NUAI: TCDC Site Economics and Tenant Scenarios

Thesis

A current preliminary TCDC model, mapped to the live Google Drive workbook, showing why nameplate capacity, tenant obligations, project financing, ownership, promote economics, and dilution must be modeled separately.

Key chart

The argument

TCDC may become valuable under several commercial structures: powered land, powered shell, build-to-suit lease, joint venture, customer-financed infrastructure, or a hybrid data-center plus power-SPV structure. Each structure allocates capex, operating cost, residual value, parent guarantees, and sponsor economics differently.

The model therefore avoids treating all nameplate MW as immediately distributable equity value. It keeps the following items separate:

• nameplate MW;
• critical IT MW;
• data-center sponsor value;
• power-SPV optionality;
• project debt and equity requirement;
• parent-company bridge items;
• direct ownership;
• promote economics; and
• diluted share count.

Key findings

• The model now treats Phase 2 behind-the-meter power as a separate power-SPV path, not as full generation capex charged to NUAI parent.
• Data-center capex, tenant/direct funding, project debt, DSCR, sponsor ownership, cash contribution, promote, and dilution are separately visible.
• Shareholder value is not gross campus value. It is the value attributable to NUAI after ownership, capital requirements, financing, and dilution.
• A signed Phase 1 lease would matter because it could de-risk later phases before those phases are fully built.
• The largest unresolved variables are ownership/waterfall economics, lease terms, power documents, financing draw conditions, and diluted share count.

Counterarguments

A more optimistic view is that scarce power-adjacent campus capacity could be strategically valuable before a lease is fully disclosed. A hyperscale or frontier AI tenant may prefer to reserve the whole campus to avoid fragmented future capacity, and that could raise option value for Phase 2 and Phase 3.

A more cautious view is that a large campus reservation can be conditional, low-economics, or capital-intensive for the sponsor. Without disclosed rent, remedies, credit support, construction obligations, power tariffs, and distribution waterfalls, gross MW can overstate equity value.

The current model handles this by showing Phase 1 signed and full-execution states separately rather than blending them into one headline valuation.

Risks

Power and generation

Company GuidanceNew Era has announced a 450 MW behind-the-meter generation plan for TCDC involving Thunderhead Energy and Turbine-X. The model uses that information to separate power-SPV value from data-center value. It does not assume final power economics, fuel supply, emissions approvals, definitive equipment purchase documentation, or retained power-SPV ownership are complete.

Tenant concentration and lease economics

A single large tenant can improve financing visibility but also concentrates bargaining power, renewal risk, and construction-risk allocation. Comparable AI lease economics help frame potential rent ranges, but they cannot establish NUAI-specific rent, tenant funding, pass-throughs, credit support, or remedies.

Capital formation

Project debt may be available only after lease, permit, power, and construction milestones are satisfied. The workbook sizes debt with loan-to-cost and DSCR constraints rather than assuming every planned dollar can be drawn immediately.

Ownership and waterfall uncertainty

The direct NUAI data-center share, cash contribution, power-SPV share, development fee, management fee, and promote assumptions are explicit. They are not final disclosed economics. A weaker sponsor share or heavier parent guarantee would lower per-share value even if project enterprise value remains attractive.

Dilution

The model uses a scenario share count of 123 million diluted shares in the base case. Per-share value must be rechecked against current basic shares, warrants, options, RSUs, converts, ATM issuance, equity facilities, and any financing dilution.

Catalysts

The highest-value evidence would be:

• an executed lease or binding customer commitment;
• a lease summary with rent, escalators, pass-throughs, tenant funding, credit support, and remedies;
• final data-center ownership and GP/LP waterfall terms;
• binding project-financing commitments and draw conditions;
• power, turbine, gas, interconnection, and emissions documentation;
• permits and notice-to-proceed milestones;
• construction progress and delivery schedule; and
• updated share-count and financing disclosures.

Scenario analysis

The linked model now emphasizes development state rather than one headline output.

The wider bear and bull outputs remain available in the spreadsheet. They are scenario outputs, not predictions or recommendations.

Valuation analysis

The workbook uses phase-level project economics and then bridges to sponsor value. The core structure is:

Project cash flow and project equity value → direct sponsor share → development and management fees → optional promote → power-SPV share → HoldCo bridge → diluted shares.

The current base model uses:

• 200 MW / 450 MW / 750 MW nameplate phases;
• 190 MW / 427.5 MW / 712.5 MW critical IT phases;
• $1.80 million of starting lease revenue per critical MW-year;
• 2.25% annual lease escalation;
• $15.0 million of gross data-center capex per critical MW;
• 25% tenant/direct capex funding;
• 75% max debt to funded cost;
• 1.40x minimum initial DSCR;
• 8.25% debt coupon;
• 11.0% equity discount rate; and
• 123 million diluted shares.

Model AssumptionThe current-platform base value is probability-weighted using 55% for Phase 1, 45% for Phase 2, and 25% for Phase 3. The Phase 1 signed and full-execution states change those probability gates rather than rewriting the entire campus as already stabilized.

Assumption evidence map

The NUAI workbook now includes a dedicated “10 Assumption Evidence” tab. That tab links material inputs to their evidence category, model location, confidence level, unresolved items, and intended website source IDs.

The highest-confidence items are the disclosed phase plan and source-backed description of the 450 MW BTM power pathway. The lowest-confidence but most valuation-sensitive items are direct economic ownership, cash contribution, power-SPV retained share, promote terms, debt drawability, and diluted shares.

Methodology

Every phase is evaluated through six gates:

1. site and property control;
2. generation and/or grid rights;
3. permits and environmental compliance;
4. customer obligations;
5. financing commitments; and
6. construction and equipment progress.

Probability adjustment captures whether the phase occurs. Discounting captures timing and required return. The model keeps these separate so that a delayed project is not accidentally double-counted or under-penalized.

Disconfirming evidence

The base case weakens if:

• a lease is delayed or contains weaker economics than comparable AI infrastructure deals;
• project debt requires more parent support than modeled;
• common infrastructure costs are materially higher;
• Phase 2 power requires parent-level capital or guarantees beyond the current assumption;
• power/fuel/emissions permitting constrains delivery;
• NUAI retains less direct economic ownership than modeled;
• promote economics are unavailable or junior to a high LP hurdle; or
• diluted shares rise materially above the current base denominator.

What would change the conclusion

The conclusion would become more constructive if NUAI discloses a binding Phase 1 lease, final customer-credit support, lower parent equity requirements, a more favorable waterfall, or evidence that later phases benefit from already-funded shared infrastructure.

The conclusion would become more cautious if the project requires more parent equity, less direct ownership, longer timing, weaker pass-throughs, lower lease revenue, higher capex, higher debt cost, or materially larger dilution.

Primary sources

The current NUAI source archive and the model-specific source map now include:

• the March 31, 2026 Form 10-Q for filed financial statements, liquidity, obligations, equity, and share disclosures;
• the May 18, 2026 SEC-filed business update for the TCDC phase plan, power pathway, proposed GP/LP structure, capitalization, and near-term workstreams;
• New Era's 450 MW behind-the-meter generation announcement for the Thunderhead / Turbine-X power pathway;
• NUAI: A New Era of Compute, reviewed from the private NUAI research archive, as a secondary synthesis of the site, partners, power question, Macquarie timing, land contribution, waterfall economics, and execution risks;
• Northland Securities' May 26, 2026 company update, used only as conflicted secondary research and risk framing; and
• TeraWulf's July 6, 2026 Anthropic lease announcement as an external AI-infrastructure lease benchmark.

The Northland report discloses conflicts and is not treated as independent confirmation. Secondary research and comparable-company data can help form questions and calibrate ranges, but only filings, executed contracts, permits, lender documents, and counterparty disclosures can upgrade uncertain inputs into verified facts.

Related spreadsheet model

Open the TCDC Site Economics Model to review the current scenario table, phase buildout, assumption evidence map, sensitivities, limitations, revision log, and change log.

Article revision summary

Version 1.0.0-preliminary moves the public NUAI model from the original demonstration scaffold to the current Google Drive workbook basis. The change adds in-workbook revision tracking, a baseline archive copy, assumption-level evidence mapping, source-backed model overrides, and updated public model limitations. It does not remove the need to verify final leases, project financing, power contracts, permits, ownership, or dilution before relying on the output.