AI’s Infrastructure War
Idaho as Ground Zero for America’s Data-Center Boom
America’s AI build-out is arriving in Idaho—fast. In less than five years, Kuna alone has gone from quiet farmland to two hyperscale data-center megaprojects: Meta’s 960,000-sq-ft campus and the 620-acre Gemstone Technology Park. This collection investigates, with receipts, what that shift really means for Idaho’s power grid, water, jobs, taxes, and local governance—and what smarter alternatives could look like.
We start by separating “critical” from “non-essential” AI. Not every GPU hour is created equal; some workloads (defense, emergency, science) deserve priority when the grid is tight, while commercial recommendation engines can be time-shifted or curtailed without risking lives. A category framework makes those trade-offs legible—and actionable for policy and planning.
Then we get specific about Idaho. Meta’s Kuna site will be among Idaho Power’s largest customers ever, with bespoke energy contracts, new solar PPAs, storage, and dedicated delivery facilities structured to avoid cost-shifts onto other ratepayers. We unpack what’s public, what’s redacted, and where the risk still sits if loads ramp or exit.
Water and land are the other pressure points. Cooling strategies pivot on the Western Snake River Plain aquifer, where historic irrigation recharge is declining and nitrate concerns are rising; we trace the well rights, reuse plans, and municipal hookups that now anchor these campuses. On land use and renewables, we size what it would take to “self-match” a ~1,050 MW AI load with solar in Idaho’s capacity-factor reality—tens of thousands of acres and billions in CAPEX—and why that matters for siting.
Jobs and tax claims get a forensic treatment. Hyperscale centers are capital-intensive but people-sparse: ~100 long-term jobs per campus on hundreds of acres, with many construction trades imported during the build. We contrast those outcomes with small-business and light-industrial scenarios that deliver far higher jobs-per-acre and broader local multipliers, and we explain how URDs and targeted exemptions reshape who actually sees near-term revenue.
Finally, we follow the process itself—NDAs, shell LLCs, code names, modified procedures—and ask whether Idaho’s three-member, no-advocate PUC model and city-level dealmaking are built to guard the public interest amid a national AI land rush. Where they aren’t, we offer a repairs list: clearer curtailment priorities, stronger credit spines in special contracts, transparency standards, and a small-business-first playbook that doesn’t trade power and water for thin returns.
AI as a National Imperative
Artificial Intelligence has vaulted beyond its origins as an IT tool—it is now a pillar of national power equal to economic strength and military might. U.S. leaders describe AI as the defining technology of our era, capable of transforming decision-making, military readiness, cyber defense, and logistics at a scale no human system can match.
Unlike traditional software, AI is trained on immense datasets using power-hungry computing clusters that rival small cities in energy demand. Training a single frontier model can consume megawatts of electricity in days, making AI development resemble an industrial enterprise more than a coding project. This infrastructure is what fuels the speed, scale, and scope of AI’s national security impact.
Five mechanisms explain AI’s game-changing role. First, it compresses decision cycles—enabling commanders to move from observation to action in seconds, outpacing adversaries operating on human time scales. Second, it turns raw data into foresight, as seen in predictive intelligence systems that shift surveillance from reactive to proactive. Third, it supercharges cyber operations, defending against machine-driven threats while enabling counter-offensives at unprecedented speed. Fourth, AI autonomy powers swarms of drones, robotic vehicles, and unmanned ships that extend U.S. reach and resilience. Finally, AI optimizes logistics and maintenance, ensuring supply chains and fleets remain battle-ready even under stress.
Adversaries understand the stakes. China has declared AI central to its military strategy, pouring resources into autonomous systems, surveillance, and massive data centers. Russia too signals AI ambition, framing it as decisive in future warfare. The National Security Commission on AI warns bluntly: to face AI-enabled adversaries without AI is “an invitation to disaster.”
The lesson is clear: AI is no longer “just IT.” It is the high ground of 21st-century power. Securing U.S. leadership in AI is not optional—it is a national security imperative.
The AI Balkanization Trap
America’s AI infrastructure isn’t being built by strategy—it’s being stitched together by NDAs, shell companies, code names, and rushed local votes. The chapter shows how a national-scale asset (hyperscale/AI data centers) is forming through thousands of fragmented city, county, utility, and state decisions—with minimal coordination and maximum opacity. The result: forum-shopping, tax-break bidding wars, grid and water surprises, and a rising community backlash that is now slowing, stalling, or scattering capacity.
You’ll see the playbook up close: early NDAs gag public officials; code-named proposals (“Project Blue,” “Project Atlas,” “Project Pine”) obscure who’s really behind a site; and shell LLCs (e.g., Brisbie LLC for Meta; Jet Stream LLC for Google) mask ownership in land and power filings until the deal is effectively a fait accompli. The table on page 4 summarizes how common these tactics are—and why they work so well inside siloed local processes.
Three case studies make the pattern undeniable. Kuna, Idaho: Meta’s $800M campus advanced via a sealed energy contract and a front company, with the utility and PUC reacting case-by-case as the region’s largest single load suddenly appeared. Then, in the same town, a developer advanced a 620-acre “Gemstone” tech park widely believed to be for Google—again without naming the end user—while offering selectively sized “community” payments that left schools underfunded by orders of magnitude. Virginia: the world’s data-center capital ran into hard limits (noise, substations, water, transmission) and a bipartisan revolt; billions were delayed or withdrawn as counties rewrote rules in real time. Mississippi: the opposite extreme—an ultra-fast, ultra-generous package for AWS (“Project Atlas”) that traded a decade of tax exemptions for minimal permanent jobs, with state-backed infrastructure financing to boot.
The through-line: fragmentation enables secrecy and speed—until it triggers mistrust, grid strain, and whiplash policy. Without transparent standards and coordination, the U.S. risks over-concentration in some regions, suboptimal siting elsewhere, mounting subsidy races, and even national-security exposure from poorly distributed critical infrastructure. This chapter argues for escaping the “Balkanization trap” before it hardens into costly, brittle reality.
AI Compute Categorization Framework
Artificial Intelligence is not just about algorithms—it’s about who gets access to the most powerful compute. Chapter 3 introduces a three-tier framework for prioritizing U.S. AI resources, dividing workloads into mission-critical defense AI, strategic science AI, and general commercial AI.
Category A-1 covers essential warfighting and national intelligence uses. This includes AI-driven targeting (e.g., the Army’s FIRESTORM system, which compresses 20-minute kill-chains into seconds), Project Maven’s real-time image recognition for ISR, and DARPA’s Mosaic Warfare vision of swarming autonomous systems. These applications demand secure, segregated infrastructure—classified networks, air-gapped clusters, and government-only cloud regions like AWS Secret and Azure Government Top Secret.
Category A-2 encompasses strategic national advantage workloads: AI for climate modeling, pandemic forecasting, advanced materials, and energy grid resilience. DOE’s exascale supercomputers, such as Frontier at Oak Ridge, are already producing breakthroughs—shrinking years of climate simulations into days and powering epidemic models that guide CDC response. These scientific gains are national security assets in their own right, ensuring resilience against global shocks.
Category B represents the vast commercial AI sector: generative models, social media algorithms, and enterprise AI services. These rely on hyperscale data centers run by tech giants, with siting and power driven by market forces. While vital to innovation, they are not privileged by national-security rules—except when public-private partnerships temporarily channel their compute toward A-2 science efforts.
Segmentation is already happening: classified enclaves for A-1, federally funded supercomputing centers for A-2, and global commercial clouds for B. But the policy challenge is ensuring the right balance—protecting critical defense and science workloads without starving the commercial innovation engine. As Eric Schmidt bluntly put it, America must “go all in on building AI data centers” even if it complicates climate goals, while others warn that over-segmentation risks stifling open innovation.
The bottom line: compute is the currency of AI power. How the U.S. allocates and safeguards it will determine whether we maintain military overmatch, scientific leadership, and global competitiveness.
Regulatory Quicksand and Lobbying Influence
Regulating artificial intelligence in the United States has become a test of political will against the most powerful lobbying machine in Washington. The tech sector now rivals, and in some cases surpasses, the traditional giants of influence such as pharmaceuticals. Collectively, the largest technology and AI firms spend hundreds of millions of dollars each year to shape policy, ensuring that proposed rules are either delayed, diluted, or steered in their favor. Amazon, Meta, Alphabet, Microsoft, Apple, Oracle, and other players in the AI race devote vast sums to lobbying, while newer entrants like ByteDance and Palantir have joined the fray with record spending.
This concentration of money and influence has turned AI into a top-tier priority for lobbyists, creating an environment where meaningful regulation faces an uphill battle. Even well-intentioned bills can be stalled indefinitely under the weight of industry resources, leaving policymakers hesitant to challenge the very companies driving the technology forward.
The stakes are high. If Washington remains entangled in lobbying pressure and procedural delays, America risks ceding leadership in AI to rivals who move faster under more flexible or centralized systems. The challenge is clear: overcoming the gravitational pull of Big Tech’s influence is essential if the U.S. hopes to set its own standards for AI rather than adapting to those imposed elsewhere.
This chapter lays bare the scale of lobbying dollars behind the AI industry—and the difficult road ahead for anyone seeking to regulate it
China’s AI Power Play
China has made artificial intelligence the centerpiece of its strategy for military power, domestic control, and economic growth. In the military domain, the People’s Liberation Army is testing AI-enabled targeting, drone swarms, and real-time intelligence fusion under the banner of “intelligentized warfare.” Civilian breakthroughs in robotics and surveillance are rapidly folded into defense through Military-Civil Fusion, blurring the line between private innovation and state security. Yet much remains experimental, and questions persist about whether the PLA can translate prototypes into reliable battlefield systems.
AI also underpins the state’s vast security architecture. The “Skynet” network of millions of cameras, paired with facial recognition and predictive policing platforms, has given Chinese authorities unmatched surveillance reach—capable of tracking fugitives in minutes but also drawing criticism for human rights abuses. These data streams double as training sets for military and intelligence algorithms, reinforcing China’s edge.
Beyond defense and security, Beijing is embedding AI across public systems and infrastructure. Smart classrooms, adaptive traffic systems, and AI-driven healthcare are spreading nationwide, backed by a talent pipeline that produces more STEM graduates than any other country. At the infrastructure level, the “Eastern Data, Western Computing” initiative channels renewable energy into massive data center hubs, providing the computing backbone for AI ambitions.
Together, these efforts reveal a state-directed model: AI as a tool of national power. Success depends on overcoming bottlenecks—talent retention, chip shortages, and overcapacity—but the scale and speed of China’s push are reshaping the global balance of technological competition.
China’s AI Strategy: Military & Security Intelligentized Warfare
China is embedding AI into military doctrine, from drone swarms to predictive logistics, but battlefield effectiveness remains unproven. Doctrine and ambition outpace tested capability.
China’s military modernization is now defined by its doctrine of “intelligentized warfare”—a vision that places artificial intelligence at the center of future combat power. The People’s Liberation Army (PLA) has formally anchored intelligentization in strategy documents and speeches from Xi Jinping, describing a transformation from mechanization and informatization toward a battlefield dominated by autonomy, machine learning, and human–machine teaming.
Evidence shows progress across several fronts. AI-enabled decision aids are being tested in command centers to accelerate targeting and logistics planning. Procurement records highlight PLA investments in intelligent ISR, automated target recognition, and predictive logistics. Swarm demonstrations of autonomous drones and loyal wingman prototypes suggest significant advances in unmanned systems, even if most fielded platforms remain human-supervised. Military-Civil Fusion is drawing on civilian firms to adapt commercial robotics, vision systems, and speech AI for defense use.
At the same time, bottlenecks remain. Access to high-end semiconductors is constrained by U.S. export controls, and PLA writings acknowledge technology gaps compared to leading militaries. Operational autonomy is still cautious—China has not fully entrusted lethal decisions to AI. Much of the doctrine remains aspirational, with uneven deployment across branches.
The chapter concludes that China is pushing hard toward AI-driven warfighting, building the foundations for faster decision cycles, saturation drone attacks, and data-driven logistics. Yet the true battlefield advantage is unproven. Intelligentized warfare could redefine the balance of power if realized—or prove a hollow ambition if systems fail under the pressures of combat.
China’s AI Strategy: Intelligence, State Security, Surveillance and Control
AI fusion platforms, predictive policing, and vast surveillance networks fuel China’s digital panopticon. Integration grows, but bias, data silos, and public backlash expose cracks.
China is fusing artificial intelligence into its state security architecture, creating one of the most expansive surveillance and control systems in the world. At the core are “AI fusion platforms” that integrate data from cameras, smartphones, social media, financial records, and biometric databases. In regions like Xinjiang, the Integrated Joint Operations Platform has already demonstrated how multi-source data can be combined to flag “abnormal” behavior—from unusual electricity use to frequent mosque visits—resulting in pre-emptive detentions. Similar city-level platforms, often built with Chinese tech giants, enable police to locate suspects in hours rather than days.
AI-driven predictive policing is another key element. Algorithms profile individuals based on travel, ethnicity, or online behavior, generating alerts before crimes or protests occur. While authorities credit these tools with preventing terrorism and unrest, they are also notorious for false positives and systemic bias. Facial recognition, gait analysis, and massive CCTV networks under Skynet and Sharp Eyes have further reduced anonymity, extending from megacities to rural villages.
The state has also applied AI to censorship and propaganda. Automated filters block banned content in real time, while generative systems help flood social media with pro-Party messaging. Combined with the social credit system’s punitive data pipelines, this forms a feedback loop of digital control: more AI surveillance produces more compliance, which in turn legitimizes further expansion.
Yet gaps remain. Data silos, overreliance on flawed algorithms, and public backlash risk undermining the system’s effectiveness. Still, China’s AI-powered security model represents a formidable fusion of technology and authoritarian governance.
China’s AI Strategy Public Systems, Talent & Infrastructure Foundations
China’s twin investments—AI education pipelines and massive data centers—form a reinforcing ecosystem. Scale and speed are unmatched, but innovation quality and sustainability remain uncertain.
ina’s public-sector AI strategy rests on two foundations: human capital and digital infrastructure. Together, these elements form a deliberate national approach to scaling both “brains and silicon.”
On the talent side, Beijing has transformed its education system to produce AI specialists at unprecedented scale. Through the “Double First-Class” university plan and dedicated AI institutes, dozens of top universities now offer AI-focused majors and cross-disciplinary “AI + X” programs. The result has been tens of thousands of AI-trained graduates each year, with China surpassing other nations in both publication volume and patent filings. While concerns remain about quality and brain drain, the sheer size of the pipeline gives China a durable advantage in staffing its AI ambitions.
Infrastructure is the other pillar. Massive data center clusters—channeled through the Eastern Data, Western Compute initiative—are relocating workloads to energy-rich inland provinces. Powered increasingly by renewables and supported by ultra-high-voltage transmission lines, these centers provide the compute backbone for model training and national-scale digital projects. Innovation in cooling systems, including undersea data centers, is reducing environmental costs even as overall energy consumption grows.
Industrial policy ties these strands together. Central targets cascade into provincial subsidies, cheap land, and credit, ensuring alignment across universities, state-owned enterprises, and private firms. The outcome is a reinforcing “AI flywheel”: talent feeds research, infrastructure supports deployment, and government policy sustains both.
This chapter shows how education, energy, and infrastructure interlock to give China the foundation for long-term AI competitiveness—though sustainability and innovation quality remain open questions.
Meta’s Kuna Datacenter Secrecy, Incentives, and Hidden Costs
Behind shell companies and NDAs, Meta secured tax breaks, clean energy deals, and infrastructure in Kuna—leaving the public sidelined until the project was irreversible
Meta’s entry into Kuna, Idaho was orchestrated with stealth and speed. Through shell company Brisbie LLC and the codename “Project Peregrine,” the company quietly acquired 325 acres, negotiated utility deals, and secured approvals while the public remained in the dark. By the time Meta’s identity was revealed in 2022, Kuna’s leaders had already created an urban renewal district, advanced infrastructure plans, and signed agreements that locked the city into long-term commitments.
Central to the deal was a bespoke Energy Services Agreement with Idaho Power. Meta became the first customer under a clean energy tariff designed to fund new renewable projects. While pitched as protecting other ratepayers from cost-shifts, the arrangement guaranteed Meta decades of subsidized energy certainty. Simultaneously, Kuna accepted Meta’s construction of a new $50–100 million wastewater plant—handed to the city but designed primarily for the data center’s needs. Officials hailed it as a community asset, though ongoing operating costs now fall to local taxpayers.
Idaho’s 2020 data center tax exemption added another layer, granting Meta a permanent sales tax holiday on equipment. Lawmakers later realized the combination of state breaks and Kuna’s urban renewal district left little immediate benefit for residents, prompting new laws to close loopholes.
This chapter shows how secrecy, incentives, and rapid approvals shaped a project touted as transformative—but whose real returns to Kuna are modest: about 100 permanent jobs, significant water draw, and infrastructure costs that may outlast Meta’s promises.
Meta’s Kuna Data Center: NDAs, Shell Companies & Oversight Failures
Kuna approved Meta’s billion-dollar data center under NDAs and shell companies, keeping residents in the dark on water, power, and costs until it was too late.
Meta’s Kuna data center deal reveals how secrecy, shell companies, and weak oversight enabled one of Idaho’s largest industrial projects with little public input. Operating under the codename “Project Peregrine,” nearly 1,100 acres of farmland were annexed and rezoned in 2021 while city officials were bound by non-disclosure agreements. Public records and hearings mentioned only “Brisbie LLC,” Meta’s shell entity, leaving citizens unaware that a Fortune 500 company was shaping the city’s future behind closed doors.
The NDA-driven secrecy extended beyond land. Kuna’s leaders approved development agreements, water arrangements, and special energy contracts without openly disclosing Meta’s role. Even the school district accepted a token mitigation payment without realizing the project’s true scale. Critical questions—water use from the local aquifer, 120+ megawatts of power demand, long-term costs of new infrastructure—were kept out of public debate until approvals were irreversible.
This pattern reflects a broader playbook: Idaho’s pro-business model encourages fast-tracking deals through tax breaks, confidentiality, and light regulation. In Kuna, officials largely acted as facilitators rather than gatekeepers, failing to demand independent studies or enforce transparency. Promises of jobs, infrastructure, and community investment were trumpeted only after the project was fait accompli, leaving residents feeling blindsided and distrustful.
This chapter exposes how manufactured secrecy and governance shortcuts tilted the balance of power toward Meta, sidelining public accountability. It sets the stage for a deeper reckoning: whether the limited benefits of 100 jobs and tax concessions justify the massive trade-offs imposed on Kuna’s community and resources.
Gemstone Technology Park Secrecy, Incentives & Kuna’s Next Data Center Gamble
Kuna’s second mega data center, Gemstone, was approved by one vote after secret negotiations—offering millions in contributions but raising fresh questions on water, power, and fairness.
Kuna’s approval of the Gemstone Technology Park shows that Idaho’s data center saga is far from over. Just three years after Meta’s controversial entry, Diode Ventures (a subsidiary of Black & Veatch) secured rezoning for a 620-acre, billion-dollar data center campus under the same shroud of NDAs, code names, and withheld details that characterized Meta’s “Project Peregrine.” Residents only learned the true scale when the deal was nearly sealed.
The project promises eye-catching numbers: more than $40 million in contributions to fire, police, and schools; 100 high-paying jobs; and hundreds of millions added to the tax base. Yet the details reveal imbalance. The school district initially accepted just $500,000—later admitting it didn’t understand the project’s scope until too late—while police and fire walked away with tens of millions. Water and power demands, likely exceeding Meta’s already immense footprint, remain largely unquantified in public documents, leaving citizens guessing about long-term impacts.
Kuna’s City Council split 2–2 on approval before the mayor broke the tie in favor, underscoring just how divisive the project has become. Supporters framed the data center as a better alternative to housing subdivisions, citing lower strain on schools and traffic. Opponents warned that approving another mega-project outside the designated industrial zone violates the city’s plan, erodes farmland, and locks Kuna deeper into a low-jobs, high-land model of growth.
Gemstone’s approval, by a single vote, cements Kuna’s path toward becoming a data-center corridor. Whether that makes the city a regional “tech hub” or merely Big Tech’s utility outpost remains an open—and contested—question.
Idaho’s AI Data Centers: Power Demands & Electrical Infrastructure
Meta retools to a 200–275 MW “AI-ready” campus; Gemstone targets 600–800 MW. Idaho must add generation, a new substation, and airtight tariffs—fast—or households pay the price.
Idaho is about to absorb one of the fastest load spikes in its history, driven by two hyperscale builds in Kuna. Meta’s 960,000-sq-ft “AI-ready” campus is no longer a 30–100 MW legacy design; triangulated methods (chip-to-rack math using 100–120 kW AI racks, modern W/ft² densities, and like-for-like comparables) defend a 200–275 MW full-build envelope. That puts a single private campus in the power class of a small city—and explains why Idaho Power created a bespoke Schedule 33 contract, tied resource procurements to Meta’s load, and ring-fenced costs so other ratepayers aren’t stuck with the tab.
Gemstone Technology Park is bigger still. At 620 acres and five planned buildings, public testimony and technical summaries place Gemstone’s full build in the 600–800 MW range, effectively a new electric city grafted onto the Treasure Valley grid. Delivery requires a new on-site substation tied into high-voltage corridors, phased feeders, hundreds of MW of new generation under “clean construction” tariffs, and a small army of backup generators to meet hyperscale reliability standards.
No explicit statutory cap throttles either project; the real “limits” are what the utility can deliver without shifting costs or jeopardizing reliability. Idaho Power’s planning now leans on accelerated wind/solar/storage plus transmission upgrades to keep pace. Bottom line: Meta sets the precedent; Gemstone stresses the system. Together, they could push ~1 GW of incremental load—about a quarter of Idaho Power’s recent peak—if both fully materialize. The question isn’t whether these megaprojects are feasible; it’s whether Idaho can scale supply, transmission, and contracts fast enough to keep lights on and rates fair while hyperscalers chase AI at unprecedented density【
Water Demands of Kuna’s Data Centers Aquifer Stress, Offsets, and Secrecy
Meta and Gemstone promise “less water than farms,” yet both pump Kuna’s aquifer year-round, evaporating 80% and returning little in summer—the season recharge matters most.
Beneath Kuna’s farmland sits the Treasure Valley aquifer, a confined basalt–alluvium system that historically relied on flood irrigation to recharge. With canals paving the city’s growth, that recharge has collapsed—and now two hyperscale data centers are set to draw on the aquifer year-round.
Meta’s 960,000-sq-ft campus, operating through Brisbie LLC, is tied into Kuna’s municipal wells. Gemstone Technology Park, backed by Diode Ventures, will drill its own wellfield on 620 acres. Both projects rely entirely on groundwater, with consumptive cooling systems that evaporate up to 80% of withdrawals. Return flows are minimal: Meta routes treated wastewater to a 30-acre reuse field; Gemstone promises to infiltrate tower blowdown through basins “treated to drinking water standards.” Neither approach replaces the massive summer recharge once provided by irrigated crops.
Developers argue these facilities “use less than agriculture,” citing IDWR’s 33% reduction rule when irrigation rights are converted to industrial. But this is annual accounting sleight-of-hand. Farms returned large volumes to the aquifer during peak summer, when water is most scarce. Data centers consume steadily all year, returning only a trickle in winter when recharge is least valuable. Offset pledges—Meta’s “water positive by 2030,” Gemstone’s “33% reduction”—are non-probative when measured against seasonal aquifer dynamics.
The risks are clear: declining groundwater levels, nitrate concentration in a priority area, and local well interference. Permits impose monitoring and quality standards, but no enforceable requirement for aquifer neutrality. The result is legalized overdraft disguised as sustainability—a slow trade-off of Idaho’s most precious resource for Big Tech’s insatiable compute.
Solar Sizing and Land Requirements
To keep Idaho’s new AI data centers humming 24/7, the math gets staggering—billions in panels, 40 000 acres of land, and a whole lot of sunlight.
In the high desert outside Boise, the new temples of artificial intelligence are rising—vast data centers consuming more electricity than some cities. To keep them running purely on sunshine would require a solar buildout so large it borders on science fiction.
A recent analysis found that powering a continuous 1,050-megawatt AI load in Idaho would demand roughly five times that much in solar capacity—about 4 to 6 gigawatts of panels covering 35 000 to 45 000 acres of land, or nearly 70 square miles. Even the optimistic scenario, using the most efficient tracking systems, still needs more than 30 000 acres and billions of dollars in hardware. Idaho’s biggest solar farm today is 120 MW on 1 000 acres; the AI era would require dozens of such projects stitched together across the Snake River Plain.
At current prices, the tab runs $7 to $12 billion—and that’s before batteries. Because Idaho’s winter sun produces barely a third of summer output, these panels would oversupply the grid by day and leave the data centers stranded by night. Utilities are already scrambling to pair big solar fields with equally massive batteries, trying to keep pace with corporate clients like Meta, whose 250 MW campus in Kuna will rely on a trio of bespoke solar deals.
Behind the scenes, the global supply chain adds another twist. More than 80 percent of the world’s solar modules come from China or Southeast Asia, and inverters are similarly concentrated. A single trade dispute or component shortage could ripple through every project on Idaho Power’s books.
It’s a grand promise—and a warning. Solar can offset the AI load, but it can’t do it alone. Meeting round-the-clock computing demand will take more than sunshine; it will require storage, resilience, and political will to match the technological ambition.
Idaho Power & Hyperscale Provisioning Meta, Gemstone, and the Utility’s Balancing Act
Meta’s ESA funds solar, storage, and its own interconnection; Gemstone is poised to follow. Idaho Power profits, but only strict oversight keeps ratepayers off the hook.
Idaho Power has become a central broker in Idaho’s data center boom, with Meta’s Kuna campus and the proposed Gemstone Technology Park forcing the utility into unprecedented buildouts. Its parent, IDACORP, benefits financially—new capital projects expand the regulated rate base, driving ~9% returns—but regulators have pressed to ensure those gains don’t come at ratepayer expense.
Meta’s Energy Services Agreement exemplifies this delicate balance. The utility procured three massive solar PPAs—Pleasant Valley 1 (200 MW), Pleasant Valley 2 (125 MW), and Blacks Creek (320 MW)—exclusively for Meta, alongside a 200 MW/800 MWh Boise Bench battery. All costs flow directly to Meta under the Clean Energy Your Way program, while renewable energy credits (RECs) are delivered to Meta for its “100% renewable” pledge. Meanwhile, Meta funded its own interconnection—new substation and transmission lines—through a Contribution in Aid of Construction, ensuring no roll-in to general rates. The contract also contains minimum billing provisions, credit guarantees, and exit penalties, locking Meta into paying for its full load even if usage falls or the company walks away.
Gemstone is expected to follow the same template: a bespoke ESA, dedicated renewables, and CIAC-funded infrastructure. But as of late 2025, no filings confirm who will carry the generation risk or how costs will be guaranteed.
The stakes are high. These hyperscale contracts let Idaho Power align customer sustainability goals with grid expansion, but they also concentrate risk. If even one mega-project falters, stranded costs and rate shocks loom. Oversight by the Idaho PUC is now the critical backstop to ensure “customer pays ⇒ ratepayers safe” remains more than a slogan.
Who Guards the Gate Idaho PUC, Missing Advocate, and Hyperscale “No-Harm” Deals
No consumer advocate. Weak revolving-door limits. Sealed contract guarantees. Idaho’s “held harmless” mantra for data centers rests on trust—without the receipts the public deserves
Idaho’s data-center era arrived before its guardrails. Unlike 47 states and D.C., Idaho has no independent utility consumer advocate—no Public Counsel in the AG’s office, no statutory ratepayer watchdog—leaving residential and small-business customers to rely on Commission Staff inside the very agency they’re asking to scrutinize the utility. Structural gaps compound the problem: Idaho’s ethics code bars commissioners from owning utility stock, but it lacks a modern revolving-door cooling-off ban, making post-service utility employment largely unrestricted compared with peers like Nevada and Washington. Transparency lags, too—key special-contract exhibits (prices, guarantees) are sealed, so the public must trust summary assurances rather than verify terms.
Against that backdrop, Idaho Power’s hyperscale deals (Meta/“Brisbie” and Gemstone) hinge on the promise that “customer pays ⇒ ratepayers safe.” On paper, Meta’s contract is strong: high minimum demand (20 MW floor), customer-funded interconnection (CIAC), dedicated renewable PPAs billed to Meta, discounted credits for surplus, and a required stranded-cost guarantee. But the proof remains redacted—the identity/amount of the guaranty, termination formulas, and any assignment limits are not public. Gemstone’s even earlier: no filed ESA yet; “no-harm” is still an aspiration rather than a binding instrument.
Meanwhile, Idaho Power is racing billions in new capex through rate cases as load surges, with residential increases outpacing large-power percentages in proposals—while special-contract customers pay bespoke charges outside many pooled costs. Without an independent advocate, Idaho’s “who guards the gate?” answer is effectively: the gatekeeper guards itself. The fix is straightforward: require redacted public versions of special contracts (including guaranties), enact a cooling-off ban, and establish a consumer counsel so the public’s case isn’t made from the gallery. Until then, “held harmless” remains a promise that Idahoans can’t fully audit.
Idaho’s Data Center Tax Break Origins, Stacking, and Open‑Ended Exposure
Sold as homeowner relief, Idaho’s data‑center tax break became a permanent subsidy: anti‑stacking helps, but no sunset and scant reporting leave taxpayers on the hook indefinitely
Idaho’s data center incentive was sold as property‑tax relief wrapped in a jobs bill. Lawmakers added Section 63‑3622VV to exempt “information technology equipment” and construction inputs from the 6% sales/use tax if a project certifies $250M investment within five years and 30 jobs at or above county wage within two—then converts to a permanent, facility‑specific exemption on all future equipment purchases. The pitch sweetened the deal by excluding the project’s value from the “new construction roll,” implying levy relief for homeowners. Guardrails barred stacking with the state’s TRI program. In practice, this created a generous, no‑sunset sales‑tax holiday tethered to modest job thresholds and minimal public reporting.
Local policy collided with the promise: Kuna initially funneled Meta’s property taxes into a new urban renewal district (URD), undermining base‑broadening. The legislature responded with HB 328, forcing an either/or choice—take the state exemption or the URD, not both—an unusually hard anti‑stacking line compared to peer states. A later bid to cap the exemption at seven years for future projects advanced in the House but stalled, leaving the break uncapped and permanent for qualifying facilities .
The price tag is large and opaque. Meta’s ~$800M build implies tens of millions in foregone state revenue; Gemstone’s planned ~$1B campus would be similar. Permanent headcount is small (≈100 jobs per site), pushing public cost per job into the high six figures, even before considering future expansions that also avoid sales tax. Locals gain a sizable property‑tax base and negotiated service contributions; the state general fund shoulders the subsidy without routine deal‑level disclosure. Whether the exemption actually tipped site selection remains unproven—company statements emphasize power, climate, and workforce, not tax policy. The net: one‑bite anti‑stacking helps, but no sunset + no transparency = open‑ended exposure
Tax Incentives and Business Attraction Subsidies Idaho’s Costly Mirage
Idaho’s subsidies promise jobs but mostly pay companies for what they’d do anyway—while taxpayers lose revenue, services, and leverage in a costly race to the bottom.
Idaho, like most states, leans heavily on targeted tax incentives and business subsidies to lure investment—data centers, factories, “transformational” projects. The political pitch is simple: tax breaks and cash packages “create jobs” and diversify the economy. The empirical record is brutal. Independent studies find that 75–98% of subsidized firms would have invested anyway, meaning billions in foregone revenue are wasted. When subsidies do “tip” a decision, the cost per job routinely exceeds $200,000, sometimes into the millions.
Economists describe the dynamic as a zero-sum arms race. States outbid one another for firms, shifting jobs geographically but not creating new ones nationally. The public sees ribbon-cuttings; what they don’t see are the trade-offs: fewer teachers hired, higher homeowner taxes, potholes left unfixed. These diffuse costs are hidden while politicians collect visible credit for “winning” projects. The imbalance persists because benefits are concentrated (companies, construction trades, local boosters) while costs are scattered across taxpayers.
Idaho’s 2020 data center exemption shows the trap. Pitched as relief for homeowners via a larger property-tax base, it quickly morphed into open-ended subsidies for Meta’s Kuna campus. Lawmakers later scrambled to limit stacking incentives and cap exemptions, belatedly acknowledging the imbalance. But the broader pattern remains: corporate welfare survives because politics reward short-term optics, not long-term economics.
The evidence is clear: subsidies distort markets, reward lobbying, and exacerbate inequality while doing little for true competitiveness. Real development comes from fundamentals—talent, infrastructure, broad-based tax reform—not from writing checks to the biggest bidder.
The AI Data Center Sales Pitch Idaho’s Promised Payoff
Cheap power, renewable energy, and billion-dollar builds—Idaho’s AI data centers are sold as job creators and infrastructure heroes. The next chapter asks: do they deliver?
Idaho’s leaders and tech recruiters frame hyperscale AI data centers as a once-in-a-generation opportunity: billions in private investment, high-skill jobs, and a stake in the nation’s AI future. The sales pitch is polished and data-driven. Idaho, they argue, offers the perfect mix—cheap, clean power, low disaster risk, and a pro-business climate—positioning the state as an “ideal location for data centers”. Meta’s $800 million campus in Kuna and Diode Ventures’ proposed $1 billion Gemstone Technology Park are showcased as proof: thousands of construction jobs, over 200 permanent tech positions, and hundreds of millions added to the tax base.
Developers and officials emphasize a suite of secondary benefits. Data centers, they claim, act as “anchor tenants” that fund infrastructure the community couldn’t otherwise afford—Meta’s $50 million water-reclamation system for Kuna, Diode’s planned substation and road improvements, and new fiber networks for future businesses. Utilities echo the message: large 24/7 electrical loads create rate stability, spread costs across more kilowatt-hours, and—thanks to custom “no-cost-shift” contracts—don’t burden other ratepayers. The companies’ sustainability promises add another layer of appeal: 100% renewable energy through new solar farms, advanced cooling technology, and even water reuse projects touted as environmental wins.
Add philanthropy, community grants, and workforce partnerships, and the picture sold to policymakers is compelling. The narrative: AI data centers pay their own way, strengthen infrastructure, stabilize power costs, fund schools and police, and secure Idaho’s place in America’s AI supply chain. It’s a confident vision—the upside case—that the next chapter will test.Idaho’s leaders and tech recruiters frame hyperscale AI data centers as a once-in-a-generation opportunity: billions in private investment, high-skill jobs, and a stake in the nation’s AI future. The sales pitch is polished and data-driven. Idaho, they argue, offers the perfect mix—cheap, clean power, low disaster risk, and a pro-business climate—positioning the state as an “ideal location for data centers”. Meta’s $800 million campus in Kuna and Diode Ventures’ proposed $1 billion Gemstone Technology Park are showcased as proof: thousands of construction jobs, over 200 permanent tech positions, and hundreds of millions added to the tax base.
Developers and officials emphasize a suite of secondary benefits. Data centers, they claim, act as “anchor tenants” that fund infrastructure the community couldn’t otherwise afford—Meta’s $50 million water-reclamation system for Kuna, Diode’s planned substation and road improvements, and new fiber networks for future businesses. Utilities echo the message: large 24/7 electrical loads create rate stability, spread costs across more kilowatt-hours, and—thanks to custom “no-cost-shift” contracts—don’t burden other ratepayers. The companies’ sustainability promises add another layer of appeal: 100% renewable energy through new solar farms, advanced cooling technology, and even water reuse projects touted as environmental wins.
Add philanthropy, community grants, and workforce partnerships, and the picture sold to policymakers is compelling. The narrative: AI data centers pay their own way, strengthen infrastructure, stabilize power costs, fund schools and police, and secure Idaho’s place in America’s AI supply chain. It’s a confident vision—the upside case—that the next chapter will test.
Forensic Rebuttal The AI Data Center Mirage
Behind Idaho’s AI data-center hype lies a hollow deal: few jobs, thin tax gains, soaring power and water demand—and subsidies masking extraction as innovation.
The sales pitch for Idaho’s AI data centers collapses under forensic scrutiny. The $800 million investment headlines and “high-paying jobs” sound transformative—until you examine the ratios. Meta’s nearly million-square-foot facility will permanently employ about 100 people, roughly one job per six acres or $8 million in capital per position. Construction booms are fleeting and largely staffed by out-of-state contractors because Idaho’s trades were already fully employed. The supposed “multiplier effect” is mostly short-term service work—security, catering, or landscaping—not sustainable high-wage employment.
The infrastructure “gifts” fare no better. The $50 million water-reclamation plant Meta built wasn’t philanthropy—it was a prerequisite to cool its own servers. Likewise, the 200 MW solar farm and substation upgrades exist to feed the project’s 24/7 power hunger, not Idaho’s green transition. Utility filings confirm that data centers will drive a 20 percent rise in statewide electricity demand, triggering a $4 billion grid-expansion plan that ratepayers will ultimately finance. The “no-cost-shift” contracts meant to protect households are only as strong as corporate permanence; if a facility downscales or exits, stranded-asset costs flow to everyone else.
Tax benefits are overstated. Kuna homeowners’ much-touted “4.8% property-tax drop” equals roughly $110 per year—while the state forfeited $30–50 million in sales tax for Meta alone. The few community grants ($500k total) are PR rounding errors beside Idaho’s lost revenue and resource strain. What’s left is a lightly staffed fortress consuming a city’s worth of power and water—its quiet presence more emblem of extraction than of progress. Idaho’s “AI revolution” looks less like a tech boom and more like a subsidized mirage.
Meta’s “Job Creation” Reexamined Imported Labor, Local Illusion
Meta’s Kuna data center promised local prosperity but relied on imported crews. Idaho paid millions in tax breaks for a project that built servers—not Idaho careers.
Idaho officials sold Meta’s $800 million Kuna data center as a jobs engine for “hardworking Idahoans.” The evidence dismantles that narrative. At peak, roughly 1,200–1,400 construction workers swarmed the site—but the majority were imported from out of state. Meta’s general contractor, Hensel Phelps of Colorado, and subs like Cupertino Electric and LPR Construction recruited nationally, offering $120–$240 per diem, travel pay, and lodging to fill the gap. Idaho’s own trades were already at full employment—unemployment near 2–3%, virtually no local slack. The state had about 1,500 electricians in the entire Treasure Valley while Meta and Micron together needed over 1,200. Union locals confirmed “historic” traveler influxes: IBEW 291 and UA 296 both drew dozens of crews from across the West just to keep the project staffed.
The supposed “local jobs boom” translated into hotel bookings and RV parks, not new Idaho careers. Hundreds of itinerant workers lived in temporary housing, sending their paychecks home. The long-term impact: 100 permanent operations jobs—roughly 0.21 jobs per acre—a level of employment so thin that a single midsize retailer in Meridian employs more people.
Meta’s incentive-driven investment thus created a mirage of job creation: a transient construction surge mostly performed by outsiders, followed by a handful of technical positions. The public cost—$30–50 million in forgone sales tax—amounts to over $300,000 per job. Idaho didn’t so much build a local workforce as host a traveling one. The state subsidized an imported labor force for a company that will leave behind servers, not careers.
Idaho’s Small-Business Solution Reclaiming Growth, Revenue, and Community
After data-center disappointments, Idaho’s solution is homegrown: small businesses that hire locals, pay full taxes, and keep profits circulating where they belong—right here in Idaho.
After years of watching megaprojects devour land and subsidies while producing little real prosperity, Idaho now has a better blueprint. This small-business-first development strategy channels investment back into the state’s own people—local entrepreneurs, makers, farmers, and innovators—rather than distant corporations. Instead of one data center consuming 600 acres for 100 jobs, a network of Idaho-owned small businesses can generate dozens of jobs per acre, multiplying both wages and tax yield. The logic is simple: every locally owned business keeps profits circulating inside Idaho; every local hire raises per-capita income; every productive acre strengthens city budgets instead of draining them through tax giveaways.
The model is measurable. A 100-acre small-business park can create 1,000 Idaho jobs where a data center might create 100. A revitalized downtown block can return ten times the property tax per acre of suburban sprawl. Neighborhood hubs and rural value-added agriculture projects keep spending local, build self-reliant communities, and close the gap between urban and rural prosperity. These firms buy from Idaho suppliers, hire Idaho residents, and pay full freight in taxes—no exemptions, no foreign owners, no NDAs.
This “Buy Local” ethos isn’t sentimental—it’s economic math. Studies show locally owned businesses recirculate up to three times more money in the community than national chains, producing stronger job growth, civic engagement, and fiscal stability. Idaho’s new path proves that real economic development isn’t about courting the biggest bidder; it’s about cultivating the broadest base. The power to rebuild Idaho’s prosperity already lives here—one small business at a time.
Small Businesses First Reclaiming Idaho’s Economic Future
Other states invest in Main Street; Idaho bets on megaprojects. The numbers prove small businesses create more jobs, more tax base, and stronger communities—without secrecy or subsidies.
Across the nation, states are rediscovering what Idaho has forgotten: that real economic strength grows from the bottom up. Over the past two decades, Colorado, Wisconsin, Pennsylvania, Vermont, and North Carolina have shifted their development playbooks away from “megadeals” toward cultivating locally owned small businesses—and the results are profound. These states have proven that investing modestly in thousands of local firms yields ten times more jobs per public dollar, revitalizes downtowns, and keeps profits circulating within their own communities.
Idaho, by contrast, has staked its future on landing a handful of large corporate projects—Meta’s data center in Kuna, Gemstone Technology Park, and others—wrapped in secrecy, subsidies, and NDAs. The outcome is familiar: a few hundred permanent jobs, millions in foregone tax revenue, and utility expansions built for outsiders. Where other states are seeding hundreds of start-ups and Main Street revivals, Idaho has no Office of Small Business Advocacy, no rural small business initiative, and no state-level commitment to its own entrepreneurs.
The evidence is overwhelming: small businesses create two-thirds of new jobs, recirculate local dollars up to three times more effectively than national corporations, and build resilience across both rural and urban economies. Vermont’s “Working Lands” program, Pennsylvania’s Ben Franklin Technology Partners, and Wisconsin’s Main Street Bounceback grants all demonstrate how modest public investments generate exponential returns.
Idaho’s path forward is clear. Instead of subsidizing data centers that extract value, the state can empower local enterprise to generate it. The choice is between building dependency or building Idaho.
Inside the Idaho PUC: Whistleblowers, Transparency Failures, and the Case for a Consumer Advocate
How Idaho Became One of the Only States Without a Consumer Advocate—and What That Means for Utility Rates, Regulatory Integrity, and Public Accountability
istleblower allegations inside the Idaho Public Utilities Commission (IPUC) have ignited the most serious crisis of public confidence in the agency in decades. Two senior officials—Executive Director Maria Barratt-Riley and policy strategist Joshua Haver—were removed after repeatedly reporting what they describe as systemic misconduct: ex parte meetings with utility executives during pending rate cases, violations of Idaho’s Open Meeting Law, internal retaliation, and even an alleged quid-pro-quo offer involving a whiskey-tasting contest.
Their lawsuit paints a troubling picture of an agency that may have strayed far from its quasi-judicial mandate. Commissioners allegedly held private meetings with Avista and Veolia executives, outside public view, while major rate increases were under consideration. A senior staffer allegedly offered favorable settlement terms in exchange for help winning a social event. And when internal staff documented these problems and notified human resources, their warnings went nowhere—and they were fired.
These revelations spotlight a deeper structural issue: Idaho is one of the only states in America without an independent consumer advocate to represent residential and small-business ratepayers in utility cases. Forty-five states and D.C. have created such offices—whether as independent agencies, divisions of the Attorney General, nonprofit Citizens Utility Boards, or legislative public counsel. Idaho has none.
Instead, Idahoans are told to “show up at hearings” and trust the PUC to look out for them. But public comments cannot cross-examine utility witnesses or present expert evidence; they are not a substitute for formal representation. Without an advocate, no party in the room exists solely to protect consumers—and no one can appeal a PUC decision on their behalf.
This investigation explores what happened inside the IPUC, how regulatory capture can occur without independent oversight, and why Idaho’s lack of a consumer advocate leaves ratepayers uniquely vulnerable at a moment when utilities are seeking some of the largest increases in state history.