If GENCO is the answer, then what is the question? Simply put, is GENCO the answer to the question of how the electric industry can get into the AI action without compromising utility finances and disadvantaging the rest of the customers?
First, some history. The word GENCO (short for generating company) was applied to independent (unregulated) generators that proliferated like rabbits after deregulation at the turn of the century. A large number of them went bankrupt due to overcapacity, demand below expectations, and too much debt. Since utilities did not own local GENCOs, neither they nor their customers had to bail out the distressed GENCOs.
Now to the AI load. Regulated utilities would love to get this business, but if they put expensive new power stations built specifically to serve AI into their rate base, everyone’s electric bill will rise, meaning that all customers would pay more to subsidize the AI load. And even worse, if the AI load did not materialize or dropped sharply, the rest of the utility’s customers would still have to pay for that newly added, expensive power generating capacity. So, how about setting up a separate GENCO to serve the AI load, affiliated with the utility and its holding company, and finance it with its own debt, separately from that of the utility? Then, if the GENCO runs into difficulties, the utility and its customers, “ring-fenced” from the GENCO, won’t be directly affected. Put the lawyers and bankers to work. Sounds great.
Related: Uniper to Sell Gas Pipeline Stake to Meet EU State Aid Conditions
Of course, its success in insulating consumers from AI risks depends on the details. For instance, does the utility contract to buy the electricity for a term of years from the affiliated GENCO and then sell it to the AI center, with a contract that passes on all costs? Well, what if the duration of the utility’s contract to purchase the power exceeds the actual (rather than the planned) life of the AI project? The utility and its customers are stuck, right? What if the utility improperly defines all the costs? Will consumers or the utility pay the shortfall? These are not academic questions. The AI firms have the money to splash around in political circles and in lawsuits, and the regulators often do not have the knowledge to judge the issues critically, leaving consumers with the short end of the stick.
From a financial standpoint, the GENCO idea is not as clean as it looks, either. First, if the utility signs a contract to buy from GENCO, GENCO will use that contract as security to raise funds. It will, in effect, borrow the utility’s credit rating in order to obtain the lowest-cost debt. That is, utility consumers support the utility credit rating, and GENCO gets the benefit, without paying the utility’s customers for the favor. Second, if GENCO gets into trouble, it will affect the financial standing of the holding company that controls GENCO and the utility; and because the holding company finances utility equity, that will affect the cost of the equity component of the utility’s capital cost structure. Finally, management facing serious financial difficulties at a major subsidiary might just find ingenious ways to move money to it, despite ring-fencing.
Of course, there is another GENCO alternative. Just build a GENCO outside of the regulated utility that sells its power directly to the AI center. One complication is making sure that the GENCO does not get any special privileges from the affiliated utility, which might leave other customers hard up when the utility favors its associate. Most public service regulators nowadays will carefully scrutinize any utility affiliates that take services from the utility in a way that might prejudice other competitors or consumers. Imagine the questions after a storm, when the utility’s affiliated GENCO and its AI customer get their service back first? Public service means public service for all on equal terms.
We have discussed credit implications, piercing the corporate veil, and management distraction from GENCO. There are two other seldom addressed. First is the implications of the size disparity of the AI companies vs. utilities, beyond what we have addressed. It’s not just that they’re asking for power. It’s that they’re asking for huge amounts with enormous financial implications. (It’s like going to a diner and ordering one hamburger. That’s fine. If you order fifty, you overwhelm the cooks, the waiters etc.). Apart from the financial implications, it’s not clear that the grid system was even built for this level of size disparity. Second, the environmental harm caused by all these new gas-fired facilities ( Musk’s facility in suburban Memphis, for example) represents a big political vulnerability. If progressive politicians win at the national level, and clean air/water standards are again enforced, these new facilities will find their operating environment much more challenging. So, a restoration of environmental concerns, as part of a political platform, may be the biggest risk here. It’s also the risk that gets the least discussion because of status quo bias and a rightward-leaning news bias in business media. (Electricity executives and American investors don’t, in general, read Mother Jones or The Guardian).
Now, back to all these warnings. You might argue that all these worries are overdone. The AI promoters know what they are doing. Utility managers will assess risks rationally. Regulators will protect consumers. AI is just one more new technology that the world will manage to absorb. Maybe, but we have seen what happens when fear of missing out (FOMO) affects decision-making, when desperation turns to fraud and regulators supinely take the company’s word for it. As for AI, we aren’t technology experts, but AI developers regularly write warnings that AI will end civilization and that many AI firms will go under; a top Wall Street bank told its clients that AI will be one of the shortest tech revolutions in history because quantum computing will overtake it sooner than previously expected; a Nobelist warned that there are low barriers to entry into AI and users of AI find the Chinese version far more economical than the American models, so you can draw your own conclusions. Mind you, we don’t object to investors pouring billions (or trillions) into AI. Investors evaluate risks and returns, and then voluntarily put up the money. We just don’t think that captive utility consumers should be subjected to risks of higher costs and less reliable service, without any compensatory payoff. It’s a simple matter of equity.
So back to the original question. To us, the GENCO is a clever cosmetic workaround that partially mitigates the AI problem but does not make it go away. In our view, the only way to protect consumers is to require the AI center to own and operate its own power facilities off the grid. Even that does not protect the utility consumer from the competition for energy resources and water by the AI center, but competition for resources is part of the free enterprise system. Dumping risks and expenses onto captive, regulated consumers is not. At least not yet.
By Leonard Hyman and William Tilles for Oilprice.com
More Top Reads From Oilprice.com
Oilprice Intelligence brings you the signals before they become front-page news. This is the same expert analysis read by veteran traders and political advisors. Get it free, twice a week, and you’ll always know why the market is moving before everyone else.
You get the geopolitical intelligence, the hidden inventory data, and the market whispers that move billions – and we’ll send you $389 in premium energy intelligence, on us, just for subscribing. Join 400,000+ readers today. Get access immediately by clicking here.
