Monday Morning Links

• The 2007-8 Everything Bubble and the 2021 Duration Bubble, for instance, were both bubbles in which the right portfolio to own if you believed there was a bubble was a portfolio that would have an unacceptably low expected return if markets were fairly priced. But the 2025 AI Bubble looks little like either of those two and much more like the 2000 Internet Bubble, in which a bubble-agnostic investor could have owned a portfolio with a reasonable risk/reward trade-off in either a bubble or a business-as-usual scenario. Today, non-U.S. equities, deep value stocks, and liquid alternatives offer returns that look reasonable or better, regardless of whether AI is in a bubble. Tilting a portfolio away from AI names and toward those assets may save investors a lot of pain if it turns out we are in a bubble without meaningfully reducing expected returns if financial markets are somehow still fairly priced today. [GMO]
• There's basically always been a cottage industry of value investors complaining that the market has gone crazy, but that's a selection effect. People like reading about the market when it's up, or when it's crashing, but not when it's had a long period of underperformance or has been grinding down month after month. But value tends to underperform during bull markets, and also tends to crash alongside everything else in bear markets. And people talk a lot about stocks during bull markets and crashes, and a lot less in year three of a bear market or after a long, choppy period of sideways trading. So you should always expect value investors, of the discretionary or systematic variety, to be complaining about things. They do best when nobody's talking about stocks. And that's probably part of the risk premium that value has earned over long periods: to be a value investor is to underperform when stocks are all anyone is talking about, and to make your money when people don’t care about it all that much. [The Diff]
• For the last two decades, datacenter construction basically co-opted the power infrastructure left over from US deindustrialization. For AI CapEx to continue growing on its current trajectory, everyone upstream in the supply chain (from people making copper wire to turbines to transformers and switchgear) will need to expand production capacity. The key issue is that these companies have 10-30 year depreciation cycles for their factories (compare that to 3 years for chips). Given their usual low margins, they need steady profits for decades, and they’ve been burned by bubbles before. If there’s a financial overhang not just for fabs, but also for other datacenter components, could hyperscalers simply pay higher margins to accelerate capacity expansion? Especially given that chips are an overwhelming 60+% of the cost of a data center. We did some back-of-the-envelope math on gas turbine manufacturers which seems to indicate that hyperscalers could pay to have their capacity expanded for a relatively small share of total datacenter cost. As Tyler Cowen says, do not underrate the elasticity of supply. [Dwarkesh Patel] 
• OEM Bergen Engines has secured a contract to supply a 400MW power plant for a new AI data center on the US East Coast. Operating fully off grid in true islanded mode, it will be the first medium-speed reciprocating engine power plant in North America designed for AI-driven workloads. [Bergen Engines]
• When I first read about the discovery of a vast new deposit of lithium in a volcanic crater along the Nevada-Oregon border, I can’t say that I was surprised. Not because I know anything about geology — but because, as an economist, I am a strong believer in the concept of elasticity of supply. Now about elasticity of supply, in which we economists tend to have more faith than do most people. Time and again over the centuries, economists have observed that resource shortages are often remedied by discovery, innovation and conservation — all induced by market prices. To put it simply: If a resource is scarce, and there is upward pressure on its price, new supplies will usually be found. Not surprisingly, the Lithium Americas Corporation put in a lot of the work behind the discovery. Searching for new lithium deposits has been on the rise worldwide, as large parts of the world remain understudied and, for the purposes of lithium, undersampled. Just as Adam Smith’s invisible hand metaphor would lead one to expect, that set off many new lithium-hunting investigations. Sometimes the new supplies will be for lithium substitutes rather than for lithium itself. In the case of batteries, relevant potential substitutes include aqueous magnesium batteries, solid-state batteries, sodium-based batteries, sodium antimony telluride intermetallic anodes, sodium-sulfur batteries, seawater batteries, graphene batteries, and manganese hydrogen batteries. I’m not passing judgment on any of these particular approaches — I am just noting that there are many possible margins for innovation to succeed. [Tyler Cowen]
• Think about the last time you read something important to you—maybe it motivated you to do something differently, or changed the way you thought about something. Did you write to the author and let them know? Personally, I have literally never done this, as far as I can remember. (Huh, maybe I should…) Similarly, you should expect that most people who love your writing aren’t going to tell you that directly. So: lower your bar for what’s worth writing about! My personal standard is anything that I’ve said more than once in a conversation. [Ben Kuhn]
• Canadian sands, some of the world’s largest oil reserves, are also steadily contributing to higher Canadian oil output, now at 5.4mbpd! Canada is a sleeping oil giant and its capable operators figured out how to produce oil at low breakevens over the years. PS: if u see this as UAE oil minister, what do you do? Yes, you panic and order a reserve audit among OPEC members to get a higher quota as you know everyone’s oil reserves (which determine output quotas) such as Kuwait are overstated, as typical for ME low trust societies. My point? UAE had it with cheaters. They too want to produce more oil or leave OPEC sooner rather than later. Other nations that are set to produce more oil regardless of oil price volatility is Guayana, Argentina, Kazakhstan, potentially also Venezuela down the road. The latter possesses the largest oil reserves globally. [Alexander Stahel]
• China’s clean energy efforts contrast with the ambitions of the United States under the Trump administration, which is using its diplomatic and economic muscle to pressure other countries to buy more American gas, oil and coal. China is investing in cheaper solar and wind technology, along with batteries and electric vehicles, with the aim of becoming the world’s supplier of renewable energy and the products that rely on it. The main group of solar farms, known as the Talatan Solar Park, dwarfs every other cluster of solar farms in the world. It covers 162 square miles in Gonghe County, an alpine desert in sparsely inhabited Qinghai, a province in western China. [The New York Times]