatfields approach is to invest in the AI infrastructure providers rather than attempt to pick winners among software companies. Amazon, Microsoft (for its infrastructure scale), Broadcom and Marvell (for ASIC intellectual property) are the names he favors. Data center REITs like Digital Realty offer a lower-volatility entry point. His core point: nobody builds a data center without signed contracts, so there’s no speculative overbuild risk comparable to the fiber bubble of the late 1990s. Demand for compute is broad enough that if one counterparty defaults, another takes the capacity.
On financing the AI infrastructure build-out.
Hatfield dismisses the short thesis that there isn’t enough capital. The U.S. capital market is not only banks — it includes public bonds, private credit, private equity, and strategic capital. When banks hit concentration limits on Oracle exposure, a 144A bond offering picked up the slack for a major Michigan data center deal. That kind of market depth doesn’t exist in Europe or Asia.
On market efficiency.
The markets get it right over the long run, Hatfield suggests, and badly wrong in the short run. Lockheed Martin dropped $100 in two weeks with no fundamental reason — collateral damage from capital rotating into semiconductor momentum trades. At 16X earnings, he thinks it should trade at 20, and his firm is buying it. The semiconductor momentum trade, in his view, is tired. MLPs still have fundamentals that aren’t fully reflected in prices.
His closing point: patient investors who focus on valuation rather than momentum will find plenty of opportunity in the current environment, particularly in midstream energy and AI infrastructure. The bear theses that circulated in recent months — no AI applications, circular financing, insufficient capital — have largely been wrong, and the noise around them has created entry points that a shorter-term focus would have caused investors to miss.