Interconnection queue converts hyperscaler capex into generator rent
gen-interconnection-queue · conviction — · status open · horizon — · as of 2026-08-10
How to read the numbers on this page
A range instead of a point. 903 of 1,848 supply weights are not disclosed by anyone. Where redrawing them across their plausible range moves a figure by more than 25%, the figure is shown as a range and marked. A tight number is ground you can stand on; a wide one is not.
Not every premise is scored. A premise citing something the model verifies on every rebuild — a filed figure, a graph edge, a computed cell — is a PRECONDITION, not a risk. It gates the conclusion but contributes no uncertainty, because charging a conclusion for being verifiable made well-evidenced arguments look weaker than vague ones.
Composed two ways. Where several premises gate a conclusion, the figure is given as "X% if independent, Y% if they move together". They are claims about one industry, so the truth is between and nobody can say where. Treat it as an ordering device, not a calibrated probability.
Dated. Each figure is computed from facts, and the page states the age of the oldest one beneath it. The full arithmetic runs from a published model config to company revenue exposure.
Exhibits
Series available as data/gen-interconnection-queue.csv
Green/blue = model marks it as CAPTURING the rent (unbound and supplies the scarce good); faded = PAYING it (bound severe or moderate). Operating margin, live.
The variant
Consensus
The 2,600 GW interconnection queue is a datacenter problem: it delays new AI capacity and forces hyperscalers to pay more for power. The constraint binds the hyperscalers and hurts their economics, so investors discount datacenter infrastructure names and wait for grid relief.
Variant
The queue is a transfer mechanism, not a universal brake. It splits the datacenter layer: hyperscalers and loss-making accelerator renters pay rising power costs into finite supply, while generators with sited capacity collect rising rents without adding load to the queue. Vistra and Talen convert operating leverage into duration; CoreWeave converts revenue growth into negative margins. The constraint binds new entrants and rewards incumbents with existing grid connections.
Differentiator
Consensus reads the queue as a symmetric cost shock. The supply chain reveals asymmetry: generators sit upstream of the constraint, hyperscalers downstream. Five-year waits do not slow Vistra; they raise what it charges per MWh from already-connected plants. The earnings analyst sees power as an input cost; the constraint view sees it as a tollgate that incumbents own.
Falsifiers
- claim: Vistra operating margin expands as queue lengthens · criterion: Vistra operating margin >22% in any of next four quarters while median queue wait >5 years · horizon: 2027-08-10 · settles: confirmed
- claim: CoreWeave reaches operating margin breakeven without reducing growth · criterion: CoreWeave operating margin >0% while revenue growth >100% YoY · horizon: 2027-02-28 · settles: refuted
- claim: Hyperscaler capex pivots to co-location deals that bypass the queue · criterion: Majority of new hyperscaler datacenter capacity announced in 2027 H1 is co-located with existing generation · horizon: 2027-06-30 · settles: refuted
Open questions
- What share of Vistra and Talen revenue is contracted vs. spot, and on what duration? Long-term contracts lock in pre-queue pricing and miss the rent.
- Do hyperscalers have disclosed queue positions per region, or do we infer exposure only from announced builds? Per-plant data would sharpen the timing.
- How quickly can generators repurpose retired coal plant grid connections for datacenter co-location? Reuse bypasses the queue and changes the competitive set.
Reasoning chain
Vistra Corp.0.95 strongPublic company, filed financialsVistra Corp. — operating margin 19.7%1.00 strongRecorded Q1 2024 figureGrid interconnection queue position supplies Vistra Corp.1.00 strongSiting model infers exposure; not disclosed per-plantGrid interconnection queue position0.92 strongMedian wait is reported aggregate; withdrawal rate unknown
Composed 0.92 via and over 2 gating premises · 2 supporting premises shown but not multiplied in — citing a filed figure should not cost conviction · Weakest link Grid interconnection queue position at 0.92
Generator sits upstream of the constraint: queue delays buyers, not sellers. Operating margin sustains or expands as scarcity raises contract prices.
CoreWeave, Inc.0.93 strongPrivate, cap estimated from disclosed roundsCoreWeave, Inc. — operating margin -2.1%1.00 strongReported operating marginCoreWeave, Inc. — revenue growth $2001.00 strongDisclosed growth rateGrid interconnection queue position supplies CoreWeave, Inc.1.00 strongNew datacenter additions face queue; existing capacity timing unknown
3 supporting premises shown but not multiplied in — citing a filed figure should not cost conviction · Weakest link Grid interconnection queue position supplies CoreWeave, Inc. at 1.00
Revenue scales 3x while margins stay negative. Each incremental datacenter build enters the queue; power is a rising variable cost with no negotiating leverage.
Hyperscaler0.97 strongEstablished category, multiple public membersGrid interconnection queue position supplies Hyperscaler1.00 strongNew datacenter regions face queue; co-location deals bypass itGrid interconnection queue position0.92 strongQueue applies to new connections; does not gate existing sited capacity
Composed 0.92 via and over 2 gating premises · 1 supporting premise shown but not multiplied in — citing a filed figure should not cost conviction · Weakest link Grid interconnection queue position at 0.92
Hyperscalers pay rising power costs into cloud pricing with 18-month lag. Unlike generators, they cannot mark contract prices to scarcity in real time.
Sources
- US data-centre peak-demand share 4.1% (2025) -> 8.5% (2027); ~125 GW of US load added — link acc 2026-07-24
- TTM revenue $19.45B, gross margin 38.64%, operating margin 19.71%, market cap $54.03B, forward P/E 17.44, TTM capex ~$2.87B — link acc 2026-07-21
- TTM revenue growth ~7.41% YoY (twelve months ended 2026-03-31) — link acc 2026-07-21
- ~2,600 GW interconnection queue (early 2026), 5+ year median waits (LBNL); ERCOT large-load applications — link acc 2026-07-24
- TTM revenue, gross/operating margin, market cap, capex — link acc 2026-07-21
- Q1 2026 earnings, capex and debt focus — link acc 2026-07-21
Write-up
Pre-filled skeleton: gen-interconnection-queue.md