Chapter 9: Gateway To The Modern Day Gold Rush

Sean M. Walsh
Dec 10, 2025
8 min read

Updated: 5 days ago

New solutions to the datacenter electricity shortage require new thinking, but represent hundreds of $billions in annual revenue, and $trillions in increased equity value.

"Opportunity is missed by most people because it is dressed in overalls and looks like work." - Thomas Edison

Key Points:

1. An enormous financial opportunity, on the order of hundreds of $billions, awaits all companies willing to adopt and scale Solar Computing Clusters to solve the American electricity shortage.

2. Whoever owns this "time to power" solution will capture a large share of this opportunity.

3. The second-order benefits from this windfall are also enormous, ensuring competitive superiority for years to come.

Key Stats:

$440 Billion - midrange estimate of increased market capitalization for filling the 44GW by 2028 expected datacenter electricity shortage.

61 Companies - number of publicly traded American companies that are obvious candidates to own and scale Solar Computing Clusters.

$2.6 Billion - midrange annual revenue for datacenter colocation per Megawatt of capacity.

The Modern Day Gold Rush

In March 2025, JLL released its North America Data Center Year-End 2024 Report with a striking observation. Andy Cvengros, Executive Managing Director and Co-Lead of U.S. Data Center Markets, put it bluntly:

"It's the new 'gold rush,' as developers, occupiers, and investors are competing for available power, land, and equipment. More power generation is urgently needed if supply is going to keep up with demand."

The 1849 California Gold Rush created San Francisco, launched transcontinental railroads, and reshaped the American West. Fortunes were made—but not primarily by the miners themselves. The real winners were those who sold picks, shovels, and provisions to the prospectors. Levi Strauss didn't pan for gold; he sold durable pants to those who did.

Today's gold rush is digital, and the precious resource isn't yellow metal—it's electricity. Specifically, it's the ability to deliver power to data centers faster than competitors can secure grid connections. The companies that can solve the "time to power" problem will capture hundreds of billions in value, while those waiting in interconnection queues will watch the opportunity pass them by.

The numbers tell the story. According to JLL's year-end report, colocation vacancy in North America plummeted to a record-low 2.6% in 2024, despite several years of record construction. Absorption totaled 4.4 gigawatts—a quadruple increase since 2020. Data center rents surged 12% year-over-year, with an 11% compound annual growth rate since 2020. Tenants renewing five-year leases are experiencing sticker shock, facing up to 50% rent increases. Landlord concessions have become increasingly rare in what is now unmistakably a seller's market.

By mid-2025, JLL's updated report showed vacancy had dropped even further—to an unprecedented 2.3%. As Andrew Batson, Head of U.S. Data Center Research at JLL, observed: "Power has become the new real estate. With vacancy effectively at 0%, virtually all absorption is the result of preleasing with delivery times extending beyond 12 months."

This is not speculation about future demand. This is documented, present-day scarcity driving extraordinary pricing power and returns.

Time To Power - Unit Economics

What is a gigawatt of datacenter capacity actually worth? Let's work through the numbers using verified industry data.

According to CBRE's Global Data Center Trends Q1 2025 report, global data center pricing rose to $217.30 per kilowatt per month on a weighted inventory basis. In high-demand markets, prices are substantially higher—Northern Virginia saw a 17.6% year-over-year increase, Chicago 17.2%, and Amsterdam 18%.

Using the global average as a baseline:

Revenue per kW per month: $217.30
Revenue per kW per year: $2,608
Revenue per MW per year: $2.61 million
Revenue per GW per year: $2.61 billion

Over a typical five-year lease term, one gigawatt of data center capacity generates approximately $13 billion in revenue.

But that's for standard colocation. AI workloads command a significant premium. Industry estimates suggest AI-focused facilities can charge 2-3x standard colocation rates due to the specialized power density, cooling requirements, and mission-critical nature of AI training and inference workloads.

Using a conservative 3x AI premium:

AI Revenue per GW per year: $7.8 billion
AI Revenue per GW over 5 years: $39 billion

Now apply this to the 44 GW shortfall:

Annual colocation revenue opportunity: $115 billion
Annual AI premium revenue opportunity: $343 billion

For every year that a solution can accelerate power delivery versus the traditional 5-7 year grid interconnection timeline, the operator captures an additional year of this extraordinary revenue opportunity while competitors remain stuck in queue.

Beyond Revenue - The Equity Value Multiplier

Operating revenue is only part of the equation. Data center assets command premium valuations in capital markets, creating substantial equity value for owners.

According to Morgan Stanley's recent analysis of a number of public market disclosed colocation contracts, the equity value increase of the vendor ranged between $6.7M and $12.6M per Megawatt of datacenter capacity contracted.

Consolidating this public data outlines a profoundly exciting opportunity for public companies in the US market... up to $12 billion in market cap increase per Gigawatt of new datacenter capacity. This underscores the incentive for exploring new solutions to the electricity shortage now.

Who Gets To Swing The Pickaxe?

In the original gold rush, the smartest operators weren't panning for gold—they were selling picks and shovels. Today's equivalent is owning the technology that solves the time-to-power problem.

The potential acquirers and partners for such technology span a wide swath of the modern tech economy. Upon an initial cursory review, we identified 62 American public companies across 10 sectors that would logically benefit from owning the Solar Computing Cluster IP that accelerates data center power delivery:

AI-Native Cloud Providers: Nebius Group NV (Market Capitalization - $9.85B), CoreWeave Inc ($45.18B), Palantir Technologies Inc ($148.92B), Applied Digital Corp ($1.52B)
Bitcoin Miners Transitioning To AI: Cipher Mining Inc ($7.73B), Hut 8 ($3.51B), Riot Platforms Inc ($5.77B), MARA Holdings Inc ($4.63B), Cleanspark Inc ($3.80B), Iris Energy Inc ($2.18B), Bitfarms Ltd ($1.80B)
Datacenter Equipment Manufacturers: Vertiv Holdings ($68.20B), Hewlett Packard Enterprise ($32.68B), Broadcom Inc ($1918.65B), Arista Networks ($163.76B)
Datacenter Operators: American Tower Corp ($84.06B), Equinix Inc ($72.72B), Digital Realty Trust ($55.91B)
Defense / Government Contractors: General Dynamics Corp ($82.94B), Northrop Grumman Corp ($78.26B), Quanta Services Inc ($68.29B), Science Applications International ($6.42B), Leidos Holdings Inc ($22.47B), RTX Corp ($158.37B), Lockheed Martin Corp ($132.85B)
Electrical Equipment Manufacturers: EnerSys Inc ($5.43B), Rockwell Automation Inc ($41.26B), Enphase Energy Inc ($4.13B), GE Vernova Inc ($169.66B), Eaton Corporation ($132.74B)
Hyperscalers / Big Tech Companies: Oracle Corp ($631.54B), Apple Inc ($4095.71B), Alphabet Inc ($3826.20B), Microsoft Corp ($3656.88B), IBM ($290.22B), Alibaba Group ($258.47B), Amazon ($2436.51B), Meta ($1655.89B)
Renewable Energy / Solar Companies: Brookfield Renewable Corp ($7.05B), NRG Energy Inc ($31.96B), Clearway Energy Inc ($3.94B), Plug Power Inc ($3.08B), First Solar Inc ($27.19B), Bloom Energy Corp ($25.88B), NextEra Energy Inc ($165.86B), Tesla ($1480.55B), Shoals Technologies Group ($1.35B)
Semiconductor / Server Hardware Manufacturers: Dell Technologies ($92.64B), Nvidia ($4495.70B), Advanced Micro Devices ($360.81B), Teradyne Inc ($29.85B), Micron Technology Inc ($284.06B), Super Micro Computer Inc ($20.91B)
Utilities / Electricity Providers: The Southern Company ($94.13B), AES Corp ($9.94B), Duke Energy Corp ($89.62B), Dominion Energy Inc ($49.92B), Xcel Energy Inc ($44.79B), Exelon Corp ($44.12B), Edison International ($21.79B), Alliant Energy Corp ($16.68B), Constellation Energy Corp ($112.16B)

A Winner-Take-Most Flywheel

The financial opportunity extends beyond direct revenue capture. In AI, compute capacity creates compound advantages that reshape competitive dynamics.

Consider a simple feedback loop: More compute enables training larger, more capable models. Better models attract more users. More users generate more revenue. More revenue funds more compute. The cycle accelerates.

But, there is a much more powerful second-order impact to solving "time to power".

Consider a second feedback loop: more compute enables the attainment of AGI sooner. The attainment of AGI sooner allows the owner to apply their AGI to any and every aspect of their own business and their own competitive marketplace. They can replicate one million PhD's in computer science and product engineering, inventing and developing every imaginable product and service and competitive attack, 24 hours a day, 365 days a year. The slower-to-power competition will never catch up.

Imagine two companies:

Company A brings 10 GW online in 2026 using off-grid solar computing clusters.
Company B waits for grid interconnection until 2030.

Per Situational Awareness projections, Company A will likely achieve AGI before Company B even gets connected to the grid. In AI, where capabilities have been doubling roughly every six months, four years of compound improvement creates an insurmountable lead. Company B isn't just four years behind—it will likely never catch up.

This is a variation of a common storyline, where companies strive to win the power law dynamics market equilibrium. On the one hand, company growth with datacenter capacity can never scale like software-only networks and marketplaces. However, the fact that all marginal new computing power is producing "intelligence"... and intelligence is an unpredictably powerful resource... this author's opinion is that the winners' outcomes will indeed follow a power law distribution.

The Sovereign Imperative

Beyond commercial considerations, as I wrote in a previous chapter, AI infrastructure has become a matter of national security. The federal government has taken notice.

The FY2025 National Defense Authorization Act established the AI Rapid Capabilities Cell with $100 million in initial funding. The FY2026 NDAA mandates an "Artificial Intelligence Futures Steering Committee" to examine the military implications of advanced AI, including artificial general intelligence.

The Department of Energy issued a Request for Information in April 2025 on developing AI infrastructure on federal lands, identifying 16 potential sites. Los Alamos National Laboratory is planning for 70 MW of capacity by 2027 and 160 MW by the early 2030s.

But, if true, these are rookie numbers, likely to be increased by orders of magnitude in coming months/years.

Off-grid, domestically powered compute offers unique advantages for sensitive applications:

Immune to grid vulnerabilities — and Chinese transformer backdoor risks
No interconnection delays — immediate strategic capability
Enormous, unconstrained supply chain — from photons to processors
Geographic flexibility — hardened, distributed deployment

If the United States is destined for a 21st-century AI rendition of World War 2's Manhattan Project, which I believe it is, capital equipment budgets will be measured as percentages of American GDP, not in dollars. When the USA really mobilizes, it can achieve remarkable feats.

Let's Review

The previous eight chapters have documented a crisis. This chapter reveals the opportunity embedded within it.

A 44 GW shortfall by 2028. Interconnection queues stretching 4-7 years. Chinese backdoor hardware in 500+ critical transformers. Power has become "the new real estate."

Demand is practically infinite. $6.7 trillion in investment required by 2030. Vacancy below 2%—effectively zero. Rents surging 12% annually with 50% sticker shock for renewals.

The capital exists. $500 billion for Stargate. $400+ billion in hyperscaler capex. Silicon Valley rumors of a $1 trillion dollar AI cluster.

The only constraint is time-to-power. Traditional pathways take 4-10 years. Every conventional solution is bottlenecked.

Solar Computing Clusters bypass nearly all problems:

No transformer crisis — off-grid architecture
No turbine shortage — solar photovoltaic generation
No interconnection queue — self-contained power systems
No grid instability - off-grid computing clusters
No high-voltage electrician shortage — modular, standardized installation
No NIMBY opposition — remote desert deployment

Solar Computing Clusters exist. They've operated successfully since 2022. The IP is protected. The math is compelling. The strategic advantages are undeniable. The AI infrastructure gold rush will last for decades. The question is not whether someone will capture this opportunity...

The question is whose flag will get planted on this new continent first.

Learn more about my work in the next chapters, and at www.639solar.com.

Sources:

North American Colocation Datacenter Vacancy Rates By Year, VoronoiApp
2025 Global Datacenter Outlook, JLL
Power, Pricing, And Pipeline: CBRE Report 2025, DataXConnect
Beyond The Bubble: Why We Think AI Infrastructure Will Compound Long After The Hype, KKR
How To Invest In Datacenters: AI Infrastructure Profits, Lean Research
The Cost Of Compute: A $7 Trillion Race, McKinsey
Breaking Barriers To Datacenter Growth, BCG
Datacenter Market Size Report 2030, Grand View Research
25 AI Datacenter Statistics & Trends 2025, TheNetworkInstallers