OpenAI’s CFO recently suggested that Uncle Sam should backstop AI chip financing—essentially asking taxpayers to guarantee the riskiest capital costs for “frontier labs.” As The Information reported, the idea drew immediate pushback from tech peers who questioned why a company preparing for a $500 billion valuation—and possibly a trillion-dollar IPO—can’t raise its own money. Why should the public underwrite a firm whose private investors are already minting generational wealth?
Meanwhile, the Department of Energy is opening federal nuclear and laboratory sites—from Idaho National Lab to Oak Ridge and Savannah River—for private AI data centers, complete with fast-track siting, dedicated transmission lines, and priority megawatts. DOE’s expanded Title XVII loan-guarantee authority sweetens the deal, offering government-backed credit and low borrowing costs. It’s a breathtaking case of public risk for private expansion, at a time when ordinary ratepayers are staring down record-high energy bills.
And the ambition goes further. Some of these companies now plan to site small modular nuclear reactors to provide dedicated power for AI data centers. That means the next generation of nuclear power—built with public financing and risk—could end up serving private compute clusters, not the public grid. In a country already facing desertification, water scarcity, and extreme heat, it is staggering to watch policymakers indulge proposals that will burn enormous volumes of water to cool servers, while residents across the Southwest are asked to ration and conserve. I theoretically don’t have a problem with private power grids, but I don’t believe they’ll be private and I do believe that in both the short run and the long run these “national champions” will drive electricity prices through the stratosphere—which would be OK, too, if the AI labs paid off the bonds that built our utilities. All the bonds.
At the same time, Washington still refuses to enforce copyright law, allowing these same firms to ingest millions of creative works into their models without consent, compensation, or disclosure—just as it did under DMCA §512 and Title I of the MMA, both of which legalized “ingest first, reconcile later.” That’s a copyright subsidy by omission, one that transfers cultural value from working artists into the balance sheets of companies whose business model depends on denial.
And the timing? Unbelievable. These AI subsidies were being discussed in the same week SNAP benefits are running out and the Treasury is struggling to refinance federal debt. We are cutting grocery assistance to families while extending loan guarantees and land access to trillion-dollar corporations.
If DOE and DOD insist on framing this as “AI industrial policy,” then condition every dollar on verifiable rights-clean training data, environmental transparency, and water accountability. Demand audits, clawbacks, and public-benefit commitments before the first reactor breaks ground.
Until then, this is not innovation—it’s industrialized arbitrage: public debt, public land, and public water underwriting the private expropriation of America’s creative and natural resources.
The smart people want us to believe that artificial intelligence is the frontier and apotheosis of human progress. They sell it as transformative and disruptive. That’s probably true as far as it goes, but it doesn’t go that far. In practice, the infrastructure that powers it often dates back to a different era and there is the paradox: much of the electricity to power AI’s still flows through the bones of mid‑20th century engineering. Wouldn’t it be a good thing if they innovated a new energy source before they crowd out the humans?
The Current Generation Energy Mix — And What AI Adds
To see that paradox, start with the U.S. national electricity mix:
In 2023 , the U.S. generated about 4,178 billion kWh of electricity at utility-scale facilities. Of that, 60% came from fossil fuels (coal, natural gas, petroleum, other gases), 19% came from nuclear, and 21% from renewables (wind, solar, hydro). – Nuclear power remains the backbone of zero-carbon baseload: it supplies around 18–19% of U.S. electricity, and nearly half of all non‑emitting generation. – In 2025, clean sources (nuclear + renewables) are edging upward. According to Ember, in March 2025 fossil fuels fell below 50% of U.S. electricity generation for the first time (49.2%), marking a historic shift. – Yet still, more than half of US power comes from carbon-emitting sources in most months.
Meanwhile, AI’s demand is surging:
– The Department of Energy estimates that data centers consumed 4.4% of U.S. electricity in 2023 (176 TWh) and projects this to rise to 6.7–12% by 2028 (325–580 TWh) according to the Department of Energy. – An academic study of 2,132 U.S. data centers (2023–2024) found that these facilities accounted for more than 4% of national power consumption, with 56% coming from fossil sources, and emitted more than 105 million tons of CO₂e (approximately 2.18% of U.S. emissions in 2023). – That study also concluded: data centers’ carbon intensity (CO₂ per kWh) is 48% higher than the U.S. average.
So: AI’s power demands are no small increment—they threaten to stress a grid still anchored in older thermal technologies.
When I say AI is running on 1960s technology, I mean several things:
1. Thermal generation methods remain largely unchanged according to the EPA. Coal-fired steam turbines and natural gas combined-cycle plants still dominate.
2. Many plants are old and aging. The average age of coal plants in the U.S. is about 43 years; some facilities are over 60. Transmission lines and grid control systems often date from mid-to late-20th century planning.
3. Nuclear’s modern edge is historical. Most U.S. nuclear reactors in operation were ordered in the 1960s–1970s and built over subsequent decades. In other words: The commercial installed base is old.
The Rickover Motif: Nuclear, Legacy, and Power Politics
To criticize AI’s reliance on legacy infrastructure, one powerful symbol is Admiral Hyman G. Rickover, the man often called the “Father of the Nuclear Navy.” Rickover’s work in the 1950s and 1960s not only shaped naval propulsion but also influenced the civilian nuclear sector.
Rickover pushed for rigorous engineering standards , standardization, safety protocols, and institutional discipline in building reactors. After the success of naval nuclear systems, Rickover was assigned by the Atomic Energy Commission to influence civilian nuclear power development.
Rickover famously required applicants to the nuclear submarine service to have “fixed their own car.” That speaks to technical literacy, self-reliance, and understanding systems deeply, qualities today’s AI leaders often lack. I mean seriously—can you imagine Sam Altman on a mechanic’s dolly covered in grease?
As the U.S. Navy celebrates its 250th anniversary, it’s ironic that modern AI ambitions lean on reactors whose protocols, safety cultures, and control logic remain deeply shaped by Rickover-era thinking from…yes…1947. And remember, Admiral Rickover had to transfer the hidebound Navy to nuclear power which at the time was just recently discovered and not well understood—and away from diesel. Diesel. That’s innovation and required a hugely entrepreneurial leader.
The Hypocrisy of Innovation Without Infrastructure
AI companies claim disruption but site data centers wherever grid power is cheapest — often near legacy thermal or nuclear plants. They promote “100% renewable” branding via offsets, but in real time pull electricity from fossil-heavy grids. Dense compute loads aggravate transmission congestion. FERC and NERC now list hyperscale data centers as emerging reliability risks.
The energy costs AI doesn’t pay — grid upgrades, transmission reinforcement, reserve margins — are socialized onto ratepayers and bondholders. If the AI labs would like to use their multibillion dollar valuations to pay off that bond debt, that’s a conversation. But they don’t want that, just like they don’t want to pay for the copyrights they train on.
Innovation without infrastructure isn’t innovation — it’s rent-seeking. Shocking, I know…Silicon Valley engaging in rent-seeking and corporate welfare.
The 1960s Called. They Want Their Grid Back.
We cannot build the future on the bones of the past. If AI is truly going to transform the world, its promoters must stop pretending that plugging into a mid-century grid is good enough. The industry should lead on grid modernization, storage, and advanced generation, not free-ride on infrastructure our grandparents paid for.
Admiral Rickover understood that technology without stewardship is just hubris. He built a nuclear Navy because new power required new systems and new thinking. That lesson is even more urgent now.
Until it is learned, AI will remain a contradiction: the most advanced machines in human history, running on steam-age physics and Cold War engineering.
If we let a hyped “AI gap” dictate land and energy policy, we’ll privatize essential infrastructure and socialize the fallout.
Every now and then, it’s important to focus on what our alleged partners in music distribution are up to, because the reality is they’re not record people—their real goal is getting their hands on the investment we’ve all made in helping compelling artists find and keep an audience. And when those same CEOs use the profits from our work to pivot to “defense tech” or “dual use” AI (civilian and military), we should hear what that euphemism really means: killing machines.
Daniel Ek is backing battlefield-AI ventures; Eric Schmidt has spent years bankrolling and lobbying for the militarization of AI while shaping the policies that green-light it. This is what happens when we get in business with people who don’t share our values: the capital, data, and social license harvested from culture gets recycled into systems built to find, fix, and finish human beings. As Bob Dylan put it in Masters of War, “You fasten the triggers for the others to fire.” These deals aren’t value-neutral—they launder credibility from art into combat. If that’s the future on offer, our first duty is to say so plainly—and refuse to be complicit.
The same AI outfits that for decades have refused to license or begrudgingly licensed the culture they ingest are now muscling into the hard stuff—power grids, water systems, and aquifers—wherever governments are desperate to win their investment. Think bespoke substations, “islanded” microgrids dedicated to single corporate users, priority interconnects, and high-volume water draws baked into “innovation” deals. It’s happening globally, but nowhere more aggressively than in the U.S., where policy and permitting are being bent toward AI-first infrastructure—thanks in no small part to Silicon Valley’s White House “AI viceroy,” David Sacks. If we don’t demand accountability at the point of data and at the point of energy and water, we’ll wake up to AI that not only steals our work but also commandeers our utilities. Just like Senator Wyden accomplished for Oregon.
These aren’t pop-up server farms; they’re decades-long fixtures. Substations and transmission are built on 30–50-year horizons, generation assets run 20–60, with multi-decade PPAs, water rights, and recorded easements that outlive elections. Once steel’s in the ground, rate designs and priority interconnects get contractually sticky. Unlike the Internet fights of the last 25 years—where you could force a license for what travels through the pipe—this AI footprint binds communities for generations; it’s essentially forever. So we will be stuck for generations with the decisions we make today.
Because China–The New Missle Gap
There’s a familiar ring to the way America is now talking about AI, energy, and federal land use (and likely expropriation). In the 1950s Cold War era, politicians sold the country on a “missile gap” that later proved largely mythical, yet it hardened budgets, doctrine, and concrete in ways that lasted decades.
Today’s version is the “AI gap”—a story that says China is sprinting on AI, so we must pave faster, permit faster, and relax old guardrails to keep up. Of course, this diverts attention from China’s advances in directed-energy weapons and hypersonic missiles which are here right now today and will play havoc in an actual battlefield—which the West has no counter to. But let’s not talk about those (at least not until we lose a carrier in the South China Sea), let’s worry about AI because that will make Silicon Valley even richer.
Watch any interview of executives from the frontier AI labs and within minutes they will hit their “because China” talking point. National security and competitiveness are real concerns, but they don’t justify blank checks and Constitutional-level safe harbors. The missile‑gap analogy is useful because it reminds us how a compelling threat narrative propaganda can swamp due diligence. We can support strategic compute and energy without letting an AI‑gap story permanently bulldoze open space and saddle communities with the bill.
Energy Haves (Them) and Have Nots (Everyone else)
The result is a two‑track energy state AKA hell on earth. On Track A, the frontier AI lab hyperscalers like Google, Meta, Microsoft, OpenAI & Co. build company‑town infrastructure for AI—on‑site electricity generation by microgrids outside of everyone else’s electric grid, dedicated interties and other interconnections between electric operators—often on or near federal land. On Track B, the public grid carries everyone else: homes, hospitals, small manufacturers, water districts. As President Trump said at the White House AI dinner this week, Track A promises to “self‑supply,” but even self‑supplied campuses still lean on the public grid for backup and monetization, and they compete for scarce interconnection headroom.
President Trump is allowing the hyperscalers to get permanent rights to build on massive parcels of government land, including private utilities to power the massive electricity and water cooling needs for AI data centers. Strangely enough, this is continuing a Biden policy under an executive order issued late in Biden Presidency that Trump now takes credit for, and is a 180 out from America First according to people who ought to know like Steve Bannon. And yet it is happening.
White House Dinners are Old News in Silicon Valley
If someone says “AI labs will build their own utilities on federal land,” that land comes in two flavors: Department of Defense (now War Department) or Department of Energy sites and land owned by the Bureau of Land Management (BLM). This are vastly different categories. DoD/DOE sites such as Idaho National Laboratory Oak Ridge Reservation, Paducah GDP, and the Savannah River Site, imply behind-the-fence, mission-tied microgrids with limited public friction; BLM land implies public-land rights-of-way and multi-use trade-offs (grazing, wildlife, cultural), longer timelines, and grid-export dynamics with potential “curtailment” which means prioritizing electricity for the hyperscalers. For example, Idaho National Laboratory (INL) as one of the four AI/data-center sites. INL’s own environmental reports state that about 60% of the INL site is open to livestock grazing, with monitoring of grazing impacts on habitat. That’s likely over.
This is about how we power anything not controlled by a handful of firms. And it’s about the land footprint: fenced solar yards, switchyards, substations, massive transport lines, wider roads, laydown areas. On BLM range and other open spaces, those facilities translate into real, local losses—grazable acres inside fences, stock trails detoured, range improvements relocated.
What the two tracks really do
Track A solves a business problem: compute growth outpacing the public grid’s construction cycle. By putting electrons next to servers (literally), operators avoid waiting years for a substation or a 230‑kV line. Microgrids provide islanding during emergencies and participation in wholesale markets when connected. It’s nimble, and it works—for the operator.
Track B inherits the volatility: planners must consider a surge of large loads that may or may not appear, while maintaining reliability for everyone else. Capacity margins tighten; transmission projects get reprioritized; retail rates absorb the externalities. When utilities plan for speculative loads and those projects cancel or slide, the region can be left with stranded costs or deferred maintenance elsewhere.
The land squeeze we’re not counting
Public agencies tout gigawatts permitted. They rarely publish the acreage fenced, AUMs affected, or water commitments. Utility‑scale solar commonly pencils out to on the order of 5–7 acres per megawatt of capacity depending on layout and topography. At that ratio, a single gigawatt occupies thousands of acres—acres that, unlike wind, often can’t be grazed once panels and security fences go in. Even where grazing is technically possible, access roads, laydown yards, and vegetation control impose real costs on neighboring users.
Wind is more compatible with grazing, but it isn’t footprint‑free. Pads, roads, and safety buffers fragment pasture. Transmission to move that energy still needs corridors—and those corridors cross someone’s water lines and gates. Multiple use is a principle; on the ground it’s a schedule, a map, and a cost. Just for reference, a rule‑of‑thumb for acres/electricity produces is approximately 5–7 acres per megawatt of direct current (“MWdc”), but access roads, laydown, and buffers extend beyond the fence line.
We are going through this right now in my part of the world. Central Texas is bracing for a wave of new high-voltage transmission. These are 345-kV corridors cutting (literally) across the Hill Country to serve load growth for chip fabricators and data centers and tie-in distant generation (so big lines are a must once you commit to the usage). Ranchers and small towns are pushing back hard: eminent-domain threats, devalued land, scarred vistas, live-oak and wildlife impacts, and routes that ignore existing roads and utility corridors. Packed hearings and county resolutions demand co-location, undergrounding studies, and real alternatives—not “pick a line on a map” after the deal is done. The fight isn’t against reliability; it’s against a planning process that externalizes costs onto farmers, ranchers, other landowners and working landscapes.
Texas’s latest SB 6 is the case study. After a wave of ultra-large AI/data-center loads, frontier labs and their allies pushed lawmakers to rewrite reliability rules so the grid would accommodate them. SB 6 empowers the Texas grid operator ERCOT to police new mega-loads—through emergency curtailment and/or firm-backup requirements—effectively reshaping interconnection priorities and shifting reliability risk and costs onto everyone else. “Everyone else” means you and me, kind of like the “full faith and credit of the US”. Texas SB 6 was signed into law in June 2025 by Gov. Greg Abbott. It’s now in effect and directs PUCT/ERCOT to set new rules for very large loads (e.g., data centers), including curtailment during emergencies and added interconnection/backup-power requirements. So the devil will be in the details and someone needs to put on the whole armor of God, so to speak.
The phantom problem
Another quiet driver of bad outcomes is phantom demand: developers filing duplicative load or interconnection requests to keep options open. On paper, it looks like a tidal wave; in practice, only a slice gets built. If every inquiry triggers a utility study, a route survey, or a placeholder in a capital plan, neighborhoods can end up paying for capacity that never comes online to serve them.
A better deal for the public and the range
Prioritize already‑disturbed lands—industrial parks, mines, reservoirs, existing corridors—before greenfield BLM range land. Where greenfield is unavoidable, set a no‑net‑loss goal for AUMs and require real compensation and repair SLAs for affected range improvements.
Milestone gating for large loads: require non‑refundable deposits, binding site control, and equipment milestones before a project can hold scarce interconnection capacity or trigger grid upgrades. Count only contracted loads in official forecasts; publish scenario bands so rate cases aren’t built on hype.
Common‑corridor rules: make developers prove they can’t use existing roads or rights‑of‑way before claiming new footprints. Where fencing is required, use wildlife‑friendly designs and commit to seasonal gates that preserve stock movement.
Public equity for public land: if a campus wins accelerated federal siting and long‑term locational advantage, tie that to a public revenue share or capacity rights that directly benefit local ratepayers and counties. Public land should deliver public returns, not just private moats.
Grid‑help obligations: if a private microgrid islands to protect its own uptime, it should also help the grid when connected. Enroll batteries for frequency and reserve services; commit to emergency export; and pay a fair share of fixed transmission costs instead of shifting them onto households.
Or you could do what the Dutch and Irish governments proposed under the guise of climate change regulations—kill all the cattle. I can tell you right now that that ain’t gonna happen in Texas.
Will We Get Fooled Again?
If we let a hyped latter day “missile gap” set the terms, we’ll lock in a two‑track energy state: private power for those who can afford to build it, a more fragile and more expensive public grid for everyone else, and open spaces converted into permanent infrastructure at a discount. The alternative is straightforward: price land and grid externalities honestly, gate speculative demand, require public returns on public siting, and design corridor rules that protect working landscapes. That’s not anti‑AI; it’s pro‑public. Everything not controlled by Big Tech—will be better for it.
Let’s be clear: the data-center onslaught will be financed by the taxpayer one way or another—either as direct public outlays or through sweet-heart “leases” of federal land to build private utilities behind the fence for the richest corporations in commercial history. After all the goodies that Trump is handing to the AI platforms, let’s not have any loose talk of “selling” excess electricity to the public–that price should be zero. Even so, the sales pitch about “excess” electricity they’ll generously sell back to the grid is a fantasy; when margins tighten, they’ll throttle output costs, not volunteer philanthropy. Picture it: do you really think these firms won’t optimize for themselves first and last? We’ll be left with the bills, the land impacts, and a grid redesigned around their needs. Ask yourself—what in the last 25 years of Big Tech behavior says “trustworthy” to you?
In a twist of post–Cold War irony, the same federal sites that once forged the isotopes of nuclear deterrence are now poised to fuel the arms race of artificial intelligence under the leadership of Special Government Employee and Silicon Valley Viceroy David Sacks. Under a new Department of Energy (DOE) initiative, 16 legacy nuclear and lab sites — including Savannah River, Idaho National Lab, and Oak Ridge Tennessee — are being opened to private companies to host massive AI data centers. That’s right–Tennessee where David Sacks is riding roughshod over the ELVIS Act.
But as this techno-industrial alliance gathers steam, one question looms large: Who benefits — and how will the American public be compensated for leasing its nuclear commons to the world’s most powerful corporations? Spoiler alert: We won’t.
A New Model, But Not the Manhattan Project
This program is being billed in headlines as a “new Manhattan Project for AI.” But that comparison falls apart quickly. The original Manhattan Project was: – Owned by the government – Staffed by public scientists – Built for collective defense
Today’s AI infrastructure effort is: – Privately controlled – Driven by monopolies and venture capital – Structured to avoid transparency and public input – Uses free leases on public land with private nuclear reactors
Call it the Manhattan Project in reverse — not national defense, but national defense capture.
The Art of the Deal: Who gets what?
What Big Tech Is Getting
– Access to federal land already zoned, secured, and wired – Exemption from state and local permitting – Bypass of grid congestion via nuclear-ready substations – DOE’s help fast-tracking nuclear microreactors (SMRs) – Potential sovereign AI training enclaves, shielded from export controls and oversight
And all of it is being made available to private companies called the “Frontier labs”: Microsoft, Oracle, Amazon, OpenAI, Anthropic, xAI — the very firms at the center of the AI race.
What the Taxpayer Gets (Maybe)
Despite this extraordinary access, almost nothing is disclosed about how the public is compensated. No known revenue-sharing models. No guaranteed public compute access. No equity. No royalties.
Land lease payments? Not disclosed. Probably none. Local tax revenue? Minimal (federal lands exempt) Infrastructure benefit sharing? Unclear or limited
It’s all being negotiated quietly, under vague promises of “national competitiveness.”
Why AI Labs Want DOE Sites
Frontier labs like OpenAI and Anthropic — and their cloud sponsors — need: – Gigawatts of energy – Secure compute environments – Freedom from export rules and Freedom of Information Act requests – Permitting shortcuts and national branding
The DOE sites offer all of that — plus built-in federal credibility. The same labs currently arguing in court that their training practices are “fair use” now claim they are defenders of democracy training AI on taxpayer-built land.
This Isn’t the Manhattan Project — It’s the Extraction Economy in a Lab Coat
The tech industry loves to invoke patriotism when it’s convenient — especially when demanding access to federal land, nuclear infrastructure, or diplomatic cover from the EU’s AI Act. But let’s be clear:
This isn’t the Manhattan Project. Or rather we should hope it isn’t because that one didn’t end well and still hasn’t. It’s not public service. It’s Big Tech lying about fair use, wrapped in an American flag — and for all we know, it might be the first time David Sacks ever saw one.
When companies like OpenAI and Microsoft claim they’re defending democracy while building proprietary systems on DOE nuclear land, we’re not just being gaslit — we’re being looted.
If the AI revolution is built on nationalizing risk and privatizing power, it’s time to ask whose country this still is — and who gets to turn off the lights.
Music Tech Solutions 