National Anchor Power: Going Beyond Hyperscalers, Recalling the Past
My National Anchor Power Program was pitched as a way to sell public clean firm power to hyperscalers in the public benefit. It needn’t be just for them though.
In my post earlier this week I summarized my recent essay for the Breakthrough Institute about transmission expansion, public power, and my proposal for a National Anchor Power Program. I also included a few points that didn’t make it into the essay. Not an hour after sending it out, however, I kicked myself for omitting a major one I’d intended days ago to write up: none of it is fundamentally tied to hyperscalers and their artificial intelligence data centers.
In the NAPP architecture, the hyperscalers play two fundamental roles:
Bankrollers: The hyperscaler companies investing heavily in data center expansion for their AI services have demonstrated a willingness to pay a premium for power infrastructure. Moreover, a number of center-left thinkers have argued that they cough up their extensive capital to fund grid infrastructure that the rest of us depend on. In the NAPP, they are not just signing long-term offtake agreements for the clean firm power, which pays off the generation investment as well as the transmission investment; with CWIP-like payments they’re also funding the Reserve Fund that's used to pool cost overrun risk insurance.
Loads: Hyperscalers are responsible for large loads — from low hundreds of MW to even GWs — in utility territories across the country. And as EPRI has studied, the large datacenter load is relatively constant: “The average annual load factor relative to nameplate capacity was 75% for the hyperscale facility, while the average across the smaller co-location facilities was 57%,” they report. It’s this additive load in a utility territory that is well-suited to being provided with firm power.
The role of bankroller is a hard one to fill with another industrial customer. Who else has the deep pockets and presents a ripe target for political agitation that would make them pay if they want their investments?
But the role of load, the offtakers of NAPP power, could be filled by myriad industrial customers with high load factors. Semiconductor manufacturing spurred by the CHIPS Act presents a grid challenge for Arizona, where the major utility serving Phoenix will see 1 GW demand from TSMC’a facility, and New York, where Micron’s facility will draw almost 2 GW.
Or the offtakers of NAPP power could be aluminum smelters, water desalination plants, clean steel manufacturers, synthetic fuel producers, or any other electricity-intensive industry, all of which will pencil better with constant power delivery than with flexible operations to match intermittent power. And unlike data centers, a lot of these will offer long-term manufacturing jobs to locals.
The main idea is that load growth is good for America and for utility customers so long as the capital investment needed to serve it is fairly allocated. When industrial customers of a utility buy power at a high load factor — that is, when their load shape is largely flat; when they purchase kWhs of energy close to their peak demand 24/7 — it helps amortize major power investments that serve everyone.
That logic goes back to the New Deal public power of old, too. When FDR signed into law the statute creating the Power Authority of the State of New York, the legislation explicitly cited this beneficial role played by “high load factor” industry.1 A couple years later, as President, he signed the TVA Act with almost identical language:2
This policy is further declared to be that the projects herein provided for shall be considered primarily as for the benefit of the people of the section as a whole and particularly the domestic and rural consumers to whom the power can economically be made available, and accordingly that sale to and use by industry shall be a secondary purpose, to be utilized principally to secure a sufficiently high load factor and revenue returns which will permit domestic and rural use at the lowest possible rates and in such manner as to encourage increased domestic and rural use of electricity.
How many people realize the phrase “high load factor” was enshrined into the laws that created NYPA and TVA?
When the Bonneville Power Administration was getting off the ground in 1938, the question of power allocation and prices was hotly contested. Advocates of public power in Washington State, which had seen numerous municipal utilities blossom, demanded that the Bonneville dam's power be sold only to the people and not to private interests. Original Bonneville administrator J. D. Ross announced intentions to sell power also to private companies, a decision which met outcry from regional advocates. But Judson King, head of the National Popular Government League and a staunch public power champion nationwide,
condemned the alleged champions of the public interest for not having thought “the thing through or faced the necessity of amortization and what must be done to repay government loans.” They had failed to recognize that even the large municipal plants of Seattle, Tacoma, and Los Angeles could not sell electricity as cheaply as they did without simultaneously selling larger blocks to private industry at still lower rates. Defending Ross’s silence on the rate question before all the facts were in, King urged the administrator to propagandize the need for a balanced load on any public utility.3
Early in its life TVA signed contracts to sell power long-term to Monsanto, the chemical company, and Alcoa, the aluminum smelter, accounting for about 28% and 21% of total sales in 1938, respectively.4 It was the only way to recover costs for the massive hydropower projects, especially while legal battles held back municipal utility takeovers in larger cities like Chattanooga. Thirty years later all industrial customers only soaked up about 23% of TVA’s sales.5 During the Cold War the largest chunk of TVA power went to the federal customers involved in uranium enrichment; see my post from early last year on the intense power needs of the Cold War vs. a hyperscaler data center today.
The National Anchor Power Program, as I’ve conceived it, offers pooled construction risk for public power utilities building and selling standardized clean, firm power to neighboring utilities, with federal involvement in the transmission to break the FERC logjams. What states like New York and Tennessee have to offer desired industries in terms of public power, the NAPP grants to other states as well. Instead of being directed by hyperscaler customers, perhaps those neighboring utilities could partner with their states in allocating slices of anchor power accordingly, for example, with competitive contracts for offtakers that meet certain criteria, like number of permanent jobs created. The specificity to hyperscalers in my essay mattered for the political and financial side, not the program architecture or operations.
Philip J. Funigiello, Toward a National Power Policy: The New Deal and the Electric Utility Industry, 1933–1941, 1973, pp. 199-200.