NUAI
NUAI · New Era Energy & Digital, Inc. · Equity Research Report · as of August 19, 2026
Key figures, each tagged by provenance. First-quarter figures are the restated ones from the Form 10-Q/A of August 14, 2026. Market figures are struck on the August 18 close, the last completed session, which is the first to fall after the business update call of the previous evening, as well as after the two filings and the release of 14 August.
| Metric | Value | Basis |
|---|---|---|
| Share price | $5.13 | Market Aug 18, 2026 close, the last completed session before this stamp |
| Market capitalization | $546.6M | Estimate 106,559,339 shares at the $5.13 close, Aug 18, 2026 |
| 52-week range | $0.32–$9.45 | Market Aug 3, 2026, 251 sessions, as the quote source reports it |
| Short interest | 12.5M sh | Market Jul 31, 2026 |
| Shares outstanding | 106,559,339 | Filed 10-Q, Aug 14, 2026, cover page, as of 10 Aug |
| Fully diluted shares | 121.8M | Filed 8-K Exhibit 99.1, Aug 17, 2026, capital structure table as of 10 Aug: 121,824,660. Refreshed from the investor presentation, the June quarter reporting no diluted count because every potentially dilutive security is anti-dilutive against a loss |
| Revenue, FY2025 | $885K | Filed 10-K, Mar 12, 2026 |
| Net loss, FY2025 | ($29.6M) | Filed 10-K, Mar 12, 2026 |
| Net loss, 1Q26 restated | ($10.8M) | Filed 10-Q/A, Aug 14, 2026. Was ($9.0M) as originally reported |
| Net loss, 2Q26 | ($20.4M) | Filed 10-Q, Aug 14, 2026 |
| Revenue, 2Q26 | $36,497 | Filed 10-Q, Aug 14, 2026, against $16.1M of G&A |
| Cash | $84.8M | Filed 10-Q, Aug 14, 2026, at 30 Jun: $69.8M unrestricted plus $15.0M restricted |
| Undrawn credit facility | ~$270M | Filed 8-K, Apr 14, 2026 |
| Land cost capitalized | $80.4M | Filed 10-Q, Aug 14, 2026, at 30 Jun |
| Accumulated deficit | ($74.6M) | Filed 10-Q, Aug 14, 2026, at 30 Jun. Restated to ($54.2M) at 31 Mar |
| Site, controlled acreage | 493 acres | Filed Company release, Aug 14, 2026, on the corridor closing |
| Designed capacity, gross | 1,507 MW | Estimate Phased design, three phases, at the Phase 2 rating of the August 2026 air permit application |
What the company sells, to whom, and how it is paid.
New Era assembles powered land: large parcels where electricity, fibre, water and permits are secured in advance, so a tenant can begin building compute infrastructure without waiting years for a utility connection. It then leases that capacity under long-term contracts. The company does not yet operate a data center and has not yet signed a customer.
The flagship asset is Texas Critical Data Centers, a contiguous campus in Ector County roughly ten miles north of Odessa, in Odessa's extraterritorial jurisdiction. Total controlled acreage is 493 acres following the corridor closing announced on August 14, 2026, with $80.4M of land cost capitalized at June 30, 2026, assembled in stages from an original parcel, then 203 contiguous acres closed in December 2025, then a 54-acre corridor. Approximately 13,000 feet of legacy oilfield flowline has been remediated and pipeline reclamation is materially complete. The site is designed for three phases totalling approximately 1,507 MW of gross capacity and 1,005 MW of critical IT load at an assumed power usage effectiveness of 1.5. Phase 2 was re-rated from 450 MW gross to approximately 550 MW by the air permit application of August 2026, and the totals move with it.
| Phase | Gross MW | IT MW | Power source | Target |
|---|---|---|---|---|
| Phase 1 | 207 | 138 | Vistra and Calpine, grid-tied and adjacent | 2H 2027 |
| Phase 2 | 550 | 367 | Behind-the-meter, Siemens turbines | 2H 2028 |
| Phase 3 | 750 | 500 | Grid and behind-the-meter hybrid | 2028+ |
| Total | 1,507 | 1,005 | Diversified | through 2030 |
Three capacity figures circulate and they are not in conflict: 207 MW is Phase 1 gross, 757 MW is Phases 1 and 2 combined after the Phase 2 re-rating, and 1,507 MW is the full three-phase design. The 650 MW figure this document carried for Phases 1 and 2 predates that re-rating; it survives elsewhere only where it describes the behind-the-meter block management discussed in June, which is a different quantity.
The competitive claim is geographic rather than technological. Connecting a new campus to the Texas grid can take three to five years, and the ERCOT West Zone covering the Permian faces both queue saturation and load-pocket constraints. This parcel offers two ways around that, and it is the combination rather than either route alone that constitutes the advantage. Front-of-meter delivery buys grid-tied power from the two generation complexes at the property line through an existing point of interconnection, without traversing long-distance transmission corridors. Behind-the-meter generation burns gas from three pipelines that cross or terminate nearby, Enterprise, ONEOK and White Water, in on-site turbines that never touch the public grid.
The mix between the two is currently moving. At an investor conference on June 23, 2026 management said the parties are actively exploring a fully islanded behind-the-meter configuration, under which 650 MW of gross behind-the-meter capacity becomes the primary focus and total site capacity could support high-800 MW+ of critical IT load. Five days later a covering firm still modelled Phase 1 as front-of-meter. No company announcement has resolved the proportions.
This is a rebalancing rather than a reversal, and the turbine count is the readable signal. Thunderhead holds 460 MW of Siemens turbines under contract, sized against a Phase 2 that was 450 MW and is now rated at approximately 550 MW on different generation equipment. A genuinely islanded 650 MW campus needs more than 650 MW of generation, because without a grid backstop it carries its own spare capacity for maintenance and failures. So either more turbines are ordered, the grid connection stays as backup, or the first islanded block is smaller than 650 MW. Islanding also carries a regulatory logic: management's stated rationale included avoiding ratepayer concerns, which is the objection driving moratoriums nationally, and a campus off the public grid largely removes that argument while trading grid curtailment exposure for gas supply and generation reliability risk.
Recent comparable leases in this industry price at $140 to $190 per kW per month on fifteen-year terms; both firms modeling the company use a conservative $120. At that rate Phase 1's 138 MW of critical IT load implies approximately $87.6M of annual revenue at a net operating income margin near 95%. Leases are expected to be triple-net, with the tenant paying operating costs, taxes, insurance and power, which insulates the landlord from commodity exposure.
Construction is not intended to be funded from the parent balance sheet. Each project is structured as a standalone special purpose vehicle with New Era as sponsor, developer and general partner, while institutional investors provide the majority of capital as limited partners and roughly 80% of cost comes from project debt sized to contracted cash flows. Ring-fencing each project isolates development and operational risk and avoids cross-collateralisation. This is why a company valued near $517M can contemplate a Phase 1 build costing approximately $1,794M.
The structure cuts both ways. Project finance limits how much of the buildout the company must fund itself, which is the only reason the plan is arithmetically possible, but it also means the company owns a minority slice of what it developed. Covering firms model 45% to 49% of the venture and one assumes the operating partner takes control at approximately 51%. That percentage has not been negotiated or disclosed, and it moves the valuation more than any other single input.
At the same June conference management described an underserved 20 MW to 50 MW segment and said it is evaluating modular solutions deployable and operational within roughly nine months of acquiring a site, aimed at customers needing capacity quickly but not at hyperscale size. The corporate website followed in July with three named product lines, turnkey solutions, hybrid colocation, and enterprise solutions for AI inference and latency-sensitive workloads, offered in blocks from 2 MW to 25 MW. Both covering firms had anticipated tenant diversification into GPU cloud providers and enterprise buyers as a later-phase strategy, and that it was discussed with analysts a month before appearing in marketing materials supports reading it as ordinary commercial development rather than a change of plan.
Each stream, with its share of the total and its durability.
The company has one revenue stream today and expects a different one tomorrow. They share almost nothing.
| Period | Revenue | Source and basis |
|---|---|---|
| FY2024 | $0.5M | Natural gas and natural gas liquids, legacy Filed |
| FY2025 | $0.9M | Natural gas and natural gas liquids, legacy Filed |
| 1Q2026 | $0.8M | Legacy operations, provisional pending restatement Filed |
| FY2026E | $2.0M | Entirely legacy Estimate |
| FY2027E | $15.4M | First lease revenue in the fourth quarter Estimate |
| FY2028E | $166.7M | Phase 1 full year plus Phase 2 ramp Estimate |
Current, legacy exploration and production. Held through Solis Partners: roughly 137,000 acres in Chaves County, New Mexico, the Pecos Slope, with approximately 15.1 Bcfe of proved reserves, plus overriding royalty interests in Howard County, Texas. Operations run through third-party processors under a month-to-month arrangement, with gas and liquids sold at index-based pricing. Following divestiture of the oil assets in 2024 the revenue base is concentrated in natural gas and natural gas liquids. FY2025 net revenue of approximately $0.9M was driven by improved realized gas pricing rather than volume growth, with net gas revenue of approximately $0.6M representing about 73% of the total and liquids revenue of approximately $0.24M. Helium, once the company's name and purpose, generates no revenue at all under existing processing agreements. The package carries approximately $3.3M of net book value, against which published valuations of $19M to $84M imply most of the value sits in acreage optionality the balance sheet does not reflect. Filed
Future, data center leases. Long-term triple-net agreements with power costs passed to the tenant. Under the venture structure the company receives its ownership percentage of the property's net operating income rather than gross rent. One covering firm additionally models a royalty stream on gross capacity operated by the power partner, 450 MW in Phase 2, a figure the August 2026 air permit application raises to approximately 550 MW, and 350 MW in Phase 3, which is a modeling construct rather than a disclosed contract term.
The step from $2.0M to $166.7M across two years is not a forecast in the ordinary sense. It is what happens if a lease is signed, financing closes, construction completes on schedule and the tenant takes occupancy. Every dollar beyond FY2026 is conditional on a contract that does not exist, and the covering firms' own 2027 estimates diverge by 42%, at $15.4M against $21.9M, purely on assumptions about when revenue begins.
The demand side is the least disputed part of this story. AI workloads consume enormous electricity, and the binding constraint is no longer chips or capital but power and the physical delivery of campuses. A hyperscaler can build a data center in 18 to 24 months; a utility takes seven to ten years to upgrade the grid and construct the high-voltage transmission needed to serve it.
All four companies identified as candidate anchor tenants reported within nine days of each other, and all four confirmed rising spending.
| Company | Reported | FY2026 capital expenditure | Cloud growth | Market reaction |
|---|---|---|---|---|
| Alphabet | Jul 22, 2026 | $195B to $205B, raised Market | Cloud +82% | Fell |
| Microsoft | Jul 29, 2026 | ~$175B, reclassified Market | Azure +43% | Rose |
| Meta | Jul 29, 2026 | $130B to $145B, low end raised Market | n/a | Fell ~9.6% |
| Amazon | Jul 30, 2026 | ~$220B, raised from $200B Market | AWS +36.7% | Rose ~10% |
Microsoft's apparent reduction is not one. It reclassified future data center leases from finance leases to operating leases and extended assumed useful life from 15 to 25 years, which moves spending off the capital expenditure line while leaving total outlay unchanged. Combined 2026 spending across the four is tracking to roughly $725B, up about 77%.
Two statements bear directly on demand for third-party capacity. Alphabet's management said the company plans to expand the use of third-party capacity in the third quarter as a bridging strategy while it builds internal capacity. Amazon's management expects capacity constraints to persist through 2027, with demand already strong for 2028, and Amazon Web Services growth accelerated for a fifth consecutive quarter against a backlog of $496B. Two named candidate tenants have now stated publicly that they need more capacity than they can build themselves, on the timeline this campus targets.
Lease economics have repriced upward on that scarcity. Disclosed hyperscale colocation leases of the past eighteen months sit in the $140 to $190 per kW per month range, and tenors have lengthened from a historical seven to ten years to a current standard of fifteen-year triple-net structures. Capital intensity has climbed in parallel, from $9M to $11M per megawatt of IT load two years ago to $12M to $14M now, reflecting higher rack densities of 140 to 200 kW against 30 to 100 kW historically, direct-to-chip liquid cooling becoming standard, and supply-chain inflation in switchgear, transformers and turbines. The company's assumed $13M per critical IT megawatt sits at the low end of that range.
Everything above is true of the industry and would be equally true if this company did not exist. One covering firm counts five to ten tier-one hyperscalers against more than 40 potential landlords. A rising tide raises the sector's multiple; it does not sign anyone's lease. There is also a countervailing signal: the market has begun punishing this spending rather than rewarding it, with Meta's capital expenditure of $31.1B against $31.9B of operating cash flow collapsing free cash flow to $784M from $8.5B. Hyperscalers under pressure to protect free cash flow have more reason to lease third-party capacity, since leasing shifts capital onto someone else's balance sheet, but a broad repricing of AI infrastructure would compress the multiples used to value developers. The demand case and the valuation case can move in opposite directions at the same time.
Local resistance is growing and has stopped being merely local. United States municipal data center moratoriums grew from approximately five to 79 over the past year. On July 14, 2026 the State of New York enacted the first statewide moratorium on new hyperscale data centers, a pause of up to one year while the state studies energy demand, utility rates, water and environmental impact, and the State of Arizona separately enacted a three-year moratorium on new sales tax breaks.
The first project caught was the development partner's own. The New York moratorium paused Stream Data Centers' proposed campus at STAMP in Genesee County, a roughly $19.4B, 2.2 million square foot project requiring approximately 500 to 600 MW, which is a like-for-like substitute for this campus at 650 MW. The delay pushed the prospective tenant toward sites that can realistically deliver by 2027, with this campus moving higher on its priority list and discussions increasingly centerd on delivering capacity as quickly as possible. That is a specific causal explanation for the urgency both firms describe.
Texas legislation has been broadly favorable and Ector County officials have supported the project, travelling with the company to meet the prospective tenant. The asymmetry holds only while that remains true, and leadership turnover at the county commissioners' court is a named risk.
Named peers on comparable metrics, not a general market description.
The competitive field did not exist in its current form three years ago. Most listed peers are former bitcoin miners that already controlled power-connected industrial sites and redirected them toward AI colocation when the economics turned.
| Group | Named peers | Position relative to this company |
|---|---|---|
| Miners turned colocation | Applied Digital, TeraWulf, Cipher, Hut 8, IREN, Core Scientific, Riot, Galaxy Digital, Bitdeer, CleanSpark, WhiteFiber | Ahead; most have signed leases and operating capacity |
| Data center trusts | Equinix, Digital Realty, Iron Mountain | Far larger and profitable, trading at 17x to 20x earnings before interest, tax, depreciation and amortisation |
| Private platforms | Stream Data Centers, Prime Data Centers, Apollo-backed vehicles | Better capitalized, and Stream is this company's development partner |
| New Era Energy & Digital | n/a | Pre-lease, pre-revenue, pre-construction |
On the metric the sector uses, enterprise value per megawatt of contracted capacity, the company trades at approximately $15.9M against a peer average near $17.6M. The discount reflects pre-lease status, and the highest multiples in the group belong to operators with binding leases in place.
Because the company has no operating history of its own, the quality of its development partner carries unusual weight.
| Measure | Detail | Basis |
|---|---|---|
| Founded | 1999, headquartered in Dallas | Open |
| Delivered | 27+ data center developments across the United States | Open |
| Customer quality | Over 90% of capacity leased to Fortune 100 customers | Open |
| Retention | Zero early customer terminations | Open |
| Pipeline | ~3.6 GW: 1.45 GW Dallas-Fort Worth, 1.08 GW Atlanta, 500 MW Columbus, 350 MW Phoenix, 260 MW Chicago | Open |
| Footprint | Actively evaluating sites in more than 20 states | Open |
| Operating model | 100% of engineering and site operations performed in-house | Open |
| Uptime target | 99.9999% | Open |
| Ownership | Majority-owned by Apollo Global Management since November 2025 | Open |
Three uptime figures are now in circulation and are worth reconciling. The development partner targets 99.9999%, roughly 32 seconds of annual downtime. One covering firm cites 99.999%, roughly 5 minutes, as the contractual standard investment-grade hyperscalers require. The company's own website advertises 99.99%, roughly 52 minutes. The most likely explanation is that the website figure describes the smaller enterprise product rather than the hyperscale campus, but for assessing whether the platform can meet an investment-grade tenant's service levels the operative number is the partner's, because the partner would be operating it.
Read for burn rate and balance sheet capacity rather than for growth. The 1Q26 column is the restated quarter from the Form 10-Q/A of August 14, 2026.
| Line, $M except per share | FY2024A | FY2025A | 1Q26A restated | 2Q26A | FY2026E |
|---|---|---|---|---|---|
| Total revenue | 0.5 | 0.9 | 0.5 | 0.04 | 2.0 |
| Operating expenses | 12.1 | 25.4 | 10.2 | 16.9 | 30.9 |
| Operating loss | (11.6) | (24.5) | (9.7) | (16.9) | (28.9) |
| Net loss | (13.8) | (29.6) | (10.8) | (20.4) | (35.2) |
| Loss per share | (1.06) | (1.04) | (0.19) | (0.21) | (0.35) |
| Adjusted EBITDA | (11.6) | (23.9) | n/a | n/a | (21.6) |
Actuals Filed from the FY2025 annual report, the Form 10-Q/A of August 14, 2026 for the restated first quarter, and the Form 10-Q of the same date for the second. Adjusted EBITDA is shown as not available for both 2026 quarters: it is a non-filed measure, the only published first-quarter figure was modelled on the withdrawn accounts, and no second-quarter figure has been published. Estimates Estimate from covering firm models; a second firm carries FY2026 adjusted EBITDA at negative $27.7M. Every FY2026 estimate here predates both the restatement and the second quarter.
The FY2025 loss includes a $12.0M impairment taken in the fourth quarter against the partially built Pecos Slope gas processing plant, reflecting the strategic exit from helium midstream. The accumulated deficit stood at $54.2M at March 31, 2026 as restated, and $74.6M at June 30, 2026. Operating cash burn ran roughly $9M in the first quarter and is projected to accelerate as development spending increases, with 2026 capital expenditure modelled at $250M to $300M.
The amendment was filed on August 14, 2026 and the 1Q26 column above is now the restated one. Revenue moved $802,353 → $514,587, general and administrative expense $7,364,387 → $9,162,195, and net loss $(8,991,887) → $(10,823,974), wider by 20.4%. The revenue cut is the part no one was warned about: the July non-reliance report described expense classification, stock compensation and the acquisition valuation, and never mentioned revenue. It appears in the amendment only under other adjustments, down $287,766 with the matching amount removed from receivables.
The two named corrections did run in opposite directions, as this report expected, and the sizes are now known. The ASC 718 correction added $3,427,662 to expense and the same amount to paid-in capital, so it is non-cash and equity-neutral; the corrected grant-date fair value of the awards is $57,336,238 against $23,538,447 as originally determined. The classification correction reduced expense by $1,629,854. Total stockholders' equity therefore rose, $10,098,363 → $11,605,613, while the loss widened. A summary of this restatement that moves both in the same direction is wrong.
What makes this serious is what it says about the control environment. The same filing repeats a disclosure from the original quarterly report: management concluded that disclosure controls were not effective as of March 31, 2026 and identified a material weakness in internal control over financial reporting. That is a deficiency significant enough that a material misstatement could occur without being caught, and the restatement is that weakness becoming visible. The material weakness therefore pre-dates the restatement rather than resulting from it.
| Measure | December 31, 2025 | After the April 2026 financings |
|---|---|---|
| Cash | $1.2M Filed | $84.8M at 30 June, of which $15.0M restricted Filed |
| Near-term obligations | ~$73.9M Filed | Sharon AI note repaid in cash Filed |
| Committed credit | None Filed | $290M facility, ~$270M undrawn Filed |
| Going concern language | Present Filed | Substantial doubt formally alleviated under ASC 205-40 in the June quarter Filed |
| Shares outstanding | 53.4M Filed | 106.6M at 10 Aug Filed |
| Total assets | $14.1M Filed | $174.7M at 30 June Filed |
The cost of that rescue was ownership. Shares outstanding roughly doubled in under five months, and the registered offering priced at $3.35 while the stock traded near $5.00. A holder of 1% of the company at December 31, 2025 who did not participate held about 0.53% by 12 May and about 0.50% by 10 August, before counting a further 15.3M shares of warrants, options and restricted stock that bring the fully diluted count to 121.8M. That count is stated as of 10 August 2026 in the capital structure table of the investor presentation furnished on August 17, and reconciles to its own total: 106,559,339 shares, 9,416,958 unexercised warrants, 5,298,363 restricted stock units and 550,000 employee options. It excludes 9,649,418 performance stock units. The June quarter itself reports no diluted share number at all, because every potentially dilutive security is anti-dilutive against a loss.
Shares, preferreds, at-the-market capacity, shelf, warrants, convertibles.
The most consequential feature of this capital structure is a term buried in a credit agreement rather than anything the company chose to advertise. On April 8, 2026 New Era closed a senior secured term loan facility of up to $290M with Macquarie, structured at the TCDC subsidiary level so the debt is secured against project assets rather than the parent.
| Tranche | Amount | Rate | Purpose and gate |
|---|---|---|---|
| Term Loan A-1 | $20M | SOFR + 5.50% | Pre-lease development, funded Filed |
| Term Loan A-2 | $30M | SOFR + 5.50% | Pre-lease development Filed |
| Term Loan A-3 | $40M | SOFR + 7.75% | Post-lease development Filed |
| Delayed draw | $200M | SOFR + 7.75% | Construction, milestone-gated Filed |
| Total commitment | $290M | Matures April 2029 | First-priority lien on project assets Filed |
Two features deserve attention. The facility carries a repayment premium of roughly 1.10x to 1.35x of capital deployed, fully earned at signing, which is owed on repayment regardless of how the project turns out. And it embeds an execution-linked covenant.
The agreement requires the company to execute a binding hyperscaler lease within six months of the 8 April close, approximately October 8, 2026, which is 50 days from this report. Failure to meet that gate allows the lender to accelerate repayment or enforce structured amortisation. A separate covenant requiring a $30M equity raise within 60 days has been satisfied. Because acceleration lets a lender declare the whole balance due at once, and because Macquarie holds a first-priority lien on substantially all project assets, the practical leverage is considerable: repayment can be demanded at the moment the company is least able to pay, with a claim on the land and development rights if it cannot.
Reasonable analysts read this term in opposite directions and both readings hold. One covering firm argues the covenant is an asset rather than a constraint, because it forces management to compress its commercial timeline into a window investors can underwrite, and a lender does not impose a six-month lease deadline on a project it expects to fail. The symmetric view is that a sophisticated lender built itself an exit ramp and pre-earned its premium.
One price signal is worth isolating. Alongside the loan, Macquarie purchased 1.0M shares at $5.00 while the concurrent public offering priced at $3.35, a 49% premium. Both covering firms read a project lender voluntarily paying above the marketed price as underwriting conviction uncommon at this development stage. It is one of the few informative data points available before a lease exists, because it involved the lender's own capital.
The 8 April agreement obliges the company to establish an at-the-market equity program with an aggregate offering price of at least $100 million. On July 17, 2026 Macquarie signed a Waiver and Consent Letter extending the deadline, now within 60 days of written notice or within five business days after the next quarterly or annual report, but the obligation itself remains. Such a program sells new shares gradually into the open market at prevailing prices rather than in one discounted block, which is flexible for the issuer and means a continuous, price-insensitive supply of stock for existing holders. Against a market capitalization of $546.6M, a $100M program is authority to issue roughly 18% of the company.
The registration groundwork appears already laid. Four Form S-3 shelf registrations and two amendments were filed during 2026: 23 January, being the $350M shelf effective 30 January, then 4 February, 10 April, and 8 May with an amendment on 18 May. An effective shelf is the precondition for the program, so the gap between notice and live share sales could be short.
Two readings hold at once. That Macquarie granted a waiver eleven days before the restatement filing is meaningful, because a lender preparing to accelerate does not extend deadlines, and it indicates continued engagement. But the same document confirms that further dilution is contractually required rather than merely likely, and the alternate trigger ties it to the next quarterly filing, which could put the program in place well before the October covenant resolves.
Seventeen statements of beneficial ownership have been filed against this issuer, all of them under the Commission's current SCHEDULE form types and none under the older SC spellings, because the company has only existed as a registrant since the December 2024 combination. Sixteen are on Schedule 13G and one is on Schedule 13D.
| Reporting person | Shares | % of class | Basis |
|---|---|---|---|
| BlackRock, Inc. | 6,224,612 | 6.1% | Filed Schedule 13G, Jul 30, 2026, event date Jun 30, 2026, Rule 13d-1(b) |
| Conversant Capital and affiliates, four persons | 5,387,220 | 5.3% | Filed Schedule 13G, Jul 20, 2026, event date Jul 13, 2026, Rule 13d-1(c) |
| Davidson Kempner Capital Management and affiliates | 5,166,249 | 5.22% | Filed Schedule 13G, Apr 16, 2026, event date Apr 10, 2026, Rule 13d-1(c) |
| Caracola Ventures Corp | 5,050,000 | 8.6% | Filed Schedule 13G/A no. 1, May 15, 2026, event date Mar 31, 2026, Rule 13d-1(c). Up from 2,711,600 at 5.1% on an event date of Jan 23, 2026 |
| Zhou Zachary Yi | 5,328,495 | 9.9% | Filed Schedule 13G/A no. 1, Feb 17, 2026, event date Dec 15, 2025, Rule 13d-1(c). Up from 3,500,000 at 6.5% three weeks earlier |
| Yang Hanju | 5,176,177 | 9.7% | Filed Schedule 13G, Jan 30, 2026, event date Dec 3, 2025, Rule 13d-1(c) |
| Susquehanna Securities and G1 Execution Services | 4,018,958 | 7.1% | Filed Schedule 13G, May 15, 2026, event date Mar 31, 2026, Rule 13d-1(b) |
| Pecos Slope Holdings LLC | 879,819 | 6.68% | Filed Schedule 13D, Dec 16, 2024, event date Dec 9, 2024. Never amended |
Six further holders filed on the combination date of December 9, 2024 and have not amended since: Joel G. Solis at 16.06%, Robert C. Solis and Casey J. Solis at 7.86% each, CR Financial Holdings at 7.43%, John Lipman at 6.41% and CHLM Sponsor-5 LLC at 5.04%. Byron Roth at 8.15%, Gordon Roth at 5.16% and CR Financial at 5.0% amended to a December 31, 2024 event date. Those percentages rest on a share count of roughly 13 million and cannot be compared with the 2026 figures above, which rest on a base near 100 million: the denominator is a different company.
Two different exemptions are in use and they are not interchangeable. BlackRock and the Susquehanna broker-dealers file under Rule 13d-1(b), the institutional route. Every other 13G holder files under Rule 13d-1(c), the passive-investor exemption, which is available to any person below 20% who acquired without a purpose of influencing control and which carries a shorter reporting deadline. Describing the register as institutional would overstate what these filings say.
Pecos Slope Holdings LLC filed a Schedule 13D on December 16, 2024 for an event date of December 9, 2024, the combination close, reporting 879,819 shares at 6.68% with sole voting and sole dispositive power. Pecos is a Texas company whose manager and sole shareholder is E. Will Gray II, and it received the shares as consideration for services previously provided to the company rather than by purchase.
Item 4 states that Pecos does not currently have plans or proposals falling under subparagraphs (a) through (j), while reserving the right to consider, discuss, evaluate or pursue such transactions with its advisers, the issuer or other persons. That is a weaker disclaimer than a Schedule 13G certification, which is why the holder is on 13D at all, and no amendment has been filed in the twenty months since.
The 6.68% was struck against the share count at the combination and the count has since risen roughly eightfold, so the position is very likely below 5% today on the same shares. That is an inference from the share count rather than a fact from a filing: no amendment states it, and none is required of a holder who simply falls below the threshold by dilution. What the filing establishes is that the statement stands unamended on the record, which is a different thing from the position being current.
Remaining items: approximately 8.2M warrants at a $2.00 strike remain unexercised and deep in the money, representing roughly $16.4M if exercised; a further 6.0M public and private warrants struck at $11.50 are far out of the money but would deliver approximately $66M. A $5.0M related-party note maturing September 30, 2026 converts to stock at maturity. Pre-change net operating losses are subject to limitation under Internal Revenue Code Section 382 following the December 2024 reverse recapitalisation.
Multiples shown with their basis and pricing date. No target price is issued in this document.
Conventional multiples do not work here. There are no earnings, so a price to earnings ratio is undefined. Revenue is under $1M against a market capitalization of $546.6M, so a price to sales ratio produces a number without meaning. The second quarter makes the point sharper: $36,497 of revenue in three months. Book value is distorted by land carried at cost and a business in transition. All three covering firms therefore value the project rather than the company: they estimate what the completed campus would earn, apply a multiple, subtract debt and take the company's ownership share.
| Firm | Analyst | Date | Published figure | Price at publication | Method and basis |
|---|---|---|---|---|---|
| Northland Securities | Mike Grondahl | Apr 16, 2026 | $11.00 | $5.11 | Market 283 MW credited at 19.0x; 45% venture ownership; 11% discount rate; 125M diluted shares. Disclosed relationship in the table below |
| Texas Capital Securities | Not established | Apr 16, 2026 | $8.60 | $5.11 | Market 10% discount to net asset value of $9.57; 1.0 GW modelled of 1.4 GW potential. Disclosed relationship in the table below |
| Northland Securities | Mike Grondahl | Jul 1, 2026 | $11.00 | $6.38 | Market Reiterated, basis unchanged, struck at $6.38, an implied upside of approximately 72% at that price |
| B. Riley Securities | Nick Giles | Jul 6, 2026 | $10.00 | $4.78 | Market Sum of parts, Phase 1 at 14x and Phase 2 at 12x; 49% ownership; 101.5M shares; $50M net debt. An aggregator dates this initiation to July 7, 2026, one session later, and the two are not reconciled |
| Texas Capital Securities | Derrick Whitfield | Aug 17, 2026 | $8.70 | $5.07 | Market Reiterated, not raised; the firm dates this figure to July 20, 2026 in its own history, after $8.60 on April 16 and $8.80 on May 27. 10% discount to net asset value. The price is the Aug 14 close, which the note labels as such and which is the session before it |
| B. Riley Securities | Nick Giles | Aug 17, 2026 | $10.00 | $5.46 | Market Reiterated, unchanged since initiation. Sum of parts at a 49% venture share: Phase 1 at 15.0x on $217M of estimated net operating income, Phase 2 at 12.0x on $524M, Phase 3 ascribed $368M, less capitalised overhead of $45M at 15.0x and $70M of net debt, over 107M diluted shares for $10.17, rounded down. The price is the same-day 17 August close, though the note’s own footnote declares the prior close |
| Northland Securities | Mike Grondahl, Logan Hennen | Aug 17, 2026 | $12.00 | $5.46 | Market Raised from $11.00, the only one of the three to move. Assumes ~505 MW of critical IT load, which is the company’s 757 MW gross at an assumed 1.5 PUE and is the firm’s derivation rather than a disclosed figure; 15-year term, ~99% net operating income margin, $12M per MW of project cost, 80% loan to cost, a 12.5x multiple on $120/kW/month, ~45% ownership, a two-year discount at 11.0% and 125M diluted shares. Price is the same-day close against a declared prior-close basis |
These are the opinions of the issuing firms, reproduced as market data rather than adopted. The clustering is less reassuring than it appears: three firms reaching $8.70 to $12.00 by different routes suggests convergence, but each runs the same two unverified assumptions, that a lease signs and that the company retains roughly half the venture. The three firms do not even agree on the share count they divide by, using 107M, 125M and 81.1M diluted, which alone moves a per-share figure by more than half. All three published on 17 August after the business update call: one raised, two reiterated. The 1 July reiteration was struck at a $6.38 share price against a $5.13 close on 18 August, and published targets are updated infrequently while the price moves daily.
One firm published the sensitivity analysis behind its figure, and it is the most useful disclosure in any of the three. Holding capacity constant at 283 MW and varying only the ownership percentage:
| Venture ownership | 10% | 20% | 30% | 45% | 50% |
|---|---|---|---|---|---|
| Implied value per share | $2.46 | $4.92 | $7.38 | $11.07 | $12.30 |
Basis: Estimate 283 MW credited, 19.0x multiple, 11% discount rate. Holding ownership at 45% instead and varying capacity gives $5.20 at 133 MW and $36.50 at 933 MW.
The single input that matters most has not been negotiated. That table spans $2.46 to $12.30 on identical capacity, identical multiples and an identical discount rate, and the only variable is a percentage in a joint venture agreement that remains a non-binding letter of intent. One firm assumes the operating partner takes control at approximately 51%, implying the company holds less than half. The 45% figure should be read as a third-party placeholder rather than a disclosed term.
The same firm published a deliberately conservative floor: Phase 1 only, a 15x multiple rather than 19x, discounted one year.
| Scenario | $/kW/month | Equity value | Per share |
|---|---|---|---|
| Bear | $100 | $492M | $3.55 |
| Base | $120 | $705M | $5.08 |
| Bull | $140 | $919M | $6.62 |
| Upper | $160 | $1.13B | $8.16 |
Basis: Estimate Phase 1 stabilised net operating income, 15x, 45% ownership, 125M diluted shares, one-year discount at 11%.
Note what this assumes: that Phase 1 exists, is leased and is generating stabilised income. Even granting all of that, the bear case sits below where the security traded on 31 July. The floor is not a floor if the lease never signs.
All three frameworks were built on the Q1 2026 quarterly report that the company said on 30 July should no longer be relied upon, and none has been revised. The corrections are non-cash and may prove immaterial to a project-based valuation, since these frameworks value future campus earnings rather than the current income statement, but the inputs have not been re-verified and the acquisition fair value review touches the carrying value of the asset all three are valuing.
Severity-ranked, most severe first.
Failure to execute a binding lease gives Macquarie acceleration rights. The most recent covering firm states plainly that a delay beyond that window would create complex restructuring dynamics. Possible consequences include facility acceleration, enforced amortisation, emergency equity issuance at a distressed price, or restructuring of partner economics. The repayment premium of 1.10x to 1.35x is earned at signing and payable on repayment either way.
Management concluded disclosure controls were not effective as of March 31, 2026 and identified a material weakness in internal control over financial reporting. Three restatement items remain open: expense classification, executive share award accounting, and the fair value assigned to the January buyout of the venture partner. The third is the one to watch, because it concerns the carrying value of the company's only material asset and an outside valuation expert has been engaged. An unquantified restatement sitting alongside an active class action alleging misstatements is harder to underwrite than either alone.
Phase 1 economics depend on a single anchor tenant, so withdrawal, material renegotiation or credit deterioration would impair the project, the financing and the venture simultaneously. The campus is essentially the whole company, with no second operating business to absorb a setback. The partner group is small and named, and because the consortium was assembled at the tenant's request, one partner departing would signal more than one relationship failing.
Total capital expenditure is projected to exceed $15B over the full buildout, requiring continuous access to project debt. Share count doubled in five months, one firm expects drawing the full facility will require additional equity raises, and the credit agreement contractually requires a $100M at-the-market program representing roughly 21% of current market capitalization. Dilution here is the base case rather than a tail risk.
Final terms covering capacity, tenor, pricing and curtailment during scarcity events could deviate from expectations. A power contract that fails during peak demand does not support a lease promising investment-grade uptime. Phase 2 additionally assumes an air permit from the state environmental regulator that has not been granted.
A 1.5 GW campus needs substantial cooling water, and no water rights or supply agreement has been disclosed. The fibre arrangement remains a memorandum of understanding rather than a definitive agreement. Neither appears in any covering firm's risk framework, which reflects analyst focus rather than confirmed resolution.
This campus would be the company's first. Cost overruns, permit delays and schedule slippage are the norm rather than the exception in first megaprojects, and the three covering firms already disagree by two quarters on first power.
New York's 14 July pause is the first of its kind and caught a project belonging to the company's own development partner. So far the effect has favored this campus by removing a competing site, but the same policy logic, that hyperscale campuses raise household electricity bills, could reach Texas. A behind-the-meter configuration is the structural defense, which is part of why islanding is being explored.
A securities class action covers purchasers from November 6, 2024 to December 29, 2025, following a December 2025 short-seller report. An independent internal investigation found no supporting evidence and institutional partners diligenced and proceeded, but the case is live and the restatement is the kind of event around which complaints are amended. Three claims against the founder personally remain pending in New Mexico; the company's own five claims settled for $1.0M with court approval final on July 9, 2026, without admission of liability.
The insider filings made since the leadership transition report compensation awards, including 325,000 restricted stock units to the Chief Accounting Officer on 23 June, rather than purchases. That is normal for newly seated executives, but it means the team has not put personal capital behind the story. A promotional layer also surrounds the security: an investor awareness firm discloses the company as a paid client.
Two escalations that the filing record does not show, each of which would be worse than what it does show. No late-filing notification has ever been filed, which is the form a company files when it expects to miss a reporting deadline and would appear before a late filing; the complete index of 175 filings holds none of the four NT types. And no auditor change: the auditor was consulted on and concurred with the restatement determination rather than resigning. A third escalation, which an earlier version of this document recorded as absent, has in fact occurred: one holder is on Schedule 13D rather than 13G, as set out in the capital structure section. None of this makes the restatement less serious, but the two escalations that usually accompany a deteriorating filing record have not occurred.
Each case with its preconditions: what must be true, not what might be.
The tenant walks, or negotiations extend past the point where the lender is willing to wait. Acceleration rights are exercised or renegotiated on punitive terms. Equity is raised at a distressed price, or partner economics are restructured. The class action survives dismissal and discovery begins, and the restatement quantifies larger than expected. What supports the price is $76.0M of capitalized land, a legacy portfolio valued at $19M to $84M, and 492 entitled acres beside two power plants, which retain strategic value to someone though likely realized through a sale rather than a lease. That is on the order of $95M to $160M of tangible anchor against a $546.6M market capitalization, and the security traded at $0.32 within the past 52 weeks.
Requires: no lease by the covenant date, or a lease on terms that do not satisfy the lender.
The signing day slips past 8 October. The lender grants a waiver or extension rather than accelerating, which is plausible because a lender holding a first lien on a half-built project generally prefers the project finished, but the terms cost something in fees, spread or ownership. The venture settles below 45%. Phase 1 energisation moves toward the more conservative first-quarter 2028 estimate. Additional equity is issued along the way, including under the mandated program. The published Phase 1 floor of approximately $5.08 per share at a $120 lease rate is a reasonable anchor for a delayed but real outcome.
Requires only: that a complex first-time megaproject takes longer than planned.
A binding lease with an investment-grade counterparty executes at or before 8 October at a rate at or above the $120 per kW per month both firms model conservatively. The venture goes definitive with the company retaining 45% to 49%. The A-3 tranche and $200M delayed draw unlock sequentially. The trading multiple re-rates from approximately $15.9M toward the $17.6M peer average, and the market begins pricing Phase 2 rather than doubting Phase 1. The published figures of $8.70 to $12.00 describe roughly this outcome.
Requires: lease signed on time, venture percentage at the high end, financing conditions holding, construction beginning on schedule.
Averaging these produces a number that describes no possible future. The covenant makes this closer to a binary than a distribution: either a creditworthy tenant signs a fifteen-year lease, or the capital structure has to be renegotiated. Position sizing rather than price targeting is the honest response to that shape.
New Era Energy & Digital is a genuine asset attached to an unproven business. The 493 acres are real, the adjacency to two operating generation complexes is real and not replicable, the cash of $84.8M at 30 June and the $290M facility are real, the Siemens turbines are contracted, and the executives hired between April and July 2026 bring operating experience from Microsoft, Amazon Web Services, TikTok, EdgeConneX, CoreWeave and Switch. The demand backdrop, roughly $725B of counterparty capital expenditure in 2026 with two named candidates stating publicly that they intend to rent third-party capacity, is as favorable as it has been.
One development since has cut in the company's favor by accident: the New York moratorium of 14 July paused a $19.4B project belonging to its own development partner, pushing the prospective tenant toward sites deliverable by 2027.
Against that: no signed lease, no executed power purchase agreements, no definitive joint venture, no disclosed ownership percentage, $885K of annual revenue, a share count that doubled in five months, a contractual obligation to establish a further $100M equity program, withdrawn quarterly financial statements with an unquantified restatement pending, an acknowledged material weakness in internal control, and a credit agreement that comes due for its first real test on approximately October 8, 2026.
The single question that resolves this security is whether a creditworthy counterparty signs a binding lease at this campus by approximately October 8, 2026. Everything else, the partner stack, the land, the turbines, the published figures and the product range, is preamble to that one signature.
Five points carry most of the weight. The going concern qualification was resolved by the April 2026 financings, so commentary still citing it describes a superseded position. The October covenant is contractual rather than aspirational and organizes the near-term calendar. The venture ownership percentage swings published fair value from $2.46 to $12.30 and has not been negotiated. The campus is a hybrid by design and the front-of-meter to behind-the-meter mix is being rebalanced, so turbine orders read more reliably than adjectives. And every published figure comes from a firm disclosing an investment banking relationship with the company, all were built on the withdrawn quarterly report, and none of the three carries a sell-equivalent rating anywhere in its coverage.
Requested categories with nothing to report, stated explicitly rather than omitted.
| Category | Status | Basis |
|---|---|---|
| Dividends | None | No dividend has been paid or declared; all three covering firms model a 0.00% yield through 2029. |
| Share buybacks | None | No program exists or is plausible while the company funds a capital program projected to exceed $15B and is contractually required to establish a $100M equity program. |
| Data center revenue | None | Checked against the filing record: no lease has been signed and all revenue through FY2026 is legacy oil and gas. First lease revenue is modelled in 4Q 2027. |
| Helium revenue | None | Helium is produced as a byproduct but generates no revenue under existing processing agreements, despite having been the company's original name and purpose. |
| Preferred stock | None | No preferred shares are outstanding in the capital structure disclosed. |
| Independent analyst coverage | Not established | Not examined. The three covering firms named in this file each disclose an investment banking relationship, but no search for unconflicted coverage beyond them was performed. Also named in what was not checked. |
| Insider open-market purchases | None | Insider filings since the leadership transition report compensation awards rather than purchases. |
| Activist ownership | None | Checked against all 17 beneficial-ownership filings in the complete index of 175 on August 3, 2026, under both the SC and SCHEDULE form types. One Schedule 13D is on file, from Pecos Slope Holdings LLC on December 16, 2024, and it disclaims current plans under Item 4 and has never been amended. Every other holder above 5% is on Schedule 13G. No holder has declared an intent to influence control, and none has converted from the passive form. |
| Late-filing notifications | None | Checked against the filing record: no Form NT has been filed, notwithstanding the pending restatement. |
| Auditor change | None | The auditor was consulted on and concurred with the restatement determination rather than resigning. |
| Foreign private issuer filings | None | Checked against the filing record and found empty because the category does not apply: the company is a Nevada corporation reporting as a United States domestic registrant, so foreign private issuer forms are not available to it. |
Pricing basis. Price and market capitalization are struck to the same session, the August 18, 2026 close of $5.13, and reconcile at 106,559,339 shares outstanding: 106,559,339 multiplied by $5.13 is $546,649,409, carried as $546.6M. That session is not the as-of date, and what the gap contains has changed. The August 18 close is the first to fall after the business update call, which was held at 5:00 p.m. Eastern on August 17 after that day’s session, and it is also after the Form 10-Q, the Form 10-Q/A and the company release of 14 August, all of which landed at 16:31 Eastern on the 14th. So the price carried here is the first the market has set with every disclosure in this document in front of it, the call and the investor presentation furnished that day included. It is not a favourable one: the shares fell 6.04% from the $5.46 close of 17 August on which this document was previously struck. The $471.44M carried at the July 31 basis did not quite reconcile at the count and close then stated, being about $0.6M below 101,290,928 at $4.66; the arithmetic is shown here so the reconciliation can be checked rather than asserted. Quote sources have disagreed materially in this security, publishing market capitalizations between $412.8M and $588.5M within a single week on a stable share count.
Price at publication in the coverage table is the last close before the note, not the close on the note date: a note carrying a given date is written against the preceding session. The file already used that convention in stating the 1 July reiteration was struck at $6.38, which is the 30 June close, and the column now applies it throughout.
The four tags used throughout this file and what each asserts. Tags are not color-coded: provenance is neither meaning nor identity, and a filed fact can be bad news.
| Tag | What it asserts |
|---|---|
| Filed | Stated in an SEC filing or company release, cited by form and date. Also covers a dated official publication by a named non-SEC issuer, with the issuer named in the source line. |
| Estimate | Derived or inferred here, or modelled by a named third party. The arithmetic or the basis is shown. |
| Open | Expected but unconfirmed. Nothing filed either way. |
| Market | Price, volume, float, short interest, and published targets and ratings, stamped with the close or publication date. A third party's target or rating is market data, attributed to the firm that issued it and never adopted here. |
Fact against assumption. Body text and table figures are sourced. Blocks marked as analyst assumption are interpretation, as is the scenario framing in the bull, base and bear section. Where covering firms disagree, the divergence is shown rather than averaged.
Estimates. Figures for FY2026 through FY2028 are third-party projections rather than company guidance. Published figures reproduced in the valuation section are the opinions of the issuing firms and are presented as market data rather than adopted; this document issues no target price and no recommendation.
Filings: FY2025 Form 10-K, March 12, 2026, with amendment; Q1 2026 Form 10-Q, May 15, 2026, restated by the Form 10-Q/A of August 14, 2026, which is the source of every first-quarter figure in this report; Q2 2026 Form 10-Q, August 14, 2026; Forms 8-K dated 16 January, 25 March, 31 March, 6 April, 8 April, 14 April, 28 May with Exhibit 99.1, 6 July, 22 July, 30 July and August 17, 2026, the last furnishing the investor presentation as Exhibit 99.1 under Item 7.01; Form S-3 registration statements of 23 January, 4 February, 10 April and May 8, 2026 with amendments of 12 March and 18 May, with prospectus supplements filed under Rule 424 against those registrations; preliminary and definitive proxy statements of March 5 and 16, 2026 with supplements of 10 April; Schedule 13G filings including that of July 20, 2026; Forms 3 and July 4 to 8, 2026.
Company communications: business update calls of 17 March, 18 May and August 17, 2026, the last held at 5:00 p.m. Eastern on the day its investor presentation was furnished; an investor conference appearance of June 23, 2026; releases via GlobeNewswire and BusinessWire of 27 February, 1 April, 8 April, 9 July and August 14, 2026, the last of which resolves the permitting catalyst and re-rates Phase 2; the corporate website and investor relations platform accessed July 2026.
Third-party research: Northland Securities initiation of April 16, 2026 with company updates of July 1 and August 17, 2026; Texas Capital Securities initiation of April 16, 2026 with a company update of August 17, 2026; B. Riley Securities initiation of July 6, 2026 with a company update of August 17, 2026.
Counterparty and market data: quarterly results and earnings calls of Alphabet, Microsoft, Meta and Amazon between July 22 and 30, 2026; quote data for the August 18, 2026 close, the last completed session before this stamp; a short-seller report of December 12, 2025; law firm releases of April and May 2026.
The seven entries above are the notes read in full, which is what allows the method behind each figure to be set out. A search on August 3, 2026 found three further actions carried only by aggregators and not read here: a $10.00 figure from New Street Research, analyst Derrick Whitfield, dated July 7, 2026; and grades from two quantitative ratings services, a Sell of May 22, 2026 and a Hold of April 17, 2026, neither carrying a target or naming an analyst. Market They are named so the four are not read as the whole of the coverage, and they are not tabled beside notes that were read, because the evidence behind them is weaker. One service publishes a consensus of $10.33 over seven contributors, reported as that service figure and not adopted.
Three firms initiated within twelve weeks, all with buy-equivalent ratings, and all three published again on August 17, 2026 at $8.70, $10.00 and $12.00. Their research is detailed and in places the best publicly available work on the company. It should still be read with the incentives visible.
| Firm | Disclosed relationship | Buy-equivalent | Sell-equivalent |
|---|---|---|---|
| Northland Securities | Makes a market; received banking compensation in the past twelve months; managed a public offering and a financing; intends to seek further banking compensation within three months Filed | 79.4% | 0.00% |
| Texas Capital Securities | Received banking compensation in the last twelve months; expects to seek more within three months; received non-banking compensation Filed | 92.0% | 0.00% |
| B. Riley Securities | Acts as market maker or liquidity provider Filed | 79.1% | 0.00% |
Two further points of provenance, neither improper. The company's Chief Financial Officer led the energy, power and digital infrastructure investment banking practice at one of the three firms from 2020 through early 2026, where that firm structured and sole-managed more than $7B in data center financings, and joined the company as Chief Financial Officer in March 2026; that firm initiated coverage roughly one month later with the highest published figure of the three, having also disclosed managing an offering for the company in the prior twelve months. Separately, the research set reached this analysis through the company's own investor relations agency, whose distribution watermark appears on one of the reports, so the available sample is curated and uniformly positive rather than a survey of all views.
Risk sections in these documents should be read as boilerplate rather than current assessment. The 1 July update still lists among its risks that as of December 31, 2025 the company did not have sufficient liquidity to fund operations, raising substantial doubt about its ability to continue as a going concern. That text was carried forward unchanged from the April initiation, whose own body text stated the April financings removed key funding uncertainties that had raised going concern risk. On the same page a data block shows cash per share of $0.02, roughly $2M, which is the 31 March balance sheet struck before the April raise closed. Where a document contradicts itself, the filing is the authority. All three initiations were built on the Q1 2026 report the company has since withdrawn, and none has been revised for the restatement.
None of this makes the analysis wrong, and small-capitalization companies are routinely covered only by banks that have done business with them, because that is how coverage gets paid for. But the absence of a bearish published view is not evidence that no bearish view exists. It may reflect who writes research on a company of this size and why.
The Form 10-Q/A, when filed, should be compared against the original Q1 2026 report; the Form 8-K of July 30, 2026 is the clearest statement of what remains unresolved; the FY2025 annual report carries the risk factors and the facility terms. Insider and ownership filings, current reports and any late-filing notification appear on the public filing system before they appear anywhere else, and a conversion from the passive to the control ownership schedule would signal a change in intent among large holders.
The filing sweep behind this document was run against the complete submissions index for Central Index Key 0002028336 and reached August 18, 2026, at 186 filings, so nothing filed up to that date is outside the evidence base. The most recent is the Form 8-K of August 17, 2026 furnishing the investor presentation. The company's own investor relations news feed was read on August 18, 2026 and carries nothing later; material published away from either source was not re-read. Whether unconflicted published coverage of this company exists beyond the three firms named above was not examined.
Further gaps, each a recorded absence rather than a clean result:
Defects identified but not yet resolved, with the text at issue quoted and the reason it remains open.
The diluted share count no longer rests on a superseded filing, and this limitation is retired. It read that the fully diluted count of “119.6M” was disclosed alongside the net loss of “($9.0M)” in the report the company said “should no longer be relied upon”, and that it could not be refreshed because the June quarter reports no diluted share number at all. The June quarter still does not. The investor presentation furnished on August 17, 2026 does, in a capital structure table stated as of 10 August, and the count is now 121.8M throughout this document. The restated first-quarter figures were already carried: the net loss is $(10.8M), not $(9.0M).
Newest first. The original build entry is never removed or rewritten.
Three firms published on the business update call and all three are now in the coverage table. Northland raised to $12.00 from $11.00, the only one to move; B. Riley reiterated $10.00; Texas Capital reiterated $8.70, a figure its own rating history dates to July 20 after $8.60 in April and $8.80 in May, so this document had been carrying an April figure that two later actions had already replaced. Each firm’s method is summarised from its own note. Nothing in any note is adopted as fact. Market
The phase table still rated Phase 2 at 450 MW gross, which this document had itself superseded two sections away. The prose says Phase 2 "is now rated at approximately 550 MW gross, taking Phases 1 and 2 together to roughly 757 MW", and the table said 450 and 300. The document stated its own capacity two ways. At the table’s own power usage effectiveness of 1.5, which reproduces its unchanged rows exactly, Phase 2 becomes 550 gross and 367 IT and the total 1,507 and 1,005. The sentence naming 650 MW as Phases 1 and 2 combined is corrected to 757. The key facts row still carrying the old gross total, and the risk paragraph that rounded the campus to the old total in gigawatts, are carried with it to 1,507 MW and 1.5 GW, so this document now states its design capacity once. The prose that reasons from the pricing session is carried with the basis as well, which it had not been: the sentence stating what the carried close does not yet contain was written of the earlier session and had become false, since the session now carried is the first the market completed after the call, and it fell 6.04% rather than rising. The sources paragraph now names the session actually used and all seven notes read, and the covenant countdown is recomputed from this stamp. This comes from the company release of 14 August, not from any note. Filed
The price-at-publication convention holds, and the notes are what break it. This document says the coverage price is the last close before the note, and measured against the exchange record that is exactly right for all four earlier rows. Of the three new notes, all three declare a prior-close basis and two print the same-day close: B. Riley and Northland both print $5.46, which is 17 August, while their own boilerplate declares the prior session; Texas Capital prints $5.07 and labels it August 14. Each new row therefore carries the price its note prints, with the session that price actually is, and says which.
Repriced to the 18 August close. $5.46 → $5.13, a fall of 6.04%, and the market capitalization $581.8M → $546.6M on the unchanged filed count, reconciling at 106,559,339 × $5.13 = $546,649,409. The at-the-market authority rises from roughly 17% to roughly 18% of the company because the denominator fell. The record did not move: nothing has been filed since the Form 8-K of August 17 and the investor relations feed carries nothing newer, both read on August 19.
What was read and not adopted. Northland reports an upsize "15% to 505 MW". That is not a different capacity from the 757 MW carried here: it is the same figure at an assumed 1.5 PUE, converted to critical IT load, and the firm says so. It is recorded as Northland’s method and nowhere else. Northland also describes the Phase 1 power purchase agreement as "substantially in final form"; no filing says so, the company said advanced negotiations, and the resolved catalyst in the companion calendar stands unchanged.
A primary source was filed on the day of the last stamp and was not in the evidence base. A Form 8-K of August 17, 2026 furnished the investor presentation as Exhibit 99.1 under Item 7.01. The previous version of this report was assembled that evening and records neither the filing nor the deck. Nothing has been filed or announced since: the submissions index stands at 186 filings with that 8-K the most recent, and the investor relations news feed was read on August 18 and carries nothing later. Filed
The covenant countdown read 66 days and had been left behind by two restamps. Sixty-six days is the interval from August 3 to the October 8, 2026 lease gate. The as-of moved to August 13 and then to August 17 without the figure moving, so a reader was told the gate was nine weeks away when it was seven. At this stamp it is 51 days. The instrument that scans for stamp-derived intervals did not flag it, because the five it does flag in this file are all unanchored and it does not read “from this report” as an anchor.
The diluted share count is refreshed and the known limitation is retired. This report has carried 119.6M since May, has said in two places that it could not be refreshed, and named it in Section 13 as the one defect still open. The June quarter indeed publishes no diluted count. The investor presentation does, in a capital structure table as of 10 August 2026 totalling 121,824,660, and the table reconciles to its own total: 106,559,339 shares outstanding, 9,416,958 unexercised warrants, 5,298,363 restricted stock units and 550,000 employee options. The count moves 119.6M → 121.8M in the masthead, the snapshot and the capital structure section. The excluded performance-gated units are stated at 9,649,418 against the 8.5M this pair carried. Filed
Repriced to the 17 August close. $5.07 → $5.46, up 7.69% on the session, and market capitalization $540.3M → $581.8M on the unchanged filed count of 106,559,339. This is the first close this document has carried that falls after the news in it. The 14 August basis was struck before the two filings and the release of that evening, and this report said so. The at-the-market authority falls from roughly 19% of the company to roughly 17% on the larger capitalization, and the price to sales and tangible anchor comparisons move with it.
What the call did not produce. The deck restates the release rather than advancing it. The Phase 1 power purchase agreement is still described as in advanced negotiations, no counterparty is named, no price and no date is given, and no tenant is named. That is the downside case the calendar stated for this catalyst before it was held. Permits do not move the Macquarie covenant date, which is now 51 days out.
What the deck adds that is not a figure in this document. It states that Phases 1 and 2 are islanded and behind-the-meter and do not require ERCOT interconnection, so they are unimpeded by the delays affecting queued projects; it breaks the Macquarie facility into Term Loan A-1 of $20M drawn, A-2 of $30M available pre-lease at the lender’s discretion, A-3 of $40M and a $200M delayed draw, both at lender discretion, against the $270M undrawn already carried; it shows Phase 1 commercialisation as four parallel workstreams rather than the three converging documents this report describes, adding construction permitting; and it states that project capital is to be raised at the asset level after lease execution, targeting roughly 80% debt. Filed
Read and not applied. The deck refers to open engagement “as state audit process advances” without naming the audit, its scope or its authority, and nothing in the filing record read here establishes any of the three, so it is recorded and not carried into the risk sections. The site survey and design renderings, the community programme detail and the partner roster are read and add no figure. The five beneficial-ownership filings of August 14 remain unread, as does the Form 8-K of July 22.
Added later the same day. Section 13 still listed the restated Form 10-Q as unfiled, in a report whose own sources block calls it the source of every first-quarter figure. It was filed as the Form 10-Q/A of August 14, 2026, four days before this stamp and one day before the revision above absorbed it everywhere else. The same instrument found this and a second instance in another ticker, and it looks for a class nothing here watched: a statement that the record contains nothing, true when written and falsified by a filing rather than by anything in the document. Filed
This report carried first-quarter figures the company had withdrawn, qualified as provisional. The amendment has landed and they were wrong, not merely provisional. Revenue $802,353 → $514,587, general and administrative expense $7,364,387 → $9,162,195, net loss $(8,991,887) → $(10,823,974), loss per share $(0.16) → $(0.19), total assets $86,484,525 → $88,012,583, stockholders’ equity $10,098,363 → $11,605,613. The snapshot, the financial performance table and the capital structure section all carried the superseded set.
The revenue cut was in none of the warnings. The July non-reliance report named expense classification, stock compensation and the acquisition valuation. Revenue was reduced $287,766 under other adjustments. This report reproduced the 8-K’s three matters and would have kept a withdrawn revenue figure indefinitely on that basis.
Equity rose while the loss widened, because the ASC 718 correction is equity-neutral and the classification correction runs the other way. The earlier text describing the corrections as pointing in opposite directions was right; the magnitudes were unknown and are now stated.
Phase 2 was carried at 450 MW in three places and is now rated at approximately 550 MW, on an air permit application filed by the power partner. The company attributes the increase to different generation equipment, which puts the 460 MW of contracted Siemens turbines in question as the Phase 2 equipment. This report does not resolve that.
The second quarter is established. Revenue $36,497 against general and administrative expense of $16,091,330, and a net loss of $(20,442,625). Loss per share is unchanged from the prior year at $(0.21) while the loss grew 5.7 times, because weighted average shares rose 481% to 98,171,489. Reporting either without the other misleads in opposite directions. The six-month revenue of $551,084 ties to the restated first quarter and not to the original, which is what proves the amendment is embedded in the second-quarter report.
Substantial doubt about going concern was formally alleviated under ASC 205-40 in the June quarter, on the April financings. Cash and cash equivalents were $69,821,390 with $15,000,000 restricted, total assets $174,695,388 and stockholders’ equity $138,584,967.
Repriced to the 14 August close. $5.10 → $5.07 and market capitalization $516.6M → $540.3M, the capitalization rising on a share count that moved 101,290,928 → 106,559,339 rather than on the price. The 14 August close predates every disclosure in this revision, all of which were published after that session ended.
Known gaps in this version. The fully diluted count of 119.6M is the first-quarter figure and is not refreshed, because the June quarter publishes no diluted count. Adjusted EBITDA is shown as not available for both 2026 quarters. Every FY2026 estimate carried here predates the restatement and the second quarter. The five beneficial-ownership filings of August 14 and the Form 8-K of July 22 are unread. The 52-week range and short interest keep their earlier dates, which are stated in the basis column.
Added later the same day. The caption under the financial performance table still sourced the actuals to “the FY2025 annual report and Q1 2026 quarterly report”, which attributed the restated first quarter to the filing this report says should no longer be relied upon, and named neither the Form 10-Q/A nor the second quarter that occupies a column of the table above it. The correction had been written by the pass that rebuilt the table and was never applied: its edit was registered after the script’s own entry point and so was never part of the plan that ran. The caption now names all three filings and states why Adjusted EBITDA is unavailable for both 2026 quarters.
Looking for that one caption found three more of the same kind, and they were worse. This report absorbed three disclosures of August 14 and its own primary sources list never gained any of them: it named the Q1 Form 10-Q as “now subject to restatement” when the amendment had been filed, did not name the Form 10-Q/A, did not name the June quarter report, and did not name the company release of that date. Its market data line still read the August 3 close after every figure in the report had been restruck on August 14. A sources list that omits the filings the figures come from states the report’s basis two ways, which is the same defect as a stale figure and harder to see. All four are corrected.
Repriced to the 3 August close. $4.66 → $5.10, up 9.44% on the session, and market capitalization $471.44M → $516.6M on the unchanged filed count of 101,290,928. The at-the-market authority falls from roughly 21% of the company to roughly 19% on the larger capitalization. The arithmetic is now shown in the methodology section: the figure previously carried did not quite reconcile at the stated count and close.
Third-party coverage extended. The four notes keep their figures and methods exactly. Added are the authoring analysts where established, and the close before each note, on the convention the file already used in stating the 1 July reiteration was struck at $6.38. Three further actions found by search are named but not tabled, being known only from aggregators: a $10.00 figure from New Street Research of July 7, 2026 and two quantitative ratings grades. An aggregator dates the B. Riley initiation to 7 July where this file has 6 July, and the two are left unreconciled.
Market capitalization was tagged as published market data and is computed in this document. The masthead figure and the Snapshot figure both read Market. Each is the 31 July close multiplied by the filed share count, and both now read Estimate. The tag asserted that a data provider had published the figure rather than that this document derived it. Price and market capitalization are struck to the same session and reconcile at the filed share count, so no restatement of the figure follows; only its provenance changes.
The legacy portfolio valuation of $19M to $84M was stated without the $3.3M of net book value it was set against, leaving a range with nothing to contrast it to. The book value is restored, and with it the point the comparison existed to make.
Filing sweep re-run on August 3, 2026 against Central Index Key 0002028336 rather than against the ticker, across the complete submissions index of 175 filings covering June 28, 2024 to July 30, 2026. The index does not paginate, so every absence recorded here rests on the whole record. Seventeen of those filings are statements of beneficial ownership, all under the SCHEDULE form types and all opened; both spellings were searched and the older SC types return nothing for this issuer. Nothing has been filed since July 30, 2026, being the Form 8-K and a Schedule 13G of the same date, and no late-filing notification has been made, so no analysis moved on new filings. The Form 10-Q/A remains unfiled, and figures drawn from the withdrawn quarterly report stay provisional.
No repricing was possible or needed: the latest close remains July 31, 2026, this report being assembled on a Sunday. The as-of stamp moves 1 August → August 2, 2026, and the covenant window is restated as 67 days.
Seven figures restored to the body, closing the backlog recorded in Methodology: the staged purchase of 203 contiguous acres and the 13,000 feet of remediated flowline in the site description, the $3.3M net book value noted above, the implied upside of approximately 72% beside the reiterated third-party figure, the prospectus rule under which supplements are filed, and the countdown to the covenant date. The market capitalization of $484 million previously carried in the executive summary is superseded rather than restored: it was struck to an earlier close than the masthead beside it, and reinstating it would reintroduce two prices for one company.
Ten figures previously carried from the companion document are now sourced here directly, each cited to the filing or published research that supports it, rather than to the calendar. None is an import any longer.
Q1 2026 financial statements withdrawn pending restatement. On July 24 the Audit Committee determined that the Q1 2026 Form 10-Q should no longer be relied upon (Form 8-K Item 4.02, filed July 30): approximately $1.4M of professional fees misclassified as G&A rather than deferred as issuance costs; ASC 718 errors on executive performance share units whose $23.5M grant-date fair value "was inappropriately calculated and understated"; and an ongoing re-evaluation of the fair value of the January TCDC acquisition, with a valuation expert engaged. The company cannot quantify the effect and says it may be material; there is no cash impact. Added as a header banner and worked through Sections 01, 07, 09, 11, 12, 13 and 14, including a note that all three sell-side frameworks were built on the withdrawn filing.
Material weakness surfaced. The filing repeats a disclosure from the original 10-Q that disclosure controls were not effective as of March 31, 2026, with an identified material weakness in internal control over financial reporting. This was present in the source filing and absent from earlier versions of this report; it is now a named governance risk in Section 11, alongside a Plain English explanation in Section 07 of what a restatement is and is not.
$100M ATM obligation added (Section 08). A July 17 Macquarie Waiver and Consent Letter extended but did not remove a Term Loan requirement to establish an at-the-market equity program of at least $100M, roughly 21% of market capitalization, triggerable within five business days of the next periodic report. A filing sweep also found four Form S-3 shelf registrations and two amendments during 2026, including a May 8 filing amended May 18, so the registration vehicle for the ATM appears already in place.
Hyperscaler capex completed (Section 05). All four candidate tenants reported: Alphabet $195–205B; Amazon raised to ~$220B with AWS +36.7% and capacity constraints expected through 2027 and demand strong for 2028; Meta $130–145B with free cash flow collapsing to $784M from $8.5B and shares −9.6%; and Microsoft's move to ~$175B identified as an accounting reclassification rather than a spending cut. A comparison table was added, with a note that the demand case and the valuation case can move in opposite directions.
Filing record characterised (Section 11). That note recorded three escalations as not having occurred, and one of the three had. A Schedule 13D is on file, from Pecos Slope Holdings LLC on December 16, 2024 for 879,819 shares at 6.68%, and the register section now sets out all seventeen beneficial-ownership filings rather than the two previously read. The other two stand: no late-filing notification has ever been filed, the statutory form that would precede a missed deadline, and no auditor change, with Weaver and Tidwell having concurred with the restatement rather than resigning. Forms 4 since the leadership transition report compensation awards, 325,000 RSUs to the Chief Accounting Officer on June 23, not open-market purchases. Verification guidance now names NT 10-Q and Schedule 13D as specific forms to monitor, and Section 15 documents which form types are swept, recording that 6-K, 20-F, 40-F, NT 20-F and NT 40-F do not apply to a Nevada corporation filing as a US domestic registrant.
Market data. Struck to the Friday, July 31 close: $4.66, market cap $471.44M, day range $4.44–$5.15, TTM EPS −$0.89. Covenant countdown 66 days to the October 8, 2026 lease gate, which is the interval from this document's own date and matches the capital structure section. It had been restated as 68 days, the figure that belonged to a 1 August stamp, while the body carried 66.
Incorporates the Northland Securities company update of July 1, 2026 and verification of its regulatory references. Power architecture reframed. The site schematic and Section 03 previously presented Phase 1 as grid-purchased power from the adjacent Vistra and Calpine plants. TCDC is a hybrid by design and all three covering firms describe it that way; what is moving is the mix. Management is "actively exploring a fully islanded behind-the-meter configuration" under which 650 MW of gross BTM capacity becomes the primary focus and total capacity could reach high-800 MW+ of critical IT load. The schematic now shows two independent paths to power rather than emphasizing the neighbours alone, and Section 03 explains that turbine order size, 460 MW contracted against a 650 MW islanded requirement carrying its own N+1 redundancy, is the cleaner signal than analyst language. New York moratorium added. The nation's first statewide hyperscale moratorium was enacted July 14, 2026, pausing Stream's own ~$19.4B Genesee County campus; Arizona enacted a three-year data center tax-break moratorium. Added to Sections 05 and 11 as a two-sided event, and to the research summary. Stream credentials upgraded. Section 06 now carries the full operating record: 27+ developments since 1999, over 90% of capacity leased to Fortune 100 customers, zero early customer terminations, a ~3.6 GW pipeline with market breakdown, 20+ states, 100% in-house engineering and operations, and a 99.9999% uptime target, with a note reconciling the three different uptime figures now in circulation. Also added. The 20–50 MW modular business line disclosed at the June 23 conference, a month before the website relaunch, strengthening the read that the small-block offering is strategy rather than marketing (Section 02); Northland's July 1 reiteration of $11.00 at $6.38 against a $4.78 close (Section 09); and a worked example of stale sell-side boilerplate: Northland's July 1 risk section still citing the going-concern doubt its own body text described as resolved, alongside a cash-per-share figure struck before the April raise (Section 14).
Report rebuilt against three sell-side initiations and current company communications. Material corrections to the July 22 draft: the FY2025 going-concern qualification was resolved by the April 2026 financings and no longer describes the company; Phase 1 power is front-of-meter via power purchase agreements with adjacent Vistra and Calpine generation, not behind-the-meter only; the October 8, 2026 Macquarie lease covenant was previously absent and is now the organizing fact of the report; capacity figures reflect a phased design (207 / 650 / 1,407 MW) rather than escalating claims; the 54-acre corridor definitive agreement is executed; 460 MW of Siemens turbines are contracted, not merely in procurement; headcount reflects six senior appointments between April and July 2026. Added: three-framework valuation comparison with published sensitivities, the Alphabet capex and third-party-capacity disclosure, coverage-conflict and provenance analysis, and corrected market data.
Initial research report covering business model, revenue streams, industry trends, competitors, financial performance, valuation, growth drivers, risks, scenario analysis, and research summary. Built on public filings and press coverage available as of July 21–22, 2026, prior to review of sell-side initiations.