Oil prices hit five-year lows… don’t expect a sustained turnaround anytime soon… why doesn’t AI energy demand prop up prices?… electricity versus oil… a big AI red flag from Eric Fry VIEW IN BROWSER Yesterday, oil prices slid to their lowest level in nearly five years. Brent Crude fell below $60 a barrel while West Texas Intermediate Crude (WTI) dropped into the mid-$50s. Both are rebounding as I write on Wednesday morning, but the overall trend here is weakness. Brent and WTI have shed 23% and 25%, respectively, over the past six months. Why? Overwhelming supply. Here’s JPMorgan: At the risk of flogging a very dead horse, our message to the market has remained consistent since June 2023. While demand is robust, supply is simply too abundant. The data backs that up. Today, we have record U.S. crude output, sustained production from OPEC+ members, and softer demand signals – notably from China’s slowing economy. Put it all together, and the global markets are drowning in a glut of excess crude. If you’re holding lots of oil and gas stocks, you should know that the big investment banks don’t expect any relief over the next two years. JPMorgan says Brent will fall to $58 next year, while WTI is expected to hit $54 (it fell to $55 yesterday). And looking further out to 2027, JPMorgan doesn’t see a rebound either – just another roughly $1 per barrel of downside, keeping prices depressed for years rather than quarters. Goldman’s estimate is similar – it sees oil hitting $56 and $52, respectively, next year. What about AI’s ravenous demand for energy? In 2023, U.S. data centers consumed roughly 176 terawatt-hours of electricity, about 4.4% of total U.S. electricity use. Projections from Congress.gov show this could double or even triple by the end of the decade as AI workloads expand. So, how can we have an AI-driven energy shortfall while we’re drowning in oil? It sounds contradictory – but it isn’t. Data centers drive demand for electricity, not crude oil. The vast majority of U.S. electricity is generated from natural gas, renewables, nuclear power, and coal. Oil plays a minor role in grid power generation. There simply isn’t a correlation between data center power consumption and oil burned at utility scale. So, when utilities talk about data centers stressing the grid, they’re talking about electric generation capacity, not crude oil consumption per se. This is why agencies like the U.S. Energy Information Administration forecast record highs in electricity demand – partly driven by data centers – even as the oil market is contending with too much crude supply and soft demand growth. | Recommended Link | | | | While CNBC promotes AI stocks, Louis Navellier's grading system shows institutions quietly accumulating energy infrastructure. The same system that wealthy firms paid $24,000/year to have him evaluate stocks with. He called Nvidia at $1 split-adjusted, Google early, Apple before iPhone— all by tracking where institutional money moves BEFORE anyone else sees it. See the grades and where the money is going now by clicking here. |  | | So, what does that mean for investors? It means we should recognize that these two markets are distinct – demand for electricity is not demand for oil. Zeroing in on the AI/electricity part of the equation, let’s turn to our technology expert, Luke Lango of Innovation Investor. He recently laid out a three-layer framework for playing AI-driven electricity demand: Utilities / IPPs Own the sellers of electricity AI will buy for years. Favorites include Constellation Energy (CEG) and Vistra (VST). Nuclear & Uranium Big reactors and SMRs are back. Buy Cameco (CCJ) for uranium; Global X Uranium ETF (URA) for basket exposure; Oklo (OKLO) and NuScale (SMR) as next-gen reactor names; Centrus Energy (LEU) and BWX Technologies (BWXT) as component suppliers. Energy Storage / Backup Data centers can’t go dark. Buy Bloom Energy (BE) for fuel cells; Fluence (FLNC) and Eos Energy (EOSE) for batteries. Storage also accelerates time-to-power: build the battery now, plug into the grid later. Capex cycles end, but if grid spend really doubles into 2030, we’re in the early innings. Over the last few days, Luke has been zeroing in on opportunities related to nuclear power. Specifically, the delay in rolling out new nuclear power plants: Big Tech is going all-in on nuclear. But traditional nuclear plants take 10-plus years to build – and AI can’t wait. Realistically, we need a solution that deploys in two to three years, while also solving the issue of nuclear waste. And I have found the one company that does both. I want to cover more ground in today’s Digest, but you can watch Luke’s free research video on the company he says has “10X potential” right here. Returning to the investment implications for oil… If you had the vision to buy Big Oil at the height of the Covid panic in 2020, congrats, you’re likely up more than 200%. But today’s supply/demand economics should prompt you to ask… Is this still the best place for my money over the next 12-24 months? If you’re patient, don’t mind waiting for the supply/demand balance to shift, and are willing to sit through further potential downside over the next year or two, oil can make sense. Just recognize that prices may fall further than expected. On that note, here’s Yahoo! Finance: At JPMorgan, strategists predict that without some market stabilization efforts, Brent could change hands in the $30s per barrel by 2027, a level not seen since the depths of the 2020 oil crash at the start of the COVID-19 pandemic. Bottom line: Investing in AI/electricity is a bright “green light” today. But for oil and its supply imbalance, let’s create a new investing slogan… Instead of “don’t fight the Fed,” for the foreseeable future, it’s “don’t fight the glut.” Recognize the domino effect of this demand for electricity…. In commodities, too much supply crushes prices. That’s exactly what we’re seeing in oil right now. But today’s dynamic in the electricity market is the mirror image… While crude markets are drowning in excess barrels, the power grid is facing too much demand – and that imbalance is starting to show up in prices in a very real way for consumers. The chart below from JPMorgan tells the story clearly. Over the past decade, electricity inflation largely tracked overall CPI. No drama. No standout pressure. But over the last five years, electricity prices have decisively broken away – rising much faster than headline inflation. And the forecast suggests that gap won’t be closing anytime soon. 
Source: JPAM, RBC So, what changed to make electricity so expensive? You already know the answer… AI and data centers. Hyperscale computing facilities don’t just consume power – they demand it relentlessly, around the clock. And unlike oil, electricity can’t be stockpiled at scale or shipped in from halfway around the globe when shortages appear. When demand surges faster than generation and grid capacity can respond, prices rise. That’s a frustrating dynamic for high-price-weary U.S. consumers. Yes, drivers may finally be catching a break at the pump. But that relief is increasingly being offset by higher electricity costs, which is hitting households through monthly utility bills and feeding into everything from rent to services to local taxes. Bottom line: For millions of households, electricity is becoming the next “can’t-avoid-it” expense. Gasoline down…but electricity up. What will it mean for the American consumer in 2026? Shifting from AI and the financial stress it puts on consumers, to AI and the financial stress it puts on corporate balance sheets… As the AI boom charges ahead, wise investors are starting to ask a critical question… Where, exactly, is the risk hiding? According to Tom Yeung, lead analyst at Fry’s Investment Report, the danger isn’t likely to come from a sudden collapse in AI enthusiasm – but from something far quieter and more familiar… Financial strain masked by accounting and structure. From Tom: The thing that concerns us most is debt and debt-like liabilities… Here, we’re beginning to see some worrying shenanigans that allowed firms like Lehman Brothers and WeWork to temporarily hide long-term liabilities. Tom writes that over the past two years, several major tech companies – including Amazon, Alphabet, and Microsoft – quietly extended the depreciation schedules on their data center equipment. By spreading costs over six years instead of three or four, cloud computing suddenly looks more profitable today – even though the cash outlays haven’t changed. Then we have the more creative financing structures… Meta’s massive $27 billion Hyperion data-center project, for example, was placed into a special-purpose vehicle, with most of the project “sold” to outside investor Blue Owl Capital and leased back for up to 20 years. Back to Tom: This is essentially a loan, because Meta receives cash today in exchange for the promise of future payments. Accounting rules allow the tech firm to classify it as an operating lease that remains off the books. We see similar dynamics with Oracle. The company recently disclosed nearly $100 billion in off-balance-sheet lease commitments, almost triple last year’s figure, with little detail on what those obligations entail. Why does this matter? Because generative AI still isn’t producing meaningful cash flow. In fact, OpenAI now expects to burn more than $100 billion in cash through 2029. As expansion accelerates, more of that funding may come from debt – or debt-like structures that are easy to overlook in good times. Which brings us to a telling signal… Back to Tom: It’s particularly noteworthy that Eric sold Oracle last month for 27% gains – a company he recommended in September 2024 and expected to hold for several years. That’s a big red flag to me… A company he expected to hold for several years. Now, that early exit isn’t a call that AI is over – Tom and Eric both believe that AI has plenty more runway. But here’s Tom with the takeaway that you and I need to factor in: The financial risks of AI are simply growing too steep… and companies are turning to some rather creative ways of hiding the true costs of expansion. In our experience, this is a recipe for something going wrong down the road. As a quick reminder, for months, Eric has urged investors to take profits on overextended tech names, redirecting capital into earlier-stage, undervalued companies that can benefit from the AI wave without excessive valuation risk. He recently released a “Sell This, Buy That” that identifies several high-conviction swaps for navigating this environment (he gives away the names of three of his top picks for free). You can check it out for free right here. We’ll keep you updated on all these stories here in the Digest. Have a good evening, Jeff Remsburg Disclaimer: I own Amazon, Alphabet, and Microsoft. | 
