Tech is Selling Off. AI Debt Fears Are Rising. So Why Does a $4B Manager See a Reverse Bubble?

Marvell Technology

Across the shows: 1 speaker in other episodes, past 30 days

• Jay Hatfield · Infrastructure Capital AdvisorsBullish

  Marvell is cheap at about 15x 2028 earnings with 60% growth; expects $450 by 2029.

  Hatfield calls Marvell his firm's favorite idea. Marvell signed an Oracle agreement implying about $120B of revenue, against about $12B of revenue this year. It is growing revenue about 60% and has guided to EPS of about $14 in fiscal 2028-2030. After an announced dilutive merger, the stock fell to $216, which he says prices it at about 7x 2030 EPS and under 18x a diluted 2028 figure. On PEG logic, the stock is about a quarter of a PEG. At 15x, the stock would be worth $450 by 2029 as earnings more than double.

  “if it trades at 15 times ... in '29 going into '30, it'll be worth $450 because that's ... it's gonna more than double in just two more fiscal years.”at 30:18

• Jay Hatfield · Infrastructure Capital AdvisorsBearish

  Marvell's dilutive acquisition cut the target to $350; the stock will be out of favor for a while.

  Hatfield's fund holds Marvell. A dilutive acquisition moves '28 EPS estimates from about $14 to about $12, so they cut their target from $425 to $350. He expects the stock to be out of favor for a while because investors are angry about the deal, so the fund will sell calls at the new target for two or three months. He still thinks the stock is worth the target even after the bad catalyst. The transcript reads the figures as $4.25 and $3.50; the EPS-times-multiple math implies $425 and $350.

  “But what we can do is when it gets announced, we redo our earnings estimates. We're estimating $14 for '28. This is a dilutive acquisition, so we can figure out pretty quickly it's gonna be 12. We've reduced our target from four and a quarter to three fifty.”at 58:30

CoreWeave

Across the shows: 57 speakers in other episodes, 14 bullish and 55 bearish calls, past 30 days

• Jay Hatfield · Infrastructure Capital AdvisorsBullish

  CoreWeave selloff is an overreaction; contracts are intact, and he sees an 80 fair value.

  Hatfield says the selloff over CoreWeave's financing was a massive overreaction. Project financing for data centers is basically investment-grade paper because the sites are leased to Microsoft and other investment-grade tenants. The company already has contracts for its data centers, superior technology, and about $13B of current EBITDA run rate. Rising financing costs will hurt economics, but he says the company is not going away. Worth $120? No. Worth zero? No. He puts fair value near 60 to 80, and his team's model gives 80.

  “So we think it was a massive overreaction and that those plants will get built... So we did the work, and we came up with an 80 target.”at 16:17

Cheniere Energy

• Jay Hatfield · Infrastructure Capital AdvisorsBullish

  Cheniere's LNG contracts at about $12/MCF versus $30 in Europe make it a buy, not a bubble.

  Hatfield says Cheniere exports natural gas under contracts priced around $12 per MCF, while European prices are around $30. He says the stock is not a bubble because when the contracts roll, Cheniere should make much more money if market conditions continue. He uses this as the test for bubbles: whether the contract price is below the market price of the commodity.

  “Cheniere exports natural gas. They have contracts about 12 in MCF. Right now in Europe, they're 30. So is that a bubble? No. You wanna go buy Cheniere.”at 23:02

Amazon.com

Across the shows: 27 speakers in other episodes, 32 bullish and 6 bearish calls, past 30 days

• Jay Hatfield · Infrastructure Capital AdvisorsBullish

  Amazon is cheap versus Walmart: 20% growth at about 20x earnings, with robots lifting margins.

  Hatfield compares Amazon to Walmart. Walmart grows about 10% and traded at about 45x earnings, a PEG of about 4.5, while Amazon grows about 20% at about 20x earnings, a PEG of about 1. Walmart has since de-rated toward a PEG of about 3. Half of Amazon's business is the same or better as Walmart's, with lower margins that robots can expand. He says the relative valuation has reversed partly, but Amazon still looks cheap.

  “Amazon grows at 20. It was trading at 20 times earnings... half of their business is the same or better, has lower margins so they can expand it with robots.”at 28:29

Apple

Across the shows: 10 speakers in other episodes, 2 bullish and 11 bearish calls, past 30 days

• Jay Hatfield · Infrastructure Capital AdvisorsNeutral

  Apple is a reasonable holding; 10% growth at 35x earnings means only about 10% annual returns.

  Hatfield says Apple is a perfectly reasonable stock that his firm owns some of, though it is not an index overweight because it is about 7% of the index. It grows about 10% and trades at about 35x earnings. If the multiple stays at 35x after five years, shareholders earn about 10% a year with little risk. That is fine, but not a stock that will crush it.

  “it's a perfectly reasonable stock, and we [owned] some... it grows at 10, and it trades at 35 times earnings.”at 29:34

Oracle

Across the shows: 24 speakers in other episodes, 4 bullish and 22 bearish calls, past 30 days

• Jay Hatfield · Infrastructure Capital AdvisorsBullish

  Oracle's capex cut was a bullish catalyst; Hatfield bought heavily on the announcement.

  Hatfield's fund had been in Oracle but exited after management overspent on capex. The stock fell about $20, and they re-added ahead of an analyst meeting, reasoning that the Ellison-run company would fix its capex problem. When Oracle announced a capex budget cut, they bought a large position and booked significant trading profits.

  “the second they announced that they're cutting their CapEx budget, we bought a ton and made a huge amount of day trading profits.”at 57:27

US stock market outlook

• Jay Hatfield · Infrastructure Capital Advisors

  S&P 500 could reach 9,300 next year on 28 earnings, and 10,000 if rates fall.

  Hatfield's 8,000 target for this year is roughly reached, so he is updating. His next-year base case applies a 28x multiple to earnings, which he expects to rise about 13% to roughly 4,250 as companies reinvest at good returns on investment because of AI. That gives about 9,300, roughly 15% above current levels. A bull case of 10,000 needs the Strait of Hormuz to reopen and rates to fall about 50bp, which would also bring multiple expansion. He says the 8,000 target is also a sell signal, and he admits the outcome of the war and the midterms is unclear.

  “So if you do that math, you get to 9,300, which is about 15% up from here.”at 4:42

• Jay Hatfield · Infrastructure Capital Advisors

  The 2017 corporate tax cut permanently raised valuations, so historical market-to-GDP ratios are misleading.

  Hatfield addresses why the stock market-to-GDP ratio is at all-time highs. The 2017 corporate tax cut, from 35% to 21% (a roughly 40% reduction), lifted after-tax return on invested capital from about 13% to 15% and boosted earnings growth. A larger share of stock value shifted from payouts to growth, which he calls a sea change that throws historical ratios off. He says historical data should be corrected for it.

  “The value of the stock market to GDP is at all times highs. What changed? The corporate tax rate wasn't cut by, like, one or 2%. It was cut from the 35 to 21.”at 50:30

• Jay Hatfield · Infrastructure Capital Advisors

  Retained earnings compounding at 15% lifts average stock earnings growth to about 10% a year.

  Hatfield says the average stock retains about 70% of earnings and reinvests them at about a 15% return on capital. That implies earnings grow about 10% a year, and the rest of the return comes from payouts. He uses this to frame stock market returns as a casino where the house takes about 10% a year.

  “the average stock is retaining about 70% of earnings, is reinvesting those earnings at about 15%. That means earnings grow at 10.”at 1:00:55

US rates & the Fed

• Jay Hatfield · Infrastructure Capital Advisors

  Fed is tightening too far; neutral rate is about 3%, yet it is pushing toward 4%.

  Hatfield says the Fed is raising rates to 4% while inflation is arguably about 2%. He argues the neutral rate is about 3%, and that the Fed claims otherwise because it does not know what it is doing. He says the Fed should only raise rates when core inflation rises, not headline. He also argues the Fed misleads markets about its control over long-term rates, since its cuts and hawkish talk move the 10-year.

  “gonna raise rates to four. Inflation's arguably two ish. So the neutral rate is three. They claim it's not because they don't know what they're doing”at 32:11

• Jay Hatfield · Infrastructure Capital Advisors

  The Fed's 2% inflation target is arbitrary and too low; the Fed is incompetent.

  Hatfield says the Fed made up the 2% target, and that the economy grows most prosperously with inflation around 3-4%. He adds that the Fed uses flawed indices (PCE, which he says is misstated by 100 basis points and includes about 30% imputed data), and that its 2021 error came from ignoring real-time housing data while rates were at zero. He calls the Fed incompetent for these three reasons: the target, the indices, and forecasting.

  “The first is they made up the 2% target. It's way too low.”at 38:24

• Jay Hatfield · Infrastructure Capital Advisors

  Warsh is a poor pick now, but may partially reform the Fed and loosen the target.

  Hatfield says Warsh looks like a horrible pick because he is a hawk, though he is caught between hawks who dislike the president. Warsh has floated 'left of the decimal' inflation language, which Hatfield reads as a loosening of the target, but committee members have pushed back. Hatfield expects Warsh to bring better indices and more flexibility on the target and forecasting, and says he won't be as bad as he looks. He expects change through indices and adjustments rather than a formal target change.

  “I think that [Warsh] ultimately won't be as horrible a pick as he seems right now because he'll help at least partially reform the Fed, get better indices, and a little bit more flexibility on the target and forecasting.”at 46:55

US Treasuries & yields

• Jay Hatfield · Infrastructure Capital Advisors

  10-year yield should stall near 5%, since it trades about 100bp over the terminal Fed funds rate.

  Hatfield argues the term premium is about 100bp in most countries. US Fed funds usually sit 75-100bp above inflation, so the 10-year trades roughly 100bp over the terminal Fed funds rate. He expects the Fed to hike once more to about 4% with inflation near 2%, which points to a 10-year near 5%. He says futures do not predict cash rates, and he cites his earlier call that the 10-year would reach 5%. He is not a bull on Treasuries at these levels.

  “I think it's gonna stall out at five because the Fed is hell bent stupidly on raising rates one more time.”at 35:28

• Jay Hatfield · Infrastructure Capital Advisors

  Treasury selloff has one cause: a hawkish Fed, not multiple factors.

  Hatfield argues against momentum-style explanations of the Treasury selloff. He says there are not ten reasons the Treasury sold off; there is one, a hawkish Fed. He offers this as a simple rule of thumb for the Fed policy rate and Treasury yields.

  “There's not 10 reasons why the treasury sold off. There's one. It's because of a hawkish fed.”at 37:06

• Jay Hatfield · Infrastructure Capital Advisors

  Inflation-expectation-driven bond selloffs would act as monetary tightening that slows the economy.

  Hatfield says the fear that a higher inflation target would make the bond market sell off is the expectations theory of inflation, which he calls discredited. If expectations do rise and yields sell off, mortgage rates tighten and the economy slows, so the bond selloff itself is a huge monetary tightening. He treats that as the key dynamic.

  “Now inflation's gonna be way higher, and things will... The bond market sells off. That's a huge monetary tightening.”at 43:36

US tax & spending policy

• Jay Hatfield · Infrastructure Capital Advisors

  Corporate tax rates strongly drive growth; cutting them lifts savings, investment and growth.

  Hatfield says corporate tax rates are roughly 80% correlated with economic growth across the world, and that Europe including Scandinavia has similar low rates. Cutting corporate taxes raises national savings, investment, return on invested capital and earnings growth. He cites an IMF study: each 10% of national savings yields about 1.5% of growth. He says the pandemic masked this effect, but it is visible across countries.

  “Corporate tax rates are 80% correlated to economic growth across the world. ... Every 10% of national savings produces 1.5% of growth.”at 53:58

• Jay Hatfield · Infrastructure Capital Advisors

  A 2028 Democratic-socialist win and 30% corporate tax rates could cut stocks about 30% over time.

  Hatfield says he is worried about 2028, because a Democratic-socialist win would bring a 30% corporate tax rate. Using his tax-growth mechanism, he expects lower growth (about 1.5% rather than 3%) and a roughly 30% decline in stocks over time, not discontinuously. He notes that a wide base of 529-style accounts would leave everyone exposed to that drag.

  “So that's why I'm worried about '28 because if the democratic socialists take over, we're gonna get 30% tax rates, probably have a 30% decline in stocks over time, not discontinuously.”at 54:33

AI bubble & tech valuations

• Jay Hatfield · Infrastructure Capital Advisors

  Market has a reverse bubble: many cheap stocks trade at steep PEG discounts, not an AI bubble.

  Hatfield says that rather than an AI bubble, many cheap stocks trade at huge discounts to their PEG ratios, so they are on sale. He says the late-1990s was the real bubble, when AOL Time Warner traded at about 500x earnings. He says bubbles should be supported with data, and that his Marvell and Amazon comparisons show value is available in the AI stack.

  “everybody says, oh, where there's an AI bubble up. I'm like, there's an AI reverse bubble... so many cheap stocks out there trading at huge discounts to their PEG ratio that I think it's actually the polar opposite of that.”at 18:07

Europe's economy

• Jay Hatfield · Infrastructure Capital Advisors

  Europe lacks growth and technology, and France's fiscal path is a cautionary tale.

  Hatfield says Europe has no circular financing and no excess debt, but it also has no economic growth and no technology. He calls France arguably bankrupt, running out of other people's money. He presents this as a case study against socialism and says avoiding circular financing means accepting Europe's lack of growth. He argues that the US AI-driven capex cycle is the stronger path.

  “They also have no economic growth and no technology... And France is arguably going bankrupt.”at 19:07

AI infrastructure & data centers

• Jay Hatfield · Infrastructure Capital Advisors

  Contracted data center returns and compute pricing show no AI bubble; contracts are worth more than market.

  Hatfield argues AI data center debt is not a bubble. He says the market price of compute is about $40B per gigawatt, while long-term contracts are about $15B per gigawatt, so existing contracts are worth more than their price. Five-year contracts yield high returns on equity, around 20% to 30% for Oracle, and the power-plant analogy means sites rarely get abandoned, adding franchise value. He contrasts this with 2000s fiber and late-1980s speculative office building, where projects lacked contracts. He says Nvidia's investments in partners like Marvell gained a lot, so vendor financing is fine when credit is strong.

  “If you think this is all gonna be a bubble, tell me what the market price of compute is. It's about 40,000,000,000 per gigawatt. What are long term contracts? $15 per gigawatt.”at 21:49

AI chips & semiconductors

• Jay Hatfield · Infrastructure Capital Advisors

  Chip stocks lack contracted revenue, so the chip trade is fragile; a hedge fund blew up on it.

  Hatfield contrasts data center and power contracts, which run five to six years and generate high returns, with chip makers, which lack contracted revenue. Chips are good until they are not, and if demand drops, prices drop and chip manufacturing capacity is worth little. He says a hedge fund overdid the chip trade and blew up. He calls chip stocks a fragile trade relative to contracted infrastructure assets.

  “Blew up because they just overdid that trade. Like, chips are okay until they're not. Like, they don't have contracted revenue.”at 25:33

Gold & precious metals

• Jay Hatfield · Infrastructure Capital Advisors

  Negative on gold; war, a strong dollar and high rates are the drivers, not loss of confidence.

  Hatfield says he has been negative on gold. He argues the metal is driven by only two factors: a strong dollar and high rates during a war. He rejects the explanation that gold rises or falls because of fiat money or loss of confidence. He says that people invent many reasons for a move when there is usually one, and that the current setup is bearish for gold.

  “I've been negative on gold. ... When there's a war going on, the dollar's strong, rates are high. Those are the two factors that drive gold.”at 33:29

US deficit & government debt

• Jay Hatfield · Infrastructure Capital Advisors

  US bonds yield more than Canada and Germany because of the US deficit; spreads are already priced in.

  Hatfield tests whether the deficit drives US yields by comparing other sovereigns. Canadian bonds trade about 120bp below US bonds. He says Canada does not have better credit or better weather. The difference comes from Canada running a deficit of about 3% of GDP against the US at about 6%. He says global spreads already reflect this, and that US bonds are high relative to the rest of the world. France, which runs a similar deficit, trades close to US yields.

  “Their bonds trade a 120 basis points below ours. It's not a better credit... they run a 3% budget deficit. We run a six.”at 36:01

US inflation

• Jay Hatfield · Infrastructure Capital Advisors

  Real-time rent data show CPI inflation near 1.2%, well below reported figures.

  Hatfield says CPI shelter uses a renewal-based rent method that the BLS deliberately delayed by six months, adding roughly a year and a half of lag, so CPI runs about two years behind market rents. Using real-time rents, CPI is 1.2% year over year, versus 2.6% reported a month earlier. He says money supply is flat year over year, rents are flat and falling, and oil at $60 would push inflation negative, so any reasonable forecast lands near 1.2%.

  “They purposely delayed six months. ... If you use real time rents, CPI year over year is 1.2.”at 39:41

US housing

• Jay Hatfield · Infrastructure Capital Advisors

  Real-time rents are flat and falling, so housing inflation is far lower than CPI shows.

  Hatfield says the Fed controls housing through rates. Real-time rents are at zero growth and dropping, yet CPI shelter lags by roughly two years because of the six-month delay and renewal method. Using real-time rents gives a CPI of about 1.2%.

  “Housing is... Rents are zero and dropping.”at 47:16

Credit markets

• Jay Hatfield · Infrastructure Capital Advisors

  Triple-C bond market melted down because the Fed raised rates irrationally.

  Hatfield argues the Fed's rate hikes were irrational, given that inflation was overstated and the Fed was reacting to a flawed target and indices. He says the resulting higher rates caused stress in the CCC (triple-C) bond market, which he describes as having essentially melted down.

  “the triple c bond market essentially melted down because the Fed raised rates, irrationally raised rates.”at 47:38