Why This Short Seller Doesn’t Want to Short the AI Bubble | Russell Clark

Alphabet

Across the shows: 63 speakers in other episodes, 57 bullish and 40 bearish calls, past 30 days

• Russell Clark · Capital Flows and Asset Markets (Substack)Bearish

  Google is very exposed to AI chatbots as search shifts to ChatGPT; ad brand power is weakening.

  Clark says Google is very exposed because he personally no longer searches on Google and instead uses ChatGPT for everything. He says the earnings showed this. He points to YouTube being packed with ads, which he reads as a sign that the advertising power of Google's brand is weakening. Google still has a lot of cash, so it has to keep investing in AI to defend itself. He sees this spending as existential defence, not a sign of strength.

  “Google in particular, I always feel like is very exposed because I don't even search from Google anymore. I just search everything on ChatGPT.”at 12:31

Oracle

Across the shows: 37 speakers in other episodes, 8 bullish and 34 bearish calls, past 30 days

• Russell Clark · Capital Flows and Asset Markets (Substack)Bullish

  Oracle's capex may be peaking; resumed buybacks could re-rate the stock higher, so it isn't a good short.

  Clark notes Oracle's share price is down about 50%, but the company has not cut capex, only reduced it, and its reported numbers are strong. Cash flow rose sharply. He thinks the share price weakness came from Oracle moving from buybacks to share issuance, which the market disliked. If Oracle says capex has peaked and it has cash for buybacks, the stock could re-rate quickly. That is why he sees it as interesting and why he thinks it is not a good short. Max adds that Oracle has just tapped equity and debt markets for $10bn each.

  “That's why I say Oracle sort of looks interesting because if they're saying that their CapEx has peaked out and they're going to have cash to do buybacks, that company could rapidly re rate higher.”at 35:18

Apple

Across the shows: 48 speakers in other episodes, 58 bullish and 37 bearish calls, past 30 days

• Russell Clark · Capital Flows and Asset Markets (Substack)Uncertain

  Apple's buybacks create value now, but its tech innovation has been marginal since Jobs; durability is doubtful.

  Clark notes that Apple has been the big winner from buybacks, which have been a major driver of the market since 2016. At current valuations and interest rates, buybacks would normally destroy value, though tax incentives may make them work. He argues that Apple has not produced a new era-defining technology since Jobs died. Its products have seen marginal changes and price increases instead. That is value creating for shareholders, but he questions whether it can continue in a technology sector where change is so fast, especially if the next era is AI agents.

  “But actually, from a tech perspective, it's been a real sort of marginal changes at best and then pushing up a price in its products as much as possible.”at 36:43

US Treasuries & yields

• Russell Clark · Capital Flows and Asset Markets (Substack)

  Pro-labor, deficit-driven policy will push long-end yields into the pain zone; bond market not yet pricing fiscal reality.

  Clark says yields were in a sweet spot but are now moving into a pain zone. He attributes this mainly to political pressure from big fiscal deficits and a focus on wages. He expects the midterm result to strengthen pro-labor and inflationary policy, and sees no tax increases coming, so the deficit will not fall. He expects the bond market to eventually take a longer-term view that fiscal consolidation will never happen, which would push yields up. He rejects Max's suggestion that AI issuance is squeezing out Treasuries, arguing instead that the US fiscal outlook implies much higher rates than current levels and that sovereign yields reflect the outlook for higher inflation and rising wages.

  “The market where I think that looks more interesting will be bond markets and when or if they sort of start taking a longer term view that fiscal consolidation is never going to happen”at 4:40

• Russell Clark · Capital Flows and Asset Markets (Substack)

  Bond yields have passed their sweet spot and reached the pain zone, where markets start pricing change.

  In a rising-yield environment, equities should suffer. France is a case in point: it is probably the worst market this year, and its yields have risen under bond pressure. Clark says yields were in a sweet spot before, but the market has now reached the pain point where it starts to price change. He says he expected this to happen earlier.

  “So there's a sweet spot for yields, which I think we've been in. And now I think we're getting to that sort of pain spot where markets go, Okay, something's got to change, and they start pricing change.”at 37:59

• Russell Clark · Capital Flows and Asset Markets (Substack)

  Rate cuts would push the long end of the bond market to sell off, because wage-driven inflation is not addressed.

  Clark expects the long end of the bond market to sell off as soon as the Fed starts cutting or considers cutting rates. The reason is that markets would conclude the Fed is not dealing with the real driver of inflation, which he describes as tight employment and rising wages. He treats wage-driven inflation as structural and persistent.

  “in my view, as soon as they start cutting rates or think about cutting rates, the long end of the bond market actually sells off because they're saying, oh, actually, you're not dealing with the real driver of inflation here”at 55:11

US inflation

• Russell Clark · Capital Flows and Asset Markets (Substack)

  Wage-driven inflation from pro-labor policy is systemic, not cyclical; prices won't fall back after energy spikes.

  Clark argues that when energy prices rise, businesses push prices up to protect margins, and those prices are never given back when energy falls. Wages also rise, and wages are very hard to cut. Policies that push wages up make inflation systemic rather than cyclical. He says he has seen no price decreases despite the recent decline in energy prices, so he was surprised when the Fed started cutting rates last year. He also notes that a strong US jobs report supports the view that prices will not fall.

  “When you have policies to push up wages, the inflation increase is really not cyclical anymore. It's a sort of systemic.”at 0:21

• Russell Clark · Capital Flows and Asset Markets (Substack)

  Wages, not oil, are the structural driver of US inflation, and the Fed misreads oil moves.

  Clark says oil moves markets in the short term, but the bigger driver of inflation is wages, which he thinks are rising. He argues that the structural cause of inflation is wages, and that the Fed reacts to oil, raising rates when oil rises and cutting when it falls. The Fed is therefore not addressing the real source of inflation, which is wage growth in tight labour markets.

  “So I feel like oil is like a distraction. In short term, it does move markets, but the bigger driver of inflation to me just seems to be wages, and I think that's going up.”at 54:39

US elections & politics

• Russell Clark · Capital Flows and Asset Markets (Substack)

  Midterms show voters want more money; pro-labor, inflationary policy likely to get another push.

  Clark says the midterm result is a rejection of the way Trump has run his second administration, not an endorsement of Democratic policy, since Democrats offered few policy specifics. He argues that voters ultimately want more money in their pockets, as they usually do when they vote for change. Since Trump is unlikely to approve tax increases, the only room for compromise is spending increases. He thinks the pro-labor and inflationary policies he has been calling for should get another push, and that the bond market will have to consider what a Democratic president and House, and extreme-left policies, mean for the 2028 nomination cycle.

  “For me, the big sort of takeaway from this is what I've been calling, so pro labor policies or inflation policies should probably get another leg up because I don't see Donald Trump approving any tax increases”at 3:57

• Russell Clark · Capital Flows and Asset Markets (Substack)

  Democrats are unlikely to unwind tariffs, industrial policy, pro-union policy or the deficit; only extreme policies may change.

  Clark expects Democrats to disagree with immigration policy but possibly soften its implementation. He does not expect them to change tariffs, industrial policy, pro-union policies or the fiscal deficit much. He views tariffs and industrial policy as designed to raise US prices and grow the economy, so he struggles to see them unwound absent a debt crisis. He says the policies that really took money from people's pockets were the problem, and those are the ones that may change.

  “I don't see tariffs really changing that much. I don't see industrial policy changing that much. I don't see pro union policies changing that much. I don't see the fiscal deficit changing that much.”at 49:36

AI infrastructure & data centers

• Russell Clark · Capital Flows and Asset Markets (Substack)

  Democratic data center restrictions would be a false trade; hyperscalers are fully booked, so demand persists.

  Clark thinks Democrats may restrict data center construction, but the China race gives the industry a strong counterargument, and hyperscalers could simply build elsewhere in Europe or Japan. He says hyperscalers keep saying they are fully booked for this year and next, so demand is still there and the politics will not stop it. He also reports that some people he has talked to think the administration over-permitted data centers so it could later sacrifice one to Democrats. Overall, he sees the politics as a trade that may be priced in short term but is ultimately false. He adds that AI's need for exponential compute growth does not change this.

  “As I can tell, the hyperscalers need more data centers. They keep saying that they're fully booked out for this year and next year.”at 6:02

• Russell Clark · Capital Flows and Asset Markets (Substack)

  No short on the AI build-out yet: rising GPU and cloud prices contradict a bust; risk-reward is finely balanced.

  Clark says industries overbuild for logical reasons, such as China's build-outs, and that the best shorts come from such overbuilding. He argues that hyperscaler CEOs are not idiots, which is where many short theses start. The bear case is that AI spending depends on capital that may dry up. But he sees no smoking gun: GPU, rental and cloud prices are rising rather than falling, which is the opposite of the usual signal that spending will stop. Equity markets are still funding the AI build-out, for example through SpaceX's $100bn equity raise for its own model. Governments may also support spending as a Cold War-style strategic race with China. He says the evidence is circumstantial, and that with short-selling risk this is not enough to act on, so he is still looking for a smoking gun.

  “Normally, causes spending to stop is prices falling, so GPU prices falling or rental prices falling or cloud prices falling. Now, everywhere I look, it's the opposite. It's going up.”at 16:44

US deficit & government debt

• Russell Clark · Capital Flows and Asset Markets (Substack)

  US budget deficit won't fall: no tax increases are coming, so any compromise will be on spending.

  Clark says he does not expect President Trump to approve any tax increases. Any bipartisan agreement over the next couple of years will therefore be on spending increases, so he cannot see the budget deficit coming down. He links this to the broader theme of governments spending more, which he sees as reinforced by the midterm result, and to the bond market eventually giving up on fiscal consolidation.

  “Where they will be able to meet in the next couple of years will be on spending increases. So I can't see the budget deficit coming down.”at 4:19

Credit markets

• Russell Clark · Capital Flows and Asset Markets (Substack)

  AI debt issuance is draining liquidity from other corporate credit by offering very high yields.

  Clark says the AI debt markets have squeezed out other corporate borrowers, taking liquidity from the rest of the corporate bond market. He cites SoftBank's offering, which paid about 7–8% on three-year money in Europe and Japan, a very high rate. Because AI-linked issuers offer such yields, they absorb demand that would otherwise go to other corporates. He sees this as one sign that credit markets are reluctant to lend more while equity markets still fund the AI boom.

  “I think the AI debt markets have definitely squeezed out other... is basically taking liquidity out of the other corporate markets because they're offering... if you take the SoftBank offering, it was like seven, 8% for three year money in Europe and in Japan, which is tremendously high.”at 22:45

Private equity

• Russell Clark · Capital Flows and Asset Markets (Substack)

  Alternative managers are already weak and face a shrinking pool of capital as foreign Treasury holdings decline.

  Clark's view is that the cost of capital is rising because the pool of capital is shrinking. Foreign reserves that China, Japan and others once held as Treasuries, and that flowed into private equity and sovereign funds, are declining. Governments are now spending and pushing rates up to protect currencies and keep wages high. As a result, the capital available to alternative asset managers is getting smaller, and corporates have to look for other sources of funding. He says alternative managers are already weak, so the short may be a bit late.

  “Basically, no one wants to hold treasuries anymore. And so the pool of capital available for those asset managers is shrinking.”at 25:14

US housing

• Russell Clark · Capital Flows and Asset Markets (Substack)

  Homebuilder margins could compress from about 15% to 5–7%, driving a substantial de-rating of the stocks.

  Clark says US homebuilders have been the weakest part of the market and look attractive as a short. Their margins are still very high, but they have not cracked. A sell-off at the long end of the bond market would crimp demand to some degree, though he does not expect a GFC-style problem. He expects builders to say that margins used to be about 15%, but labour costs, higher interest rates and energy prices have cut them to 5–7%. He thinks that would drive the stocks down substantially.

  “they come out and say, Do you know how we used to make 15% margin on our homes? Well, now because of labor costs, because interest rates have gone up, because energy prices have gone up, actually, we only really can make 5% to 7% margins”at 27:07

Consumer & retail

• Russell Clark · Capital Flows and Asset Markets (Substack)

  Franchise restaurant cash flows are reversing as franchisees are squeezed; many restaurant chains are weak.

  Clark says he still likes restaurants as a short-side idea. The franchise model gets capital from franchisees who buy the restaurants and pay a cut of sales. But those franchisees are being squeezed by higher energy costs and higher interest rates. He cites McDonald's last quarter, when it had to pay money back to franchisees for store renovations, so the cash flow is starting to reverse. He says many restaurant chains are very weak and that the wider debt built up to fund assets over the last few years has not been fully reckoned with.

  “But these franchisees are really feeling the squeeze of higher energy and higher interest rates. And so what we saw with McDonald's last quarter is they actually had to pay money back to the franchisees.”at 27:55

US dollar

• Russell Clark · Capital Flows and Asset Markets (Substack)

  A Fed QE programme to cap long-end yields would be bad for the dollar and add inflation pressure.

  Clark says the only policy he sees that could help Democrats with affordability is persuading the Fed to start QE to cap the long end of the bond market, since broad household grants would be offset by a sell-off in long bonds. He argues that QE would be pretty bad for the US dollar and would create inflationary pressure. He links this to the broader weak-dollar trade he sees forming as Europe and Japan move toward pro-growth policies.

  “the only thing I could see them do perhaps is convincing the Fed to begin a QE program to cap long end of the bond market. That would be pretty bad for the US dollar.”at 29:59

Banks

• Russell Clark · Capital Flows and Asset Markets (Substack)

  Japanese banks look attractive: large deposit franchises, still cheap, and rising yields let them earn more on deposits.

  Clark still likes Japanese banks because they have a large deposit franchise and the stocks are still very cheap. As capital becomes scarcer, deposit-rich banks benefit. Rising yields raise the return on reinvesting those deposits into higher JGB yields or, if they choose, higher Treasury yields. He notes Japanese banks were poor investments for about thirty years until recently, so the re-rating is a recent development he still finds attractive.

  “So I still quite like Japanese banks. They have a large deposit franchise. It's still very cheap.”at 38:26

AI chips & semiconductors

• Russell Clark · Capital Flows and Asset Markets (Substack)

  Chip prices are rising as capacity runs out, but equities lag because markets doubt capacity will stay tight.

  TSMC has reported rising semiconductor prices for years, and that increase has flowed into semiconductor equipment. Then prices passed into DRAM and NAND flash once production reached capacity limits, which drove huge share-price gains for DRAM and NAND makers this year. Clark notes that equities then decouple from the price rise, because the market keeps asking whether new capacity will arrive and whether pricing will fade. He is looking for other areas where full capacity could push prices higher.

  “What we've seen is TSMC has been reporting increasing semiconductor prices for years, and that's fed through into semiconductor equipment. We suddenly saw that sort of pass into DRAM and NAND flash.”at 39:00

Earnings & corporate profits

• Russell Clark · Capital Flows and Asset Markets (Substack)

  US buybacks are the one trade still holding up, though he questions whether they make economic sense.

  Clark says buybacks in the US are the one trade that still holds up, even though he is not sure they make sense. He also sees companies warning about cost pressures and the market reacting by punishing shares, because those warnings imply less cash for buybacks. The support for US equities therefore rests largely on buybacks, which he regards as questionable.

  “The only thing that seems to be working is buybacks. And again, I question whether they make sense, but buybacks in The US seem to be the one trade that still holds up.”at 40:31

International equities

• Russell Clark · Capital Flows and Asset Markets (Substack)

  Japan is the most attractive market because it lacks political pressure to restrain house prices or fight inflation.

  In most markets rising property prices are a political negative, because younger voters, who now hold political weight as boomers fade, want cheaper housing or higher wages. Japan does not show that pressure. After decades of stagnation, Japanese voters are comfortable with rising house prices and are less worried about inflation. Clark sees this as making Japan a more attractive market, though he calls it a personal view rather than a firm recommendation.

  “One market I do like, for whatever it's worth, is Japan, purely from the point of view that I think in most markets, increasing property prices is seen as sort of as a political negative”at 42:15

Japan's economy

• Russell Clark · Capital Flows and Asset Markets (Substack)

  Japan's 1990s stagnation was US-driven, and renewed investment and fab-building now point to better growth.

  Clark argues Japan's stagflation followed US pressure in the 1980s. Japan led in DRAM and NAND, and the 1986 semiconductor agreement capped Japanese market share. The best Japanese companies then stopped investing at home and moved capital to Korea, Taiwan and China, which produced stagnation. Now that China is seen as the enemy, investing in Japan is acceptable again, and Japan looks like a strong place to build new fabs. He thinks the current inflation and growth are also politically driven, so he expects growth to be good.

  “So the whole Japan stagflation era, which sort of began in the '90s, was the byproduct of the Americans not being happy with Japan doing so well in the '80s.”at 43:45

Middle East & Iran

• Russell Clark · Capital Flows and Asset Markets (Substack)

  The Iran war was poorly executed, showing US weakness, and continuing it made little sense.

  Clark initially supported the war because Iran is a bad regime, and he expected a better-executed campaign given the decades of planning. He says it turned into a dogfight that was hard to understand, while oil and diesel prices rose. He thinks the continuation made little sense and that the poor execution has made the US look weak. He thinks this probably cost the Trump administration votes.

  “But this one's been very poor and really sort of showing The US to be weak, which is always a bad look.”at 51:37

Natural gas & LNG

• Russell Clark · Capital Flows and Asset Markets (Substack)

  Falling long-dated US gas futures signal that the market expects shale-driven oversupply in a few years.

  US shale rigs and production are rising again, partly because of the Iran war's effect on oil. Long-dated US natural gas futures are falling. Clark reads this as the market expecting new production to create an oversupply in a few years, because the US shale industry is gearing up.

  “if you look at long dated natural gas futures in The States, they're actually declining. So the market is seeing the production coming on and saying, actually, we're going have an oversupply again in a few years' time.”at 53:11

Oil & fuels

• Russell Clark · Capital Flows and Asset Markets (Substack)

  Oil is structurally a dead commodity: road-fuel demand has peaked, so energy use and GDP are decoupling.

  Road oil demand has already topped out in China at a far lower level than in Europe or the US, and it is now topping out and starting to decline in those markets too. Clark says the link between energy consumption and GDP is breaking. That is why no new US refineries are being built: long-term return on equity looks poor because the end value is low and the demand is not there. He allows that the price can still rise and that oil moves markets in the short term, but he treats oil as a distraction for inflation and a weak long-term asset.

  “when I look at oil, I sort of feel like we're sort of looking at like a dead commodity in a way.”at 53:46