Emergent Archetypes
Christine Primmer, a Senior Advisor with Clarum Advisors, is an investor/operator with diverse expertise helping venture-backed technology organizations build scale and resilience. She has closely advised, operated, and partnered with twenty-plus venture-backed tech companies in energy, transportation, smart buildings, and IOT, ranging from Series A to E, helping them grow revenue and scale operations using data driven strategies, partnerships/M&A, and capital.
As a Director with Clarum Advisors, Molly Podolefsky leverages her experience in economics, decarbonization, corporate sustainability, corporate finance, and the power and utilities sector to help innovative, clean tech companies.
In Part One, we traced how the utility model — founded to deliver safe, reliable, and affordable power under public oversight — is being challenged by the pace and scale of demand created by AI data centers. We also looked at other recent examples where business-model innovation solved demand-supply imbalances.
We are seeing another business model evolution happen in real time now. AI-driven load growth is increasing faster than traditional planning and regulatory cycles can accommodate. Moreover, this load growth looks different: AI load density is reshaping how, where, and by whom power is planned for, financed, and delivered.
Utilities remain the anchors of safety and reliability, but the ecosystem forming around them looks increasingly different — an echo of parallel inflection points such as the rise of TowerCos (such as American Tower) and independent power producers (IPPs) explored in Part One of this series. When incumbents reach structural limits, new archetypes emerge to bridge the gap.
It’s already happening. The question is: Will these archetypes complement utilities, or compete with them? So far, the evidence is mixed.
The New Paradigm
AI datacenter development now faces bottlenecks at nearly every stage of the power value chain:
Christine Primmer: Utilities remain the anchors of safety and reliability, but the ecosystem forming around them looks increasingly different – an echo of parallel inflection points such as the rise of TowerCos and independent power producers. When incumbents reach structural limits, new archetypes emerge to bridge the gap.
Generation — Limited available capacity in key regions;
Transmission and Distribution — Long lead times for upgrades, interconnection delays, and congestion on critical corridors;
Siting and Permitting — Zoning, environmental, and water constraints stretching timelines to five years or more;
Supply Chain — Long-lead equipment shortages — especially transformers and switchgear — have become binding limits on grid expansion, rivaling permitting delays; and
Molly Podolefsky: The AI energy wave tests more than grid capacity – it tests the covenant that underpins the utility model itself. The founding premise of affordability, reliability, and safety still matters. But meeting it now requires new partners, new contracts, and new governance models that share risk at AI speed.
Behind the Meter — Developers building onsite or adjacent resources — mixing renewables, storage, and thermal — to bypass the queue entirely.
The result is a decentralization of problem solving. Innovation is migrating from centralized planning rooms to distributed partnerships, wherever capital and capability can move fastest. Once again, specialization is driving speed.
Emerging Partnership Archetypes
None of these models replace utilities. Each reflects a different way of sharing the load — financially, operationally, and politically. In the AI era, as load becomes more concentrated, faster moving, and more siting constrained, four partnership archetypes have emerged.
Archetype 1 — Hyperscaler Led Capacity Partnerships.
Definition: Hyperscalers initiate or underwrite new generation, storage, or transmission specifically to serve their datacenter campuses. These deals are driven by load rather than corporate PPAs.
Examples:
Google + NV Energy Clean Transition Tariff — Created a fast-track procurement option for large customers — beginning with Google — to enable new clean-energy capacity aligned with data-center load growth.
Amazon (AWS) + Dominion Energy — Dominion’s recent IRPs identify hyperscale data-center growth — led by AWS’s Northern Virginia campuses — as the region’s largest incremental load driver.
Meta + TVA — TVA procured three hundred seventy-seven megawatts of new solar tied directly to Meta’s datacenter growth in Gallatin and Huntsville.
Archetype 2 — Infrastructure Capital + Developer Platforms.
Definition: The “TowerCo” of this era. Developer-capital partnerships that build replicable, near-load generation platforms targeted at hyperscale clusters. These platforms specialize in fast siting, flexible capital, and the ability to take merchant or hybrid risk.
Examples:
Cloverleaf Infrastructure — Private-capital-backed developer planning hundreds of megawatt-scale flexible plants at data-center corridors in PJM and ERCOT.
Calpine + Infrastructure Capital Partners — A private-capital-backed flexible-generation platform growth, anchored initially by a one hundred ninety megawatt ERCOT campus colocated with Calpine’s Thad Hill Energy Center.
Archetype 3 — Utility-Anchored Joint Ventures & Co-Development Models.
Definition: Utilities maintain governance and operational control while integrating hyperscaler capital, third-party investment, or co-ownership to accelerate delivery of load-serving capacity.
Examples:
Meta + Entergy — Meta pre-funds utility-owned generation dedicated to data-center expansion.
PPL + Blackstone JV — Joint venture to develop dedicated gas-fired generation for hyperscale load.
Oncor + Private Capital Partnerships — Oncor collaborates with private-capital-backed transmission developers to accelerate high-voltage infrastructure supporting rapid data-center expansion in Texas load pockets.
Archetype 4 — Dedicated Virtual Power Plant (VPP) + Developer Partnerships
An adjacent, demand-side model:
Definition: A handful of VPP providers have begun to offer specialized VPP models to provide hyperscalers and data center developers, often in partnership with utilities, with localized dispatchable capacity by aggregating DERs, including thermostat-controlled HVAC loads, solar + battery storage and commercial and industrial flexible capacity. While not a substitute for large-scale supply, these platforms can accelerate speed-to-power by easing local constraints.
Example:
Voltus + Cloverleaf Infrastructure — VPP provider Voltus and developer Cloverleaf partner on a Bring Your Own Capacity (BYOC) VPP model enabling data centers to bring their own capacity stack to utilities, bypassing capacity constraints and increasing speed to power.
A first-of-its-kind model, Cloverleaf and Voltus unveiled this initiative only three months ago. If proven successful, the model is likely to be emulated by others as hyperscalers seek scalable solutions to utility capacity constraints.
Enablers of Success
Past infrastructure leaps were fueled not just by new technology but by new structures: specialization, risk-sharing, and partnership. The same playbook applies today, but the AI era demands faster moves and more flexible models.
Specialization at Scale — Developers, infra funds, and hyperscalers each focus on a distinct part of the value chain. Like TowerCos and IPPs before them, specialists can turn grid constraints into opportunities. Furthermore, energy specialization will be critical to ensuring the delivery of these projects at scale.
Risk-Sharing Models — The next wave of capacity won’t fit cleanly into rate recovery or merchant markets. Hybrid models — capacity leasing, reliability-as-a-service, prefunded JVs — spread financial exposure across utilities, investors, and corporates.
Collaborative Platforms — Partnerships are evolving from one-off PPAs to broader ecosystems that align utilities, hyperscalers, and regulators on siting, transmission, and interconnection.
Institutional Imagination — Meeting AI-driven demand safely and affordably means blurring old categories: rate-based vs. merchant, front- vs. behind-the-meter. Regulators and utilities must co-design governance that allows speed without losing oversight.
Closing Thought
The AI energy wave tests more than grid capacity — it tests the covenant that underpins the utility model itself. The founding premise of affordability, reliability, and safety still matters. But meeting it now requires new partners, new contracts, and new governance models that share risk at AI speed. Utilities have the opportunity to help shape this transition by embracing new financial and business models that benefit the entire ecosystem.
