NUAI’s PPA and the Spark Spread: Who Captures the Value?

Thesis

NUAI’s power purchase agreement can make its Texas campus more valuable by giving prospective tenants a credible route to electricity. The investment question is how much of that advantage reaches shareholders after power charges, development spending and financing costs.

A spark spread measures the gap between electricity prices and the gas required to generate that electricity. NUAI’s Phase 1 subsidiary is the purchaser under its disclosed agreement. Its economic exposure therefore depends on two contracts: what it must pay the power supplier, and what a tenant must pay or reimburse it.

The strongest supported interpretation is that contracted supply can improve the project’s leasing proposition. Public disclosures do not establish a fixed electricity price or an entitlement to collect a merchant generator’s spread. NUAI has separately described an intended structure that passes power costs to tenants. Whether the eventual agreements achieve that protection remains a contractual question.

Model AssumptionEvery market price, heat rate, utilization rate and hypothetical tariff in the five charts below is an analytical assumption. The charts contain no live prices, forward-market curves or disclosed NUAI tariff estimates.

What the PPA establishes

The September 21 filing identifies a definitive agreement executed three days earlier. Its principal supply terms are:

NUAI Form 8-K — September 21, 2026

The announcement describes firm, contracted supply and expected availability in Q3 2027. That expected delivery window is company guidance; the December conditions deadline serves a different purpose. Neither means delivery has already begun. Firm supply also does not establish fixed pricing.

Company announcement — September 21, 2026

The public PPA summary leaves its pricing formula, escalators, minimum charges and resale rights unspecified. Those omissions prevent a reliable allocation of commodity exposure.

The alternative supply provisions also matter: this is not evidence that NUAI owns the Odessa plant or that its campus will always consume electricity generated exclusively behind its own meter.

Three different spreads

The merchant spark spread is electricity revenue less the corresponding gas cost. Heat rate measures fuel used per unit of electricity; a higher rate means more gas consumption. The standard calculation excludes other operating and financing costs, so the result is not net profit.

Spark spread ($/MWh) = electricity price − heat rate × gas price.

EIA — Spark Spread

The regional gas basis compares prices at different gas delivery points. Waha in West Texas can trade below Henry Hub when gas production exceeds available transport capacity. EIA’s historical discussion explains how additional pipeline capacity can narrow that discount. This mechanism does not establish today’s discount or a future price path.

EIA — Regional gas price differentials

The campus margin retained by NUAI is a separate result: tenant receipts less the project’s obligations and costs. A regional gas discount may improve a supplier’s economics without being fully transferred into the PPA. A favorable PPA may reduce a tenant’s bill without producing an electricity markup for the landlord.

An illustrative decomposition makes the distinction concrete. Assume benchmark power at $80/MWh, Henry Hub gas at $4/MMBtu, local gas at $3/MMBtu and a heat rate of 7 MMBtu/MWh. Add a hypothetical nonfuel charge of $29/MWh.

Figure 1. Original illustrative calculation. The $29/MWh charge is a simplifying assumption, not a disclosed tariff or independently estimated cost.

Benchmark spark spread is $52/MWh: $80 − 7 × $4. The regional fuel benefit adds $7/MWh: 7 × ($4 − $3). Subtracting the assumed $29 charge leaves a $30/MWh benchmark cost difference. The hypothetical purchase price is $50/MWh: $29 + 7 × $3.

This identity helps explain a possible advantage; it does not allocate it. Nor does it establish a fully delivered saving. Taxes, delivery charges, reliability, volume, settlement location and other services must match before two power offers are commercially comparable.

The pricing formula determines the exposure

A twenty-year term tells investors how long a relationship can last. The tariff tells them how the bill changes.

These are conditional structures, not identified NUAI terms. Consider two simplified purchase arrangements: a fixed $50/MWh tariff and an indexed tariff equal to $29/MWh + 7 × local gas price.

Figure 2. A flat hypothetical fixed tariff compared with a gas-price sensitivity. The horizontal axis is a scenario input, not a timeline or forward curve.

Both arrangements cost $50/MWh at $3 gas. At $5 gas, the indexed arrangement costs $64/MWh while the fixed arrangement remains $50. Below $3 gas, the indexed arrangement becomes cheaper.

Fixing a price can protect against increases while surrendering some benefit from decreases. Gas indexation can leave fuel exposure while protecting against expansion in the supplier’s conversion margin. Neither structure is universally preferable: the initial charge, exclusions and tenant recovery all matter.

If electricity market prices rise, a favorable contract could become more competitive. Turning that advantage into resale income would require rights that cannot be presumed. Its practical value may instead appear in the rent or lease terms a tenant is willing to accept.

Waha convergence can offset a wider benchmark spread

Suppose the Henry Hub referenced spark spread increases. That alone does not prove that a campus buying local gas-linked electricity becomes more competitive. Local gas could rise faster than Henry Hub as the regional discount disappears.

The next example holds the illustrative heat rate and $29 nonfuel charge constant. It changes both commodity prices and the regional basis.

Figure 3. Separate hypothetical states, not forecasts. The indexed purchase formula remains $29 + 7 × Waha throughout.

The benchmark spark increases by $13/MWh, but the regional fuel benefit falls by $21/MWh. The resulting advantage shrinks by $8/MWh. Indeed, the spark calculated using local gas declines, despite the increase in the Henry Hub referenced measure.

This is a useful counterexample to a broad “power rises faster than gas” thesis. The relevant gas is the index or delivered fuel actually governing the transaction. The relevant electricity benchmark is the comparable product at the relevant location and hours.

A campus can remain attractive even after its gas discount narrows. Timely availability, suitable capacity and acceptable service terms could outweigh part of the higher energy bill. That is a commercial proposition to test against a tenant’s alternatives, not a reason to assume permanently discounted gas.

Tenant recovery determines the landlord’s exposure

NUAI’s annual report describes intended leases in which electricity consumption is passed through to tenants or paid by them directly. This is affirmative evidence of the proposed business model. It is not proof that a signed tenant agreement already covers every obligation under the PPA.

NUAI 2025 Form 10-K — filed March 12, 2026

For scale, assume consumption equivalent to 207 MW at 90% utilization across a non-leap year. Annual energy is 207 × 8,760 × 90%, or 1,631,988 MWh. At the illustrative heat rate of 7, a $1/MMBtu gas increase produces a $7/MWh fuel-cost equivalent, equal to 0.7 cents/kWh.

Figure 4. Original physical-cost sensitivity: approximately $11.42 million annually per $1/MMBtu. This is not a forecast of NUAI’s PPA expense or earnings.

The 207 MW is a supply reference, not 207 MW of verified IT equipment load. No extra cooling multiplier is applied to energy already measured at the assumed campus supply point. Actual contracted billing could also include minimum charges that do not fall proportionately with utilization.

If that entire hypothetical $11.42 million increase reaches the project’s bill, the operating impact depends on reimbursement:

Figure 5. Full recovery leaves no unrecovered commodity increment; 75% recovery leaves approximately $2.86 million; no recovery leaves approximately $11.42 million. These percentages are assumptions.

Complete recovery can protect the landlord’s operating margin while exposing the tenant to a higher cost of computation. It does not necessarily eliminate cash-timing risk. NUAI could owe its supplier before collecting the corresponding payment, and tenant default could leave obligations unmatched.

Conversely, cheaper electricity may benefit NUAI through stronger lease demand or better negotiated rent. Capturing that benefit requires evidence. Adding the full electricity saving to landlord earnings while also assuming pass-through to the tenant would count the same advantage twice.

Contracted power has a capital cost

The October 8 filing reports a $116 million letter of credit supporting the PPA, secured by approximately $118.3 million of initial cash collateral. Required collateral was 102% of the undrawn amount. It also identifies a 1% issuance fronting fee and a 2% annual LC fee.

NUAI Form 8-K — October 8, 2026

On that face amount, the stated rates correspond to $1.16 million initially and $2.32 million annually, before separate financing costs. The collateral is restricted liquidity supporting performance; it is not an electricity prepayment or automatically a permanent expense. Its economic burden depends partly on funding cost, permitted earnings and release timing.

The September prospectus also describes a 5% non-voting, fully diluted interest for Vistra in the relevant project ownership structure following power delivery. This is project-level participation, not a disclosed issuance equal to 5% of NUAI’s parent shares. The same prospectus describes tenant negotiations and contingent commercial terms.

NUAI Form 424B5 — September 24, 2026

These obligations explain why an attractive electricity arrangement cannot be valued from its energy discount alone. The relevant comparison includes capital committed, ownership retained and the dates on which shareholders can receive cash.

The strongest opposing interpretations

The constructive case is that supply control matters more than commodity arbitrage. A tenant facing expensive delays could accept commercially attractive rent for a suitable powered site, even if the electricity price simply covers the supplier’s costs and return. A favorable allocation of power risks could further support financing. Neither outcome requires NUAI to trade electricity.

The opposing case is that the supplier and tenant capture most of the benefit. The supplier could receive adequate tariff compensation and project participation while the tenant demands low rent or extensive guarantees. NUAI would then retain construction and capital exposure without a large commodity margin.

A third possibility is that the tariff is attractive but the lease and PPA do not match. Different commencement dates, volume commitments, escalation rules or interruption remedies could create exposure that a headline pass-through description conceals.

These are competing contractual outcomes, not probability-weighted forecasts. Strong supply terms support the project, but the price paid for those terms and their translation into tenant cash flow remain central.

Evidence that would change the assessment

The most consequential next evidence is a detailed tariff and an executed tenant arrangement. The review should establish:

1. Indexation: fixed amounts, gas or electricity benchmarks, heat-rate provisions and escalation.
2. Scope of recovery: which capacity, energy, transport, balancing and replacement-power charges the tenant bears.
3. Timing and volume: minimum commitments, ramp schedules and gaps between payment obligations.
4. Control: dispatch, resale and curtailment rights, including who receives any related proceeds.
5. Capital: collateral release, infrastructure funding and ownership participation through the project structure.

A favorable answer could justify more confidence in value retained from the power arrangement. Adverse mismatches could weaken that conclusion even if wholesale spark spreads remain wide. The NUAI TCDC Site Economics Model provides the related financial framework; these illustrative power calculations do not revise its valuation assumptions.

Methodology and source scope

Evidence was reviewed as of October 19, 2026. The seven primary sources linked alongside the discussion establish the contract summaries, intended commercial structure and market definitions. They do not supply a disclosed NUAI tariff. EIA’s older explanatory pages support definitions only.

All five visuals are original calculations. Their shared heat rate is 7 MMBtu/MWh; their hypothetical nonfuel charge is $29/MWh. Except for the explicitly varied scenario inputs, prices and assumptions are held constant. There are no forecast probabilities, market quotations or forward curves. Hourly dispatch, taxes, financing, outage costs and contractual billing exceptions are not modeled.

Readers can inspect the chart inputs and calculations in JSON or download the CSV.