Peaceful Uneasy Feeling

Any long-term sports fan knows the moment in their own way. That moment when you start to doubt a winning outcome. It’s not over yet. The score is still close. Maybe your team is ahead, but the game on the field does not look right. Strange things happen in games, but often you can see (even when the score is close) that one team has some plays that the other team isn’t stopping. Time to burn clock, if this can be managed… though sometimes going “conservative” does not work either.

Much like The Eagles’ classic “Peaceful Easy Feeling,” markets today seem calm on the surface. Yet beneath that laid-back rhythm lies an undercurrent of tension, a melody that sounds familiar, but feels off-key. Investors may be humming along, but the tune could change without warning.

For readers who do not have their noses up against the glass peering directly into financial markets at close range, this is my state of mind in mid-October 2025. The exuberant expectations baked into stock prices are obvious. The aggressive spending on a ‘future so bright you need sunglasses’ thesis is evident. It’s well beyond the AI eats the world and related data and infrastructure buildout themes. The whole construct of today’s momentum-y market leadership feels vulnerable. Like a sports team clinging to a lead that it is bound to cough up.

That’s one loose analogy for how markets feel at the moment. Thematic categories replete with vague, open-ended thinking – some positive, some negative – have been levitating since the Spring (really for the last 2+ years). Gold. Crypto. AI stocks. Uranium. Power utilities intending to serve the AI power demand. Sports team valuations. It’s not nearly so extreme outside the U.S., where open-ended optimism of military spending in Europe has popped valuations for a very small group of European defense names, for example. China has a handful of legitimate AI plays that are up this year, but their stock movements are barely a wiggle over a longer arc. It’s mainly in America, and it’s big.

Coming into 2025, a number of investment themes stretching back into the past have come unglued. The democratization of luxury. Consumerism in general. Globalization of manufacturing and trade. Software eats the world. Maybe the theme seekers are seeking themes that work and aren’t finding many?

Markets are fueled by optimism. Vague and succinct forms of pessimism tend not to be well-rewarded except for short-lived downdrafts. Such periods are great if you have unusual timing, but these episodes tend to end quickly (think “vaccine Monday” 5 years ago). There were moments of abject pessimism in early 2020, imagining an ongoing health crisis and closed public spaces. But people were never going to live this way. There were moments of similar pessimism in 2022, as an inflation dragon not heard from in decades roared to life with an uncertain impact and lifespan. While inflation has not fully retreated to a sub-2% trend, Central Banks have acted, and the inflation problem has become far less meaningful.

The uneasy feeling is that none of this feels intellectually coherent. Sure, AI-generated landscapes, images, and searches are pretty cool. It looks like this will render various business processes much more efficient, from coding to credit scoring to sales pipeline management to visual design. Maybe entry-level investment bankers become AI coders. It’s a breakthrough technology, no pushback there. But where are the apps that are powerful enough to generate $trillions in revenue? Can the cost of power, land, and leading-edge silicon possibly be offset by paid revenue? At the heart of recent explosive moves in AI-related stocks is OpenAI, which has inked deals with Oracle, Broadcom, and AMD for 8, 10, and 6 gigawatts (GW) of server capacity, respectively, in the past few months. According to the Wall Street Journal, OpenAI believes it will need about 100 GW of power and the corresponding silicon to support its various projects. The buildout cost for one (1) GW of power, servers, data centers, connection costs, etc., is about $50 billion. So that’s $1.2 trillion of capital costs agreed to by OpenAI, just to get rolling. Ditto similar figures for Anthropic, Google, Meta, and others built from the ground AI capabilities? The servers inside will depreciate very quickly. The energy equivalent (in natural gas) of 100 GW is about 740 bn cubic feet of gas annually. It’s a high daily rate (of sustained energy and capital costs) to run.  OpenAI is at a current run-rate of about $14 billion in revenue and loses money. It will surely grow, but profits need to materialize considerably, and it’s extremely hard to cook up a revenue number where spending of this magnitude could be financed through much better profits, given the capex requirements, physical constraints (on power systems), and competitive intensity.  Sure – even if these are all overtly over-the-top spending dreams versus reality – we are going to need something truly never seen before in terms of apps and associated value.

There are some old patterns echoing the telecom bubble of the 1990s, and any number of “freemium” software products that are hard to shake. Broadband was correctly understood in the mid-1990s as an essential infrastructure foundation that all businesses and homes would need. It just took a long time to fill the fiber-optic pipes with traffic, and in the interim, pricing assumptions baked into the buildout proved outlandish. The average family and small to medium business were not going to shell out big bucks for these things. Freemium software does not have a great history of converting to a high percentage of paying users. By some WSJ estimates, 97% of current AI users are using the free version. Even if the paid rate quintuples, that’s 85% of the population not paying anything. The most familiar pattern of past infrastructure booms relates to where in the economic value chain value tends to be trapped by new technologies such as this. It tends to be downstream – from the days of Standard Oil to the hotels and manufacturers positioned along railroad lines to non-linear TV content companies such as Netflix today with respect to broadband.

Well – enough on that. The commentary is getting louder that the numbers don’t really add up here. It’s just a very large amount of global market cap caught up in this narrative. It’s an uneasy feeling, watching this continue to swell.

Over in a different corner of the market, gold, crypto, collectibles, thematic non-earners, and meme stocks are seeing their own parabolic moves. There will always be a speculative fringe of stocks; that’s not new. The long-short hedge fund crowd can be rolled by a propensity to overconcentrate in thematic shorts, leading to the first rise of meme stocks in 2021. Here we go again. There is no coherent explanation to look for here, it would seem…

Let’s now focus on gold because this particular parabolic move has been so odd. Personally, I’ve never been a “gold bug” or someone who takes the underlying sentiment (such as fear of imminent hyperinflation) seriously. It’s a non-productive asset with no industrial use case. The main attributes are that it has a distinctive look as a metal and is rare. Would one try to transact at Walmart in a hyperinflationary episode by shaving flakes off of a gold bar in a checkout line? It’s very hard to picture that. In a world dominated by e-commerce, it’s even harder to figure out the quasi-monetary use case.

Gold tends to be the “antidollar”; i.e., it tends to appreciate when the dollar is depreciating and/or when dollar-denominated business assets (like stocks) are not working. Think 2008. The best rational explanation for gold price action is “relative seigniorage”. When real interest rates are high, dollar savings earn a good return, and gold (which pays no interest) must deteriorate in value to offset the real value it is competing with in the form of interest payments. Kind of like a zero-coupon bond. Conversely, when negative real interest rates emerge (through inflation or loose central bank policies), the same zero-coupon bond features become relatively more valuable, and gold appreciates.

None of these normal drivers of value (sustained dollar weakness, negative real interest rates/low seigniorage, or poorly performing financial assets) are at work in 2025. The DXY index (a basket of currencies against which the dollar trades) is roughly at its average since 2014. Inflation could be incrementally better, but few inbound signals suggest a sudden leap upward. Money supply growth at 4.8% is well below its average rate of about 6% in the non-inflationary 2010s, and the Fed Balance sheet as a % of GDP is back to where it was in the 2010s. There’s certainly talk of “currency debasement” given yawning deficits and little political will to address them.

But given the above, this seems like a reflexive memory of the inflation problem in 2021-22. Notionally, if AI is ¼ of what it’s believed to be, it should be a deflator of costs in a variety of fields. Technology has a way of doing that. So, what is the driver here?

Nothing to see here that looks like an imminent debasement of the currency…

We don’t know. Gold has little clear use case apart from being “alternative money”. Interest in alternatives to the dollar has perhaps been heightened by the United States government’s recent tendency to bar other nations from using dollar payment systems for political reasons. Still, it does not really compute apart from being a bit of a mania. A mania that would typically be catalyzed by poor economic and financial conditions, but yet the opposite seems to be true. “Speculation” – That seems to be a more realistic narrative.

Saying “hey, wait a minute” to a thundering herd ignoring practical realities seldom works, but at some point, the herd tends to tire out. And a different set of questions comes to predominate. We are not attempting to be market strategists. We have an investment philosophy and discipline built around businesses that are reasonably explainable and intend to stay in that lane. Just pointing out things that don’t appear to make much sense that have become rather large of late.

Brian M. Barish, CFA

Brian Barish is the President and CIO at Cambiar Investors and is responsible for the oversight of all investment functions…