Question 1. Is the target vendor a contracted-power provider, an enabling supplier or an emerging generation platform?

The three tiers have different underwriting disciplines. Do not apply a Tier 1 lens to a Tier 2 name (or vice versa).

Question 2. Which firmness bucket does the vendor sit in?

Existing nuclear, SMR, gas turbine, BTM dispatchable or 24/7 renewables. Each bucket has its own price ceiling, own binding constraint and own falsification trigger.

Question 3. Is the binding constraint at this vendor a regulatory clock, a supply-chain clock, or an allocation queue?

Regulatory clocks (NRC, uprate approvals) are opaque and long. Supply chain clocks (HALEU, gas-turbine forgings, LFP cell) have public visibility and shorter cycles. Allocation queues (gas turbine reservations, ISO queue positions) can be tracked deal by deal. Underwriting quality of the constraint is the underwriting quality of the whole thesis.

Question 4. What is the counterparty structure with the hyperscaler?

Direct PPA, offtake framework agreement, tolling agreement, equity investment plus offtake, utility rate-tariff structure. Each carries different termination optionality, price-adjustment mechanics and counterparty concentration. Structure dominates headline price in the risk-adjusted return.

Question 5. Where does this vendor sit in a hyperscaler's sequencing plan?

Sequencing beats archetype selection (see Section 05). A vendor whose offering plugs into a hyperscaler's near-term bridge (BTM dispatchable), medium-term firm (gas turbine, existing nuclear) or long-term base (SMR, 24/7 CFE stack) has a specific place in a portfolio strategy. Vendors that can bridge across sequencing horizons (portfolio developers, incumbent platforms) capture disproportionate value.

Question 6. What is the P1 plus P2 exposure at the specific delivery point?

Contracted firm power that cannot be delivered is unpriced. Interconnection queue position, transmission-line capacity and GIA status at the specific node all matter. A P1 PPA at a node with a cleared GIA is a fundamentally different exposure from a P1 PPA at a node with a multi-year queue.

Grid-to-Chip P1 · Electrons: firm generation for the AI campus

27 August 2026 · 59 min read · Updated 6 Oct 2026

Grid-to-Chip P1 Electrons feature: five firmness archetypes as vermillion price-spectrum ladder with three-tier vendor taxonomy footer — The five firmness archetypes on the AI-campus P1 spectrum.

More on

The intermediate bus battleground: IP litigation and the possible futures for 48V-to-12V

Point-of-load: the hidden AI buildout winner

Copper hits the wall in AI racks: the migration to optical, CPO and silicon photonics