Markets diverge as France rattles Europe and AI rebounds

Good morning and welcome to Julius Baer's Moving Markets Podcast. It's Tuesday, 20 October, my name is Bernadette Anderko and I'm delighted to be joined by two colleagues this morning to bring you the latest developments in financial markets. After the recent rout in artificial intelligence stocks, I'm thrilled that Enrico Canelo from our next generation research team is going to be joining me to discuss what was behind it and the implications for investors now, but first, as ever, I'm looking forward to catching up on the hot topics of the last 24 hours. This morning brought to us by my colleague, Lucia Chucholovic. Good morning to you, Lucia. Good morning, Bernadette. Well, we've got a lot to talk about yesterday was a day of two halves, US tech stocks bounce back sharply. Well, European markets sank under the weight of the violent unrest in France and meanwhile, oil prices whipsawed on geopolitical tensions. Why don't we start with the US rebound? What drove the bounce there? Yeah, I was a classic oversold rally. The nested composite led Wall Street higher, rising more than 1%. The S&P 500 gained around half a percent while the DA was little changed. Technology and consumer cyclical stocks led the way. Investors stepped back into the market of the last week's sharp sell-off, which was fueled by concerns that higher interest rates could slow AI spending. But as one analyst put it, rubber bands can only stretch so far. Now, the debate isn't over yet. Five major tech players are expected to increase combined capital spending to roughly 1.2 trillion US dollars next year. And with meta and Tesla reporting tomorrow and SpaceX today, investors are still digesting the AI-CAPEX question. France in turmoil, perhaps you could paint the picture for us later. Yeah, it was a stark contrast. The stock 600 fell nearly 1%. France's capital dropped over 1%. It's worse than nearly two weeks as anti-government protests escalated into riots over the weekend. This unrest is seen as boosting the national rally leaders' chances in next year's presidential election, which then raises questions about France's EU membership. But Bernadette Spain also rattled investors. The I-Bex fell 1.5% after a SNAP election was called. Now, overall banks were the biggest leggers in Europe while tech was the only sector in positive territory. You're shifted towards non-Eurozone markets like the UK, Switzerland and Scandinavia. And by the way, French equities are heading for a 16th straight monthly outflow. Yeah, and that political turmoil in France clearly report into bond markets. How did fixed income react? Yeah, the French bond market came under pressure with 10-year yields rising and the spread over German bonds widening to around 147 basis points, which is a level not seen since the European debt crisis. Now, US Treasury yields also moved higher as the rebound in AI-related stocks reduced demand for safe haven assets. Meanwhile, the Fed is in its blackout period ahead of its October meeting. Markets show over 80% odds of no change this month, but 91% odds of a hike in December. Yeah, we'll have to see how that plays out. But why don't we turn to commodities and our loot sphere? Oil prices had a world session, didn't they? What happened there? Yeah, that's right. Oil had a volatile session yesterday, Brent briefly slumped by more than three US dollars a barrel before ending little change, while WTI rose more than 2% and is currently trading just below 104 US dollars. The moves were driven by geopolitical tensions. Once again, after the Houthis claimed a tax on Saudi energy, airport and military targets, including an Aramco facility, now at the same time, crude flows through the strait of Homoos fell to roughly 8.5 million barrels per day last week, which is down from over 20 million the prior week. All right, what about the broader commodity space and crypto? Gold slipped around half a percent yesterday, pressured by the firm $1 and elevated yields. Copper fell on rising inventories and the end of a mine strike in Chile. Silver, platinum and palladium all gained and Bitcoin bounced over 2% back above $81,000 US dollars, lifting crypto-related stocks and iron ore also rose on China's GDP beat. And how did currency markets handle all of this? Yeah, the euro was dragged lower by the France crisis and higher oil prices falling a quarter per cent against the dollar, sterling also slipped against the dollar, but gained against the euro, the Australian dollar rallied and the Japanese yen is also in focus after the Japanese central bank governor said underlying inflation is quite close to 2%, signaling scope for further rate increases. All right, why don't we take a look at what happened in Asia last night? We obviously saw a quite cautious session shaped by three dominant forces, this unresolved French bond crisis caution ahead of these heavyweight AI earnings and will continue to fall out from Indonesia's MSCI crisis, but how did markets perform and was there any new story that stood out for you? Well Bernadette, I see a sea of red, Japanese shares are around 1.5% lower, mainland China shares are down about half a percent, Hong Kong is over 1% lower and India is in the red too, the clear and the performer is the cost be however down over 4%. And one new story that really stood out was in South Korea shares in AI services firm Ellis Group plummeted 30% in their market debut today, the company provides AI cloud services as well as AI based education and business software. Wow, okay, let's look at the day ahead then, what should investors be watching out for? So we already received German producer price index full September which came in higher than expected and we also got UK labour market data which surprised positively with the unemployment rate falling. Now later this morning we get the Germans at the W survey expectations for October and in the US this afternoon we have real estate figures including housing starts and building starts. And of course the earnings season continues, today's focus will be on Netflix and SpaceX. Now finally looking at the futures board, US stock futures are currently in the red, so let's see where we end up today. Indeed, let's see, thank you very much for bringing us all at all of the news stories this morning, Lucia, great round up. Thanks for having me. Now let's take a little closer look at artificial intelligence, good morning to you Enrico. Morning Bernadette. Well, what can I say? Last week was a rough one for AI leverage stocks. What would you say is behind the latest selloff and should we be worrying about market sentiment now? Well, in the last week was pretty well for AI stocks, but beyond the stocks moves the context matters in my view and the immediate trigger was not weaker demand for AI solutions but rather concerns about certain parts of the infrastructure build out are being financed. And the pressure started with the Neo Clouds, which is a subset of data center developers that borrow heavily to build new projects and then rent that computing capacity to AI companies to run their models. And as borrowing costs went up and on the back of management commentary that did not provide the clarity the market was looking for, investors began questioning how sustainable these business models are and whether future revenues will arrive quickly enough to cover that debt. We saw the selloff speeding over to certain parts of the AI value chain, though in a more limited manner. So I would describe the selloff as a repricing of leverage and reference in risk rather than evidence that the AI technology or demand cycle is suddenly broken. So one big Neo Cloud core we've dropped 25% last week after its corporate bond sold off signaling investor concerns about the company's mounting debt and future borrowing costs. Could this deterioration of sentiment around Neo Clouds spread across the AI ecosystem? So, in short, we do not consider that the most likely scenario and the key distinction here is between the more leverage Neo Clouds and the large hyperscaters. So for context, some Neo Clouds have net depth of roughly two to three times EBITA, relatively concentrated business models and a continuing need for external financing to sustain the business model. And this in our view makes them particularly sensitive to higher borrowing costs or weaker refinancing conditions. The hyperscaters on the flip side are in a different positions. They add diversified revenue streams, strong operating cash flow and considerably more balanced sheet flexibility. These should be able to contain their direct speed over effect. And if we take a step back here, the broader lesson is that strong AI demand does not mean every financing model will eventually succeed. The upcoming reporting cycle will tell us more about plans on leverage, capital intensity and related execution strategies. So, there's more to come on that. We had Mensur on the show yesterday saying that from a technical perspective, things still look good for the AI space. So, I'm wondering what you and the next generation research team think and whether you remain constructive on AI after the sell-off? Yes, absolutely. And that's mainly because the underlying indicators continue to point to strong demand. Demand for AI computing capacity remains very high. Processing. or rental rates or firm, and adoption and investment plans to continue to point up. This is also not the first period of volatility in the broader AI cycle. DeepSick, tariff, concerns, and recurring questions about capital expenditure have all cost temporary sell-offs, but infrastructure build-out and adoption cycle continue and eventually became stronger. What would make us more cautious would be a broad and sustainable decline in utilization rates, process surprising and capital spending intentions, and we are not seeing that combination yet. For now, credit concerns in our view are a reason to be more selective. They are not by themselves evidence that the AI demand cycle has ended. That's clear, but to add to the uncertainty and through pick reported exceptional revenue growth, there is some nuance behind the headline numbers, and does that answer the concerns about artificial intelligence economics? Well, it's certainly a never-boring, and the headline figures are really impressive. An anthropic reported legend generated almost 17 billion dollars of revenue in the third quarter this year, which is an increase of about 45% from the previous quarter, and indicated that it was profitable on an adjusted basis. The debate now is shifting beyond these headline figures from the amount of growth to the quality of that growth. The implied annualized revenue round rate is approximately 67 billion dollars, only slightly above the 65 billion dollar round rate reported at the end of July. Growth therefore remains exceptionally strong, but it may no longer be accelerating at the same pace. The figures are also not audited, and investors still need more clarity on the profitability adjustments and related mechanics, when it comes to revenue and underlying compute cost, given anthropics and normals long-term cloud commitments, the critical question is whether durable margin can grow quickly enough to support these commitments going forward. I can't let you go that quick question on cyber security. This has been top of mind in the artificial intelligence age, and there were reports of a recent attacks to these French supermarket chains. What can investors learn from these events? So this incident certainly demonstrates that cyber risk is no longer confined to data computers or payment systems, and digital attacks can now disrupt physical operations. Just for context for these attacks on the French supermarket chains, the attackers reportedly gain access to connected refrigeration controllers, force prolonged defrosting cycles, and then destroy stock across hundreds of offstores. In this case, it illustrates how every connected operational device, including a sensor or a refrigeration controller, has effectively become a cyber security endpoint, and the difficulty here is that operational technology often runs on legacy hardware, receives infrequent security updates, and is not visible to conventional IT security tools. As AI makes it cheaper to identify vulnerable devices and then scale up attacks, automated cyber defense becomes increasingly important, and this supports the structured demand outlook for the broader cyber security theme, even evaluations in parts of the sector remain elevated. Super, you've given us a lot to think about today, Enrique, but thank you so much for joining the podcast. Well, that's it for today's podcast. Thank you all for listening. I hope you enjoyed the show today, and of course, please do tune in again tomorrow when Helen Freer will be back with more of our experts to bring you up to speed on what's moving markets. So don't miss that. Meanwhile, good luck today, and goodbye for now. The information and opinions expressed in this podcast constitute marketing material, and are not the result of independent financial or investment research. Please refer to www.juleasbear.com/legal/podcasts for further other important legal information. [BLANK_AUDIO]