King Coal's Not Dead Yet

And neither is US offshore wind

It looks like Nvidia Corp (NSDQ: NVDA) will carry the US stock market on its back a while longer. Following robust calendar Q4 results and guidance for artificial intelligence related sales, the stock last week recorded a one-day market value gain of $277 billion.

 Nvidia’s market capitalization is now larger than the GDP of all but 11 countries. And the Wall Street consensus is about as bullish as it gets, with 59 research firms and investment banks tracked by Bloomberg Intelligence rating the stock “buy” versus just six holds and zero sells.

 If you’ve been investing for any length of time, chances are you’ve seen this movie before. An utterly dominant company emerges in an industry with apparently inexhaustible growth and every investor wants a piece of it. The stock acquires what looks to be unstoppable momentum and any measure of business valuation gets thrown out the window—before something happens to burst the bubble and it’s look out below.

Cisco Systems (NSDQ: CSCO) the past couple decades plus is a pretty good example, transforming the world forever as a business but with a generally mediocre performing stock. And at 65 times trailing 12 months earnings, 46 times actual equity value and nearly 70 times cash flow, Nvidia is definitely attaining the heights of investor expectation Cisco did for a previous generation—even as AI is upending entire industries.

Is there an Nvidia crash in our future? It would be unprecedented if there isn’t. And given how heavy a weighting it now has in the S&P 500, any serious decline will take a massive bite out of any S&P 500 ETF—and by extension anyone who’s now investing passively in stocks, rather than actively managing their fate.

 The Nvidia/AI revolution is, however, already having a profound impact on perhaps America’s most essential industry: Electricity.

 “Electrify everything” has been the mantra of business and government globally for the better part of a decade now—the idea being electricity is a clean source of energy with fungible ways to produce and therefore far more predictable prices than the fossil fuels that would otherwise be used.

 Artificial intelligence, however, is a massive power hog. And that means massive and largely unplanned for new demand for electricity from rapidly proliferating data centers, which handle the applications.

Some estimate AI’s suck on power will boost data centers’ share of total US electricity demand to as high as 7.5 percent by 2030, versus 1 to 1.5 percent just a couple years ago. And that share is of a system that was already preparing to handle annual demand growth of 2 to 3 percent a year.

 That’s the fastest growth rate in decades for power. It’s in part from off again, on again adoption of electric vehicles. But increasingly the bigger driver is electrification of industries including leading shale oil and gas companies like Pioneer Resources (NYSE: PXD), which has electrified its drilling process to cut costs.

Bottom line: AI and data centers have an exploding share of a rapidly growing market. And as a result, it’s effectively “all hands on deck” in the US electricity sector.

Among the biggest beneficiaries are coal mining companies, along with owners of coal-burning power plants that are being granted at least a temporary reprieve from closing. Earlier this month, for example, West Virginia regulators ruled FirstEnergy Corp (NYSE: FE) can keep two major coal power plants running a while longer, reducing the immediate burden of shutting them down. And other states are allowing utilities the same flexibility, thereby reducing the ultimate financial cost of replacing the lost energy.

The benefits to coal mining companies can also scarcely be overstated. The fact the current coal power plant fleet will eventually be shuttered hasn’t changed. The vast majority of facilities are simply aging past the point where it’s possible to operate them economically. There’s also absolutely no appetite to build new coal-fired power plants in the US, given the cost of ensuring environmental standards comparable to available natural gas-fueled facilities. And as we saw from 2017-2020, even the most pro-coal White House is powerless to reverse coal closures.

But a slower pace does mean supply contracts stretching out just a bit longer. And that’s greatly improving financial and operating flexibility for mining companies, giving them critical time to cut costs and debt, carve out new markets overseas and even transition to new fuels like natural gas.

Another major beneficiary of the AI boom’s draw on power: US Offshore Wind.

The projected boom has at this point become a bust, with all but a handful of previously planned projects either shelved or cancelled. And the combination of higher interest rates and construction/materials costs well above expectations has basically frozen the announcement of new projects, while making potential customers unwilling to commit to purchasing output.

Nonetheless, one major facility has entered service, Avangrid Inc’s (NYSE: AGR) 800-megawatt capacity Vineyard 1. And the state of Massachusetts is already reporting savings for the plant’s utility customers from the replacement of pricey LNG imports.

This week, Dominion Energy (NYSE: D) announced it will sell a 50 percent ownership interest in its 2.6-gigawatt capacity Coastal Virginia Offshore Wind facility to private capital firm Stonepeak. The deal basically amounts to a 50-50 sharing of development costs.

At the expected close in Q4, Stonepeak will pay Dominion $3 billion, an amount that’s roughly half of construction costs projected to be incurred to that point. If costs exceed the $9.8 billion currently estimated to complete the project—a number that’s held since Virginia regulators approved CVOW in August 2022—the partners will split costs 50-50 up to a total of $11.3 billion.

If the project winds up costing between $11.3 billion and $13.7 billion, Dominion will bear between 67 and 83 percent of the difference. Stonepeak will have the option not to make additional capital contributions, though at the cost of a reduced ownership stake in the project.

I expect this deal to win relatively smooth approval from Virginia and North Carolina regulators, as well as the Bureau of Ocean Energy Management and other federal agencies. That’s in equal parts because of the Biden Administration’s support for offshore wind and the fact that the states have already approved a rate deal that includes utility responsibility for any costs over the budgeted $8.9 billion.

The key issue for both Dominion and Stonepeak is keeping a lid on the final construction costs for CVOW. In my view, the potential for overruns of the magnitude to hurt either is low.

For one thing, Vineyard Wind apparently did not incur significant unforeseen costs in its final stages of construction. Dominion has locked in over 92 percent of remaining costs, with an unused “contingency fund” of $351 million substantially reserving against overruns on the remaining 8 percent. And risk the key remaining variable continues to drop—timely delivery in “late 2024 to early 2025” of the company’s offshore wind construction vessel Charybdis.

The vessel is now reportedly more than 82 percent complete and total projected costs unchanged from most recent estimates at $625 million. Owning Charybdis eliminates risk to Dominion of having critical equipment called away if deadlines are missed and contracts expire, always a possibility for large construction projects—which has in fact already happened to at least one US offshore wind developer. And the partners will be able to lease it out to other projects, once the heavy lifting is done at CVOW 

Bottom line: There’s still a lot of work to do at CVOW. But Dominion has struck an amicable deal with a seasoned infrastructure investors that has upside. And impressively, it’s done so at a time when other US offshore wind developers have written billions of dollars off the value of their projects.

The includes EverSource Inc (NYSE: ES), the New England utility that had partnered with Orsted (Denmark: ORSTED, OTC: DNNGY) in three offshore wind projects. The partners have now written down the value of their development efforts by several billion dollars, reflecting the combination of unexpected project costs and buyers’ unwillingness to raise the price of contracts to compensate.

EverSource has now announced a deal to divest its remaining ownership in two projects currently under construction. That includes cost sharing of potential overruns from here on out, a far less advantageous arrangement than Dominion’s deal with Stonepeak.

One reason Dominion got a much better deal: CVOW is a regulated project with guaranteed returns if the partners bring it into service in line with the current budget. That’s a stark contrast with the Revolution and South Fork projects EverSource is exiting, which depend on signing contracts for output at prices that reflect costs.

In my view, that’s a tacit endorsement of the more than 120-year-old integrated regulated utility business model—when it comes to executing on multi-year, multi-billion dollar infrastructure projects. But in any case, Dominion’s deal means offshore wind isn’t dead yet in this country. And the more AI pushes up electricity demand, the more likely we’ll see more projects.