Goldman Sachs trading head interprets institutions' significant AI position reduction: power gap and cost of capital in focus

On October 22, Rich Privorotsky, Goldman Sachs' head of European fixed income, currencies and commodities (FICC) and equities trading, gave an interview to the podcast program "The Markets" in Goldman Sachs' trading room in London. Privorotsky disclosed a highly striking figure: "Last week, on our prime brokerage book, we saw the largest technology stock sell-off in history."

Facing this wave of selling, the ongoing third-quarter earnings season appears particularly crucial. The market's focus has shifted from simply "telling stories" to strict performance delivery. Privorotsky noted that the market expects AI capital expenditure (Capex) to reach as high as $1.1 trillion next year, but that is no longer enough to get investors to pay.

"It is not enough to simply say you are spending a lot of money; you also need to prove that outstanding orders are converting into substantive revenue," Privorotsky stressed. "They must show the returns on that spending while pushing capital expenditure. This is not an easy situation." He said that cloud business growth, monetization capability and the speed of achieving free cash flow breakeven will have a major impact on the currently extremely sensitive credit market.

Infrastructure and funding dual constraints: a "40-50 gigawatt" power gap next year

The more realistic physical bottleneck lies in power delivery capacity.

He presented a set of key figures: "Next year's power gap is about 40 to 50 gigawatts. If you cannot obtain the energy that actually powers data centers, can you still consume the chips you have already ordered in advance?" These concerns have pushed investors into "risk-off mode," reducing growth expectations by shorting AI and AI capital expenditure trades, while capital floods into the bond market for safety.

Macro paradigm reshaping: from "savings glut" to "spending glut"

From the perspective of the macro cycle, Privorotsky painted a picture of a profound paradigm shift.

He said that his entire career had been spent in a "savings glut" environment—there was so much money in the system that people did not know what to do with it. But now, a "seismic shift" has occurred.

"Over the past year, we expect to spend more than $1 trillion on AI capital expenditure, while most governments around the world are running fiscal deficits. We are in an era of 'spending glut.'"

This shift has pushed up the cost of capital and naturally crowded out marginal, lower-quality borrowers.

He observed that this pressure is reflected not only in the sharp widening of CCC-rated credit spreads, but has even spread to the sovereign debt sphere (such as French government bonds facing large fiscal deficits). Combined with persistently high energy prices, overall market risk appetite is being significantly weakened.

Amid the complex interplay of macro and micro factors, investor anxiety has shown a peculiar "polarization" phenomenon.

Privorotsky shared what he encountered in conversations with investors this week: "I met some investors who said, 'I am worried we cannot deliver computing power, and that scares me'; then I met another group of investors who said, 'I am worried we will over-deliver computing power, and that scares me.' Somehow, two completely mutually exclusive outcomes together make up the sum of market concerns."

Because energy prices have a major impact on the economy (although the S&P 500 has performed steadily, the median stock has actually fallen more than 10%), the financial conditions index (FCI) has tightened sharply, forcing the Federal Reserve to do more.

Privorotsky once firmly believed that "As goes oil, so goes everything else," but he now admits that in the current market context, the logic has evolved into "As goes AI, so goes everything else."

Despite short-term pressure, Privorotsky is not entirely pessimistic. He believes that spot market computing power prices are hitting new highs every day, and demand remains "greedy and endless." On trading strategy, he said that although conviction is currently not high, assets related to "electrification" that have suffered major downgrades, civil aerospace assets and some transportation sectors now have considerable margins of safety and deserve cautious optimism.

Below is the full translation of the podcast conversation:

Chris Hussey 00:00

This is "The Markets." I am Chris Hussey, and today is a Thursday in October. I am joined from the Goldman Sachs trading floor in London by Rich Probarowski, who is the head of European equity trading for FIC and equities under Prevo. Thank you very much for joining. Thank you for having me. Hussey, let's get straight to the point, because, you know, I have read everything you wrote, and you made a good point—earlier this summer, that pain trade, the trade everyone had to chase, was long AI and short everything else. It has somewhat become a very big trade. Now as we enter the third-quarter earnings season, this trade seems to be falling apart. What exactly happened? How do you think it will evolve next? There has been extremely severe concentration in the equity market, and I have to say, the past few weeks have been especially painful for this kind of consensus trade.

Rich Privorotsky 00:48

Positioning is highly concentrated in AI, semiconductors and hyperscalers. I would say that right now the market is probably dealing with three different issues. First, obviously, we have spillover effects in the cost of capital. I think this is almost everywhere, whether it is CCC spreads, European sovereign spreads, or the more leveraged, most cutting-edge parts of the AI trade—among them Neo cloud (Neo clouds) are especially prominent. So, this becomes a supply problem. You know, will these companies find it difficult to find financing to deliver all this AI spending? The second point is also about supply: do we really have the ability to power up and deliver gigawatt-scale electricity, the market has questions about this, the range for next year is roughly 20 to 40, 50 gigawatts. So, if you can't get the electricity to actually power the data centers, can you really, can you really absorb those chips you've already ordered in advance, so.

Rich Privorotsky 01:38

These are all concerns hanging over the market, and all of this combined has pushed everyone into a risk-off mindset; the concrete manifestation of this mindset is that growth expectations have been pushed down through the AI trade and the AI capex trade. And on the other side, as growth expectations were revised down, bonds have been rallying—this was completely not in anyone's script.

Rich Privorotsky 01:56

Okay, there's a lot of information here. But let's focus on the near future, which is actually the present—the Q3 earnings season. We're already in it, some large tech stocks have reported, but there are many more to come. What are you watching most closely during Q3 earnings season? Yes, I think we're about to enter the core part of hyperscaler earnings, and I think what really matters is not just seeing capex stay at a high level—the market broadly expects $1.1 trillion next year—but most importantly, seeing, you know, continued growth and monetization of the cloud business. So, just saying you're spending a lot of money is not enough, you also need to prove that backlog is meaningfully converting into revenue. So I think this will be very critical for the market. I think how quickly they achieve the transition to free cash flow breakeven—relative to market expectations. It could have a major impact on the credit market, and the credit market has become extremely sensitive. So for me, really, there's a lot to look at. I would say, from other areas of the economy. But I think the market's focus will really land on the hyperscalers, and then on the quality of delivery, not just the spending. It's too hard for these companies to get over this hurdle. They have to keep pushing capex higher while also showing you the returns on that capex. This is not an easy situation, is it? Yes, it's not. But look, I mean, the current backdrop is this: you see compute prices hitting new highs every day in the spot market. Demand for what they sell is insatiable, so I would say.

Chris Hussey 03:23

Looking at it a bit more optimistically, they should have a pretty good story to tell. Okay, Rich, it's been a while since we've had you on the show. Your last appearance was in January. But throughout this year, you've been one of the people holding to that mantra—however oil goes, everything else goes. Is that still the mantra? Yes, I think we've run into an extremely tricky backdrop in energy. And, you know, I think.

Rich Privorotsky 03:46

The market, the stock market, I think, has done its best to ignore and isolate the rise in energy prices in a corner. But that's actually not quite fair, because the S&P index has done a good job of isolating it, but the median stock has actually fallen well over 10%. So this has had an impact on rates. It has also affected FCI. In fact we went through a massive FCI tightening, which in a sense also forced the Fed to do more. And we repriced an entire rate-hike cycle. So I think energy has accounted for a large part of this. I think we are still carrying a lot of risk premium because of energy. Obviously, how these things evolve over the next few weeks will be very critical for risk assets overall. And my optimistic view is that we will get some form of resolution within the next few weeks. But I have to admit, I thought this should have happened two months ago, and I was a bit off on that. I think we all were. Of course, when you say FCI, you mean the financial conditions index, which is.

Chris Hussey 04:35

Sort of a measure of whether the economy, or all economic impulses, are tightening or loosening. It's not just about rates. When you look at an FCI, sometimes it might be oil, sometimes it might be stocks. Okay, that's a good point. On the oil front.

Chris Hussey 04:48

But related to that oil trade, there's also something you said at the start of the year, which was: I want to be long nominal assets—in an expansionary world where GDP growth is strong and still looks very strong, and inflation is high, I just want to be long this kind of thing, meaning a lot of commodities. How do you see this nominal asset trade today? Is it still a place worth allocating to now?

Rich Privorotsky 05:11

Rich Privorotsky 05:19

Because the reality is, when we were repricing it, we repriced everything, including commodities, including nominal assets. My instinct is that we've somewhat thrown the baby out with the bathwater, and we've clearly started worrying about every dimension of it. To be specific: this week I met some investors who said 'I'm worried compute can't be delivered, and that scares me.' Then I met another group of investors who said 'I'm worried we'll over-deliver compute, and that scares me.' Somehow, we've got two mutually exclusive outcomes, and at the same time they together make up the sum of the market's worries.

Rich Privorotsky 05:50

So I think, from the big picture, nominal assets are still where you want to be, unless we get into an anti-inflationary bust, where capex slows, everyone has to retrench, and we get a real credit cycle. Three months ago I would have said that risk was low; now it's certainly much higher. But my instinct is that over the medium to long term, that's still where you want to be. Though I think adding a bit more of a barbell strategy to the portfolio now might be wiser. It sounds like you're saying that before, maybe oil went one way and everything else followed, and now it's — AI goes one way.

Chris Hussey 06:22

And everything else follows. Is that right? I think, at the risk of oversimplifying, that's probably right. Yes, Rich, thank you for being willing to oversimplify. I'm Mr. Oversimplification myself, so I'm glad you could get there. Of course, part of this will be a big credit cycle, and three months ago you weren't worried about a credit cycle, and now you're a bit more worried. Talk to us about that. Are you seeing cracks in credit? Are you referring to what's happening in European sovereign debt, or more to the credit issues we're seeing on the corporate side?

Rich Privorotsky 06:51

My entire career, you could say, has been in a — savings glut. You know, there's so much money in the system we don't know what to do with it. And just over the past year, we're projected to spend over $1 trillion on AI capex.

Rich Privorotsky 07:05

Most governments around the world are running fiscal deficits. We're now in a spending glut. In terms of how things work and what the cost of capital will become, this is a very big structural shift. So I think it naturally squeezes out those borrowers on the edge, the ones who aren't the highest quality. I think you can already see that in CCC spreads — though to be fair, given the composition of what's happened there, it's somewhat idiosyncratic. But they've moved a lot, and it's already spilled over a bit into the broader credit market. And then obviously you have the French sovereign debt concerns, namely: as a sovereign carrying a huge fiscal deficit in this environment, they too will all get squeezed out. So I think all of that is having an effect, and at the same time, going back to the oil trade — as long as energy prices stay high, it will weigh on risk appetite. And I think for this kind of asset, it's now a bit of a buyer's market. Yes, from savings glut to spending glut, that's a brilliant distillation. That's so right. And that savings glut was only five years ago. The speed at which we've shifted is remarkable.

Chris Hussey 08:04

Okay. A lot has happened. You've just told us all about it. Let's wrap up. How do you do the trade now?

Rich Privorotsky 08:10

The picture is a bit messy, and I have to say confidence isn't especially high, but I think there are some parts of the electrification space that we think look interesting. Their valuations have come down a lot, and I still think A27 has plenty of visibility, which gives you a big margin of safety. I think a lot of aerospace assets have pulled back, especially on the civil aviation side, which is also very attractive to us. And then even parts of the transport sector, frankly, if the energy problem could somehow be resolved, given how low net positioning has gotten and how much selling we've seen over the past few weeks, there's a lot of, I'd say, convexity in that trade. For example, last week the tech selling in our prime brokerage business was the largest ever. So I think there's plenty of reason to be concerned, and I don't want to downplay these risks too much, but I still lean toward being cautiously optimistic at this point. Okay, Rich, let's fast-forward to next week. What are you watching next week that could affect the market? I think the key is those numbers, the hyperscaler numbers we discussed. I think, you know, let's see the monetization, let's see the capex numbers. I think that will be very critical for the tape. And I think, you know, how this progresses will determine how people think about credit, how they think about potential demand, and the other data points we get from the various labs and suppliers. Rich, always a wealth of information. Thank you very much for joining us today. Thank you. That.

Chris Hussey 09:30

That's it. That's the end of this episode of The Markets. I'm Chris Hussey. Thanks for listening. Risk Disclosure and Disclaimer

Markets carry risk; invest with caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their particular circumstances. Invest at your own risk.