Layer 1 · Where the power comes from
Grid interconnection
Everyone, eventually. Every other option on this list is either a bridge to this or a way around it.
The shape of it
- Lead time
- 4-5 years typical, approaching 5 on average
- Capital cost
- Utility-funded upgrades billed back; site cost varies wildly by region
- Carbon
- Whatever the local grid mix is
- Maturity
- proven
Verified 2026-09-06.
The catch
The US queue exceeds 2,600 GW and roughly 80% of projects withdraw before energising. A queue position is not a power supply, and treating it as one is the single most common planning error in this industry right now.
What it actually decides
If your interconnection date is 2031, your 2026 chip decision is really a 2031 chip decision. Plan the silicon around the power date, never the reverse.
What this is powering
A GB200 NVL72 rack draws 120–132 kW and a GB300 pushes 135–200 kW, against a 2026 average rack of about 27 kW. The silicon decision and the power decision are the same decision, made eighteen months apart.
Layer 2 — Silicon · GB200 NVL72 · The 150 W inference option
Weighed against
- Small modular reactors (SMR) vs Grid interconnection — Two answers to the same question, both arriving in the 2030s.
- Behind-the-meter generation vs Grid interconnection — Skip the queue and take the regulatory risk, or wait five years and take none.
Others in where the power comes from
- Nuclear PPA (existing plants) — 1-3 years to contract; the plant already exists
- Small modular reactors (SMR) — 2030s for meaningful capacity
- On-site natural gas turbines — 12-24 months, the fastest firm power available
- Behind-the-meter generation — Depends on the generation, but it skips the queue
- Solar + battery storage — 18-36 months
- Wind PPA — 24-48 months