Stock Analysis · Bulls vs Bears

Soluna Holdings SLNH

🐂 Bulls vs Bears 🐻Report #110 Oct 2026
★ Willem owns this stock · 0.1% of the portfolio
72
high risk
🐂 6vs9 🐻
Bears lead
34%
▲ Chance of the biggest tailwind the forecasts miss
48%
▼ Chance of the biggest headwind the forecasts miss
The verdict

Soluna is a small, cash-burning bitcoin host with real, hard-to-get Texas power positions. Owning it is a bet that one AI lease, financed mostly with project debt, lands before more equity issuance eats the upside. The power assets are real, but the tenant, the money and the management track record are all unproven, so the evidence leans to the cautious side today.

The Bulls vs Bears podcast is being made
Two AI hosts debate this report, NotebookLM-style. It appears here automatically within a few hours.

Written by Claude (Anthropic’s AI) from public sources · Willem owns SLNH · AI can make mistakes — numbers in brackets point to the sources · Not financial advice

🐂 Bull arguments 6

  1. 1Scarce, energised and queue-protected Texas powerMedium
  2. 2Kati 2 is a genuinely advanced AI projectMedium
  3. 3A real cash cushion after the 2026 raisesMedium
  4. 4Bitcoin hosting volume is growing and sites are fullMedium
  5. 5Control of the Dorothy campusWeak
  6. 6Insiders bought shares in the open marketWeak

🐻 Bear arguments 9

  1. 1Relentless dilution, with more authorisedHigh
  2. 2The AI pivot is unproven: no lease, no money, partial power, and a failed first attemptHigh
  3. 3The core business barely covers its overhead and moves with bitcoinHigh
  4. 4Customer concentration on short contractsMedium
  5. 5Debt covenants and legacy liabilitiesMedium
  6. 6The Briscoe wind farm adds merchant power riskMedium
  7. 7Management credibility and governanceMedium
  8. 8Texas rules are tightening for co-located data centersLow
  9. 9Nasdaq $1 minimum bid risk is backLow

Bull arguments show how well they survive the devil’s advocate (strong / medium / weak); bear arguments show the damage they could do after the counter-argument (high / medium / low).

Bull case — all positive points (most important first)

🐂 Bull #1Scarce, energised and queue-protected Texas power

  • The point: AI developers can't get grid power fast. Soluna has power running today, plus queue positions ERCOT has protected.
  • Evidence:
    • 206 MW is operating [9].
    • ERCOT conditionally classed Kati (166 MW incl. Kati 2) and Dorothy (100 MW) as Base Load in Batch Zero. Their capacity is kept "on the basis of prior study work" [9,10].
    • The ERCOT queue held >438 GW of large-load requests by mid-2026, ~89% data centers [31].
    • Only ~2.2 GW was approved to energise in a 12-month period [41].
  • Why it matters: queue position plus land plus nearby generation is the thing AI tenants pay for. Peers with AI leases trade at ~12.9x enterprise value to next-year sales, versus 3.7x for miners without [32].
  • 😈 Devil's advocate:
    • Both classifications are "conditional" and can be removed if ERCOT's checks aren't satisfied [9].
    • Kati 2 gets only 83 MW of protected interconnection, against 100 MW of computing load in Phase I and 350 MW total. Phase II needs new power and substation work [2,9].
    • Texas now offers "bring your own generation" routes to anyone, which narrows the edge [41].
  • Verdict: Medium The power position is real but smaller than the marketing, and it is not yet monetised.

🐂 Bull #2Kati 2 is a genuinely advanced AI project

  • The point: One signed lease would transform a ~$30M-revenue company.
  • Evidence:
    • The JV with Metrobloks was signed 2026-06-03. Soluna controls it and gets capital back plus a 14% return before profits are split [16,2].
    • A general contractor is onboarded, design is ~50% done, long-lead equipment letters are signed, and the tenant letter of intent is exclusive [2].
    • Rosenblatt (Buy, $4, 2026-09-17) expects a first contract "before year-end" and sizes 118 MW at >$200M of yearly revenue [24].
    • H.C. Wainwright (Buy, $4) also cites an AI deal as the catalyst [25].
  • Why it matters: a 10–15-year lease with a creditworthy tenant would let Soluna borrow against the rent, like Cipher and Applied Digital did at 6–7% coupons [33,34]. That would replace dilutive equity.
  • 😈 Devil's advocate:
    • H.C. Wainwright flagged a deal for "later in summer 2026". That window passed with nothing signed [25].
    • The 10-Q gives "no assurance" of a lease [2].
    • The tenant's identity and credit are unknown.
    • Soluna's illustrative economics assume a ~$1.2–1.3B build for ~100 MW with ~$250M of Soluna equity. That is more than twice its June cash [3]. Even a signed lease therefore implies heavy further raising.
    • H.C. Wainwright is also the company's paid share-sales agent [2].
  • Verdict: Medium It is the whole thesis, but it is one counterparty away from either outcome.

🐂 Bull #3A real cash cushion after the 2026 raises

  • The point: Soluna is no longer weeks from running out of money.
  • Evidence:
    • $113.4M of cash at 2026-06-30 (vs $76.4M at year-end), and $23.6M more of share sales after quarter-end [2].
    • On 2026-08-10 it prepaid $19.1M of Generate debt, including a $3.9M penalty [2].
    • The going-concern language in the FY25 10-K is not repeated in the Q2 10-Q [2].
    • Interpretation: net cash, excluding the HPE liability and preferred dividends owed, is roughly $100M. Against a ~$220M market value, that implies the business and pipeline are valued at ~$120M.
  • Why it matters: it pays for Kati 2 design, land and pre-lease work without a fire sale.
  • 😈 Devil's advocate:
    • Operating cash burn was $11.6M in H1, and stock pay was $19.7M [2].
    • The cash was raised by more than doubling the share count.
    • It still covers less than half of Soluna's illustrative equity cheque for one 100 MW AI phase [3].
  • Verdict: Medium It buys time, not the AI build.

🐂 Bull #4Bitcoin hosting volume is growing and sites are full

  • The point: the base business has real demand and executes.
  • Evidence:
    • Hosted hashrate was 4,944 PH/s in Q2, versus 1,807 a year earlier (+174%) [2].
    • Kati 1 is fully subscribed: Galaxy 48 MW, Bitdeer 35 MW [9,12].
    • Revenue has grown for five straight quarters, and Kati 1 made its first gross profit in Q2 [40].
  • Why it matters: it provides a cash-flow base and a reference list of large customers for AI tenants.
  • 😈 Devil's advocate:
    • Volume growth hasn't produced profit (Q2 gross profit $0.77M).
    • Co-mining shifts more bitcoin-price risk onto Soluna.
    • Listed miners' cash cost ($75.5k/BTC) was above the realised price ($72k) in Q2'26, so customer credit quality is weakening [32].
  • Verdict: Medium Weak. The growth is real, but the economics are poor.

🐂 Bull #5Control of the Dorothy campus

  • The point: Soluna now owns the wind farm, 100% of D1A and D1B (bought out its partners in April–May 2026) and 397 acres for Dorothy 3 [2,13,14].
  • Evidence: the Spring Lane stake in D1A was bought for $16.5M and the Navitas 49% of D1B for ~$8.8M [2,14]. Management says full ownership is needed to convert the campus to AI [23].
  • Why it matters: an AI tenant wants one counterparty that controls land, power and site.
  • 😈 Devil's advocate:
    • Briscoe lost money in its first quarter: a gross loss of $0.8M with ~$1.5M of turbine repairs [2].
    • It sits in the Panhandle, where the ERCOT market monitor says prices are "often" negative because of wind congestion [29].
    • Dorothy 3 is at the environmental-study stage [2].
  • Verdict: Weak Medium. It is strategically sensible, but so far it costs money.

🐂 Bull #6Insiders bought shares in the open market

  • The point: the CEO bought 50,000 shares at $1.20 (2026-08-26). The CFO bought 150,000 shares between May and August 2026 at $1.13–1.63 [22].
  • Evidence: Form 4s, code "P", which means open-market purchases [22].
  • Why it matters: it signals that management believes in the value at ~$1.
  • 😈 Devil's advocate: together the purchases are ~$0.26M. The same people received millions of free restricted shares on 2026-06-01 (CEO 3.0M, Executive Chairman 7.3M) [22]. The buying is symbolic.
  • Verdict: Weak

Bear case — all negative points (most important first)

🐻 Bear #1Relentless dilution, with more authorised

  • The point: equity sales pay for growth, and the price per share keeps falling.
  • Evidence:
    • Shares outstanding were 102.5M at 2025-12-31, 225.8M at 2026-06-30 and 244.6M at 2026-08-10 [2].
    • H1 share sales raised $113.5M net, at roughly $1.53 per share. After quarter-end sales were at roughly $1.26 [2] (our arithmetic from filed proceeds).
    • There are 24.3M warrants and 2.4M restricted share units on top [2].
    • A $500M share-sales registration and a $250M Yorkville equity line remain available [2].
    • The proxy asks to raise authorised shares from 375M to 1B on 2026-10-16. It states the purpose is "future financing options" [21].
  • Why it matters: even a successful AI lease may leave each share with a small slice of the value.
  • 🛡️ Counter-argument: management says it will fund AI builds 70–80% with project debt once a lease is signed [23]. If so, the equity need per project is limited. Raising at ~$1.50 to repay 15%+ debt and buy out partners was arguably value-accretive.
  • Severity: High Nothing stops further issuance before a lease exists, and the stock ($0.90) is below every recent raise price.

🐻 Bear #2The AI pivot is unproven: no lease, no money, partial power, and a failed first attempt

  • The point: the thesis rests on Kati 2, which lacks a tenant, financing and its full grid connection.
  • Evidence:
    • No lease: the deck says "Soluna does not currently have a lease in place" [3].
    • Third-party financing "has not been obtained" [2].
    • Only 83 MW of Kati 2's connection is Base-Load protected [9], against 100 MW of computing load in Phase I and 350 MW planned [2].
    • Soluna's 2024 GPU cloud with HPE ended in 12 months, with HPE terminating "for cause" and a $19.3M liability [2].
  • Why it matters: without a lease, Kati 2 and Dorothy 3 are land and paperwork, and the company stays a sub-scale bitcoin host.
  • 🛡️ Counter-argument:
    • The Kati 2 model is different: a lease-only landlord, not a GPU operator. It has a hyperscale-experienced development chief, and there are many tenants seeking power.
    • Peers signed similar leases in 2025–26 [33,34].
    • The 83 vs 100 MW gap may reflect different units: protected grid power versus marketed computing load, or the planned gas firming. That is not confirmed.
  • Severity: High Execution risk is high on every one of the three legs: tenant, money and power.

🐻 Bear #3The core business barely covers its overhead and moves with bitcoin

  • The point: after six years, bitcoin hosting still doesn't make money at group level.
  • Evidence:
    • Q2'26 gross profit was $0.77M and adjusted EBITDA −$1.6M [2,4].
    • Q2 G&A excluding depreciation was $15.2M, up from $5.4M a year earlier [2].
    • Average hashprice fell 34% ($51.09 → $33.63) [2].
    • The halving is expected in April 2028; it cuts bitcoin miners' income per block in half [2].
  • Why it matters: the base business can't fund the AI plan, and a bitcoin downturn hits revenue directly via profit-share and co-mining.
  • 🛡️ Counter-argument:
    • Some Q2 costs were one-off: $1.5M of Briscoe repairs and a $4.2M debt-repayment loss.
    • Hashprice recovered from a record-low monthly average of ~$27.7 in June 2026 to ~$38 by September [32].
  • Severity: High

🐻 Bear #4Customer concentration on short contracts

  • The point: four unnamed customers are 80% of hosting revenue, on 12–24-month contracts [1,2].
  • Evidence: In FY24 the biggest customer (28% of revenue) left in Q4 and was replaced only by March 2025 [1].
  • Why it matters: one departure or failure could erase a large share of revenue. Weak miners are a credit risk (cash cost above BTC price in Q2'26 [32]).
  • 🛡️ Counter-argument: customers are among the strongest names in mining (Galaxy, Bitdeer), and sites are full.
  • Severity: Medium

🐻 Bear #5Debt covenants and legacy liabilities

  • The point: the balance sheet has several hidden claims ahead of common shareholders.
  • Evidence:
    • The Generate loan has a "forward contracted DSCR" test. It compares contracted revenue to debt payments and must stay at or above 1.20x. Management says missing it is "probable" at test dates in the next 12 months. The June waiver required a $19.1M prepayment [2].
    • The $19.3M HPE liability sits in a subsidiary with "no remaining assets" [2].
    • The $6.0M Green Cloud note carries a parent guarantee [3].
    • $35.1M of Series A preferred dividends had not been paid as of 2026-06-30, and they are still accruing at ~$2.8M per quarter [2].
  • Why it matters: a covenant default can trigger cross-defaults. The preferred arrears rank ahead of common in any sale or wind-up.
  • 🛡️ Counter-argument: remaining Generate debt is small (~$9.6M) and could be repaid from cash. Interpretation: HPE's claim seems confined to a subsidiary with no assets, though the full guarantee terms were not checked.
  • Severity: Medium

🐻 Bear #6The Briscoe wind farm adds merchant power risk

  • The point: Soluna bought a 150 MW wind farm for $53M. That brought negative-price, repair and decommissioning risk onto a small balance sheet.
  • Evidence:
    • 10 of 81 turbines were down at closing, with ~$1.5M of Q2 repairs and a Q2 gross loss of $0.8M [2].
    • The 10-Q notes negative nodal pricing from "severe congestion" in ERCOT West [2].
    • The ERCOT market monitor confirms frequent negative prices in the Panhandle [29].
    • West Texas transmission relief (765 kV lines) is targeted only around 2030 [38].
  • Why it matters: it is a single asset that must earn the $6–11M of yearly EBITDA management promised, or it will drain cash.
  • 🛡️ Counter-argument: most of Briscoe's output feeds Soluna's own Dorothy sites, which turns negative-price hours into cheap power for hosting.
  • Severity: Medium

🐻 Bear #7Management credibility and governance

  • The point: promises are often missed, pay is high, and controls look loose.
  • Evidence:
    • The HPE cloud failure and the reversal from positive EBITDA in 2024 to negative [7,4].
    • Three CFOs in about two years [1].
    • CEO 2025 pay of $6.0M [1].
    • The 7.3M-share grant to the Executive Chairman [22].
    • A "CONFIDENTIAL DRAFT" header on the exhibit of an SEC-filed press release (2026-09-28) [12].
    • The auditor changed from UHY to KPMG in March 2026, with no disagreements reported [19].
  • Why it matters: the AI plan requires a big, complex financing. Lenders and tenants judge the team.
  • 🛡️ Counter-argument: recent builds (Kati 1) were on time and on budget. A hyperscale-experienced development chief and KPMG as auditor are upgrades.
  • Severity: Medium

🐻 Bear #8Texas rules are tightening for co-located data centers

  • The point: new rules raise the cost and conditions of big new loads.
  • Evidence:
    • SB6 (2025) lets ERCOT curtail large loads in emergencies and requires new loads ≥75 MW to be curtailable [30].
    • The PUCT large-load rule (in force 2026-10-08) requires a $100k study fee and $50k/MW of security [35] (secondary).
    • A July 2026 PUCT order says co-located sites must be able to cut their full load within 30 minutes in grid emergencies, unpaid [36] (secondary).
  • Why it matters: AI tenants want near-100% uptime. Unpaid curtailment duties may reduce the appeal of co-located sites, or force backup generation that adds cost.
  • 🛡️ Counter-argument: Kati 2 already holds conditional Base Load status. Soluna has years of experience running curtailable load, and all Texas competitors face the same rules.
  • Severity: Low Medium.

🐻 Bear #9Nasdaq $1 minimum bid risk is back

  • The point: the stock closed at $0.90 on 2026-10-09 [27].
  • Evidence: Soluna got a deficiency notice on 2026-04-10 and regained compliance by 2026-04-30 [17,18]. It also had episodes in 2023 and 2025 [18].
  • Why it matters: a new notice would come after 30 trading days below $1. Interpretation: a reverse split usually follows and tends to weigh on sentiment for heavily diluted names.
  • 🛡️ Counter-argument: Soluna has cured this three times, and any AI lease headline would likely lift the price above $1.
  • Severity: Low

Tailwinds NOT in the forecasts

Consensus (2 analysts) expects FY26 revenue of ~$61.8M and EPS of −$0.41 [26]. Interpretation: that reflects the Kati 1 ramp, Bitdeer co-mining and Briscoe, not an AI lease. The $4 targets from H.C. Wainwright and Rosenblatt do assume one [24,25].

  • Signed and financed Kati 2 lease (≥100 MW, 10+ years, creditworthy tenant) — 34% chance within 36 months · Impact if it happens: High

    • What it is: a definitive lease plus project debt, so Phase I gets built.
    • Why it's not priced in: consensus revenue has no AI line. At ~$220M market value, the stock trades "like a stressed bitcoin miner" (Rosenblatt) [24].
    • What would trigger it / the sign to watch: an 8-K naming the tenant, then a project-debt term sheet. Watch the Q3 call (~2026-11-12).
    • Why this probability: several large tenants are signing with ex-miners [33,34]. Soluna already has an exclusive letter of intent and a protected queue position. But it must clear three hurdles in a row (lease, ~$1B financing, full power), with a weak balance sheet and an HPE-shaped track record.

  • Bitcoin and hashprice recovery lifts hosting and co-mining — 41% chance that the average hashprice is above $50/PH/day for a full quarter before the April 2028 halving · Impact if it happens: Medium

    • What it is: a higher BTC price raises profit-share and co-mining revenue at little extra cost.
    • Why it's not priced in: consensus is built on 2026 hashprice near record lows [32].
    • Sign to watch: monthly hashprice indices, and BTC versus the ~$75k miner cash cost.
    • Why this probability: hashprice was above $50 as recently as mid-2025, but the network keeps adding machines and the halving cuts income per block.

  • A larger player buys a site or the company for its power positions — 16% chance · Impact if it happens: High

    • What it is: a buyer pays for the queue-protected MW and land: Kati (166 MW Base Load) and Dorothy (100 MW).
    • Why it's not priced in: no one has made a public approach. The analysts' cases are stand-alone.
    • Sign to watch: sale of a stake in Kati 2 or Dorothy 3, strategic-review language, or 13D filings.
    • Why this probability: consolidation among ex-miners is active. But preferred arrears, the many layers of financing and the small scale make a deal complex.

  • Dorothy 3 AI lease — 17% chance within 36 months · Impact if it happens: High

    • What it is: a second AI campus (~300 MW) on owned land, powered by the owned Briscoe wind farm.
    • Why it's not priced in: it is at environmental-study stage, and tenant marketing starts only this autumn [2,23].
    • Sign to watch: a tenant letter of intent and confirmation of the extra 50 MW Dorothy request that the utility is studying [10].
    • Why this probability: it is earlier stage than Kati 2 and depends on Kati 2 proving the model first.

Headwinds NOT in the forecasts

  • AI-build funding pushes the share count above 450M (+85% vs Aug 2026) within 24 months — 48% chance · Impact if it happens: High

    • What it is: equity raised at today's ~$0.90 to fund the ~$250M Soluna equity cheque, land ($19.1M for Kati 2 Phase II) and overhead [2,3].
    • Why it's not priced in: consensus EPS uses the current share count. Analysts' $4 targets assume financing mostly with debt.
    • Sign to watch: passage of the 1B authorised-share vote (2026-10-16), the Q3 10-Q share count and use of the Yorkville equity line.
    • Why this probability: the authorisation and facilities exist, and the company has always raised when it could. A lease could shift funding to debt, so this is close to even.

  • The Kati 2 letter of intent lapses with no lease — 33% chance within 12 months · Impact if it happens: High

    • What it is: the exclusive tenant walks away or exclusivity expires.
    • Why it's not priced in: both covering analysts treat a lease as near-term [24,25].
    • Sign to watch: silence past year-end 2026, a "new prospective tenant" in filings, or changes to Kati 2 scope.
    • Why this probability: the "summer 2026" timing already slipped, and tenants are picky about curtailment rules and partial grid power. But exclusive letters of intent at advanced design stages usually close.

  • Generate covenant trouble forces another costly waiver or prepayment — 52% chance within 12 months · Impact if it happens: Low–Medium

    • What it is: the forward-contracted DSCR test fails again, given short hosting contracts and zero credit for uncontracted revenue [2].
    • Why it's not priced in: it is a footnote issue that consensus does not model.
    • Sign to watch: Q3 10-Q debt note (the 9/30 test date).
    • Why this probability: management itself calls a breach "probable" absent new contracts. Bitdeer's contract may help, and the balance is now small, so the damage is likely a fee, not a crisis.

  • New Nasdaq deficiency and reverse split — 46% chance within 12 months · Impact if it happens: Low

    • What it is: 30 business days below $1 triggers a notice. A reverse split often follows.
    • Why it's not priced in: it is a technical event, but it tends to cause selling and narrative damage.
    • Sign to watch: closing prices below $1 through mid-November, an 8-K Item 3.01, or proxy language.
    • Why this probability: the stock is already at $0.90. An AI headline would likely cure it.

  • A customer failure or exit in bitcoin hosting — 27% chance within 24 months of losing a top-four customer · Impact if it happens: Medium

    • What it is: a large miner leaves or defaults, which is plausible with miner cash costs above the BTC price [32].
    • Why it's not priced in: consensus assumes full sites.
    • Sign to watch: concentration notes, site utilisation and new "co-mining" swaps for fixed-fee deals.
    • Why this probability: contracts are 12–24 months, so churn is constant, but the current customers are among the strongest miners.

Investment Risk Score: 72/100

Risk area Weight Score (0–100) Why
Business & competition 25% 75 Commodity bitcoin hosting with 22% gross margin and no AI revenue. Far larger rivals already hold 10–15-year AI leases.
Financial health & dilution 20% 85 Shares 2.4x in 7 months, and 1B authorised up for vote. Cash burn, $35.1M preferred arrears, a "probable" covenant breach and a $19.3M HPE liability.
Execution (delivering the plan) 20% 72 Kati 1 was delivered on time. But the first AI attempt failed, the plan needs ~$1B+ of financing, and the company has had three CFOs in two years.
Dependency & concentration 15% 74 Four customers are 80% of hosting revenue on 12–24-month contracts. Nearly all MW are in Texas, and Kati relies on one third-party wind farm.
Regulation, policy & geopolitics 10% 55 SB6 emergency curtailment and new PUCT security rules apply. Batch Zero status is only conditional. ASIC and equipment tariffs.
Sentiment & expectations 10% 50 At $0.90 the market expects little. But the $4 targets lean on an AI lease, and the analysts include the company's paid share-sales agent.

Overall = 72/100 (high). Weighted: 18.75 + 17.0 + 14.4 + 11.1 + 5.5 + 5.0 = 71.75. Soluna is a small company that must clear three uncertain hurdles in a row (tenant, ~$1B financing, full power) while funding itself mostly by selling shares below $1. The power positions and the $113M of cash keep the score below "very high". This is the first deep dive, so nothing has changed versus a previous one.

Conclusion

  • What kind of company this is: a sub-scale bitcoin-hosting landlord with real, hard-to-replicate power sites in Texas. Owning it is a call option on one AI lease at Kati 2, financed mostly with debt, arriving before dilution and cash burn erode the option.
  • The 2–3 things that will decide the outcome:
    1. Whether the Kati 2 letter of intent becomes a lease with a creditworthy tenant.
    2. Whether that lease can be financed mostly with project debt rather than shares.
    3. How much more equity is issued before then, which the 1B-share vote enables.
  • Signposts that would confirm the bull case:
    • A named tenant and lease of 10+ years for ≥100 MW.
    • A project-debt commitment covering ≥70% of build cost.
    • ERCOT confirming (not "conditionally") the 166 MW and 100 MW Base Load positions.
    • Q3/Q4 adjusted EBITDA turning positive as Kati 1 and Bitdeer ramp.
  • Signposts that would break it:
    • No lease by Q1 2027, or a change in the Kati 2 tenant or scope.
    • The share count above ~350M before any lease.
    • Another Generate waiver with a costly prepayment, or a default notice.
    • Loss of a top-four hosting customer, or Briscoe staying loss-making after repairs.
  • Balanced verdict:
    • Bull side, strongest version: in a market where AI developers can't get power, Soluna holds 206 MW running and 266 MW of protected queue capacity. It has an exclusive tenant in advanced talks and enough cash to reach a signing. One lease could be worth several times today's ~$220M market value.
    • Bear side, strongest version: Soluna has never turned its power positions into profit. Its last AI venture failed. It funds itself by selling shares at ever-lower prices and is asking for permission to issue up to 1B shares. Even a signed lease needs about twice its current cash in equity alongside the debt.
    • Where the evidence leans: the power assets are genuine. But every value-creating step (lease, financing, full grid connection) is still a promise, and this management has a mixed record on promises. The evidence leans to the bear side until a lease is signed.

Open questions & data gaps

  • The tenant's identity, credit quality, rent per MW and lease length: not found. These matter most, because they decide whether project debt is possible.
  • The 83 MW protected connection vs 100 MW Phase I computing load (350 MW total) at Kati 2: not reconciled. It decides how much of Kati 2 can actually be powered from the grid.
  • No verbatim earnings call transcripts were available. Call quotes come from an AI-assisted Investing.com account and aggregator summaries. Prior-call promises could not be checked word for word.
  • All-in power cost ($/kWh) per site and hosting margin by site: not disclosed, so the "lowest-cost" claim is unverified.
  • The top-four customer identities and contract renewal dates: not disclosed.
  • Share sales since 2026-08-10, and why the stock fell from ~$1.20 to $0.90 in September–October: not found. The Q3 10-Q will show the share count.
  • Whether the Briscoe/Dorothy setup (Soluna owns both the generator and the load) needs PUCT approval under SB6: not determined. It needs a legal reading.
  • Full terms of the Green Cloud parent guarantee, and HPE's collection efforts: only partly read.
Sources
  1. Soluna Holdings Form 10-K FY2025 (Item 1, 1A, Part III, going-concern CAM) — SEC EDGAR — 2026-03-30 — https://www.sec.gov/Archives/edgar/data/64463/000162828026021847/slnh-20251231.htm
  2. Soluna Holdings Form 10-Q Q2 2026 (MD&A, debt, VIE, equity, risk factors) — SEC EDGAR — 2026-08-13 — https://www.sec.gov/Archives/edgar/data/64463/000162828026056702/slnh-20260630.htm
  3. Q2 2026 Highlights and Business Update (investor deck) — Soluna — 2026-08 — https://www.solunacomputing.com/wp-content/uploads/2026/08/SLNH-Q2-2026-Review-and-Business-Update.pdf
  4. Soluna Q2'26 results press release (8-K Ex. 99.1) — SEC EDGAR — 2026-08-13 — https://www.sec.gov/Archives/edgar/data/64463/000149315226037734/ex99-1.htm
  5. Soluna Q1'26 results press release (8-K Ex. 99.1) — SEC EDGAR — 2026-05-18 — https://www.sec.gov/Archives/edgar/data/0000064463/000149315226024060/ex99-1.htm
  6. Soluna FY2025 results press release (8-K Ex. 99.1) — SEC EDGAR — 2026-03-30 — https://www.sec.gov/Archives/edgar/data/0000064463/000149315226013456/ex99-1.htm
  7. Soluna Q1'24 results press release (8-K Ex. 99.1) — SEC EDGAR — 2024-05-16 — https://www.sec.gov/Archives/edgar/data/0000064463/000149315224021002/ex99-1.htm
  8. Soluna Holdings Form 10-K FY2024 — SEC EDGAR — 2025-03-31 — https://www.sec.gov/Archives/edgar/data/64463/000164117225001756/form10-k.htm
  9. Soluna Finishes Kati 1 and Secures Conditional Base Load Position for All 166 MW in ERCOT (8-K Ex. 99.1) — SEC EDGAR — 2026-09-08 — https://www.sec.gov/Archives/edgar/data/0000064463/000149315226041699/ex99-1.htm
  10. Soluna Secures Conditional Base Load Position for Project Dorothy's 100 MW (8-K Ex. 99.1) — SEC EDGAR — 2026-09-14 — https://www.sec.gov/Archives/edgar/data/0000064463/000149315226042479/ex99-1.htm
  11. Soluna and Bitdeer Enter 28 MW Agreement at Project Kati 1 (8-K Ex. 99.1) — SEC EDGAR — 2026-08-25 — https://www.sec.gov/Archives/edgar/data/0000064463/000149315226039964/ex99-1.htm
  12. Soluna and Bitdeer Expand Agreement to 35 MW at Project Kati 1 (8-K Ex. 99.1) — SEC EDGAR — 2026-09-28 — https://www.sec.gov/Archives/edgar/data/0000064463/000149315226044444/ex99-1.htm
  13. Soluna Closes Purchase of 397 Acres for Project Dorothy 3 (8-K Ex. 99.1) — SEC EDGAR — 2026-08-24 — https://www.sec.gov/Archives/edgar/data/0000064463/000149315226039781/ex99-1.htm
  14. Soluna Acquires Full Ownership of Project Dorothy 1B (8-K) — SEC EDGAR — 2026-05-20 — https://www.sec.gov/Archives/edgar/data/0000064463/000149315226024511/ex99-1.htm
  15. 8-K: YA II promissory note, warrants, Dorothy 1A buyout — SEC EDGAR — 2026-04-17 — https://www.sec.gov/Archives/edgar/data/64463/000149315226017877/form8-k.htm
  16. 8-K: Kati 2 JV with Metrobloks — SEC EDGAR — 2026-06-09 — https://www.sec.gov/Archives/edgar/data/64463/000149315226027852/form8-k.htm
  17. 8-K Item 3.01: Nasdaq bid-price deficiency notice — SEC EDGAR — 2026-04-15 — https://www.sec.gov/Archives/edgar/data/64463/000149315226016620/form8-k.htm
  18. 8-K Item 8.01: Nasdaq compliance regained — SEC EDGAR — 2026-05-01 — https://www.sec.gov/Archives/edgar/data/64463/000149315226020814/form8-k.htm
  19. 8-K Item 4.01: auditor change (UHY → KPMG) — SEC EDGAR — 2026-03-31 — https://www.sec.gov/Archives/edgar/data/64463/000149315226013710/form8-k.htm
  20. 8-K: CEO compensation / CDO appointment — SEC EDGAR — 2026-07-24 — https://www.sec.gov/Archives/edgar/data/64463/000149315226034600/form8-k.htm
  21. DEF 14A: 2026 annual meeting proxy (authorised share increase to 1B) — SEC EDGAR — 2026-08-24 — https://www.sec.gov/Archives/edgar/data/64463/000162828026058611/slnh-20260824.htm
  22. Form 4 filings, insiders, 2026-04-08 to 2026-09-24 — SEC EDGAR — https://data.sec.gov/submissions/CIK0000064463.json
  23. Earnings call transcript: Soluna Q2 2026 (AI-assisted account, not verbatim) — Investing.com — 2026-08-13 — https://www.investing.com/news/transcripts/earnings-call-transcript-soluna-shares-fall-after-q2-2026-revenue-surge-93CH-4859493
  24. Rosenblatt initiates Soluna stock with Buy rating on HPC pivot (note itself not read) — Investing.com — 2026-09-17 — https://www.investing.com/news/analyst-ratings/rosenblatt-initiates-soluna-stock-with-buy-rating-on-hpc-pivot-93CH-4904924
  25. H.C. Wainwright initiates Soluna stock with Buy rating, $4 target (note itself not read) — Investing.com — 2026-07-20 — https://www.investing.com/news/analyst-ratings/hc-wainwright-initiates-soluna-stock-with-buy-rating-4-target-93CH-4800673
  26. SLNH forecast and consensus — StockAnalysis — 2026-09-17 — https://stockanalysis.com/stocks/slnh/forecast/
  27. SLNH quote (close 2026-10-09, market cap) — StockAnalysis — 2026-10-09 — https://stockanalysis.com/stocks/slnh/
  28. Soluna analyst ratings / short interest — MarketBeat — 2026-10-09 — https://www.marketbeat.com/instant-alerts/soluna-nasdaqslnh-receives-buy-rating-from-hc-wainwright-2026-07-20/
  29. 2024 State of the Market Report for the ERCOT Electricity Markets — Potomac Economics (ERCOT IMM) — 2025-06 — https://potomaceconomics.com/wp-content/uploads/2025/06/2024-State-of-the-Market-Report.pdf
  30. Texas SB 6 (89R) enrolled bill text — Texas Legislature — 2025 — https://capitol.texas.gov/tlodocs/89R/billtext/html/SB00006F.htm
  31. Can ERCOT's new Batch Zero process alleviate the burden of 438,000 MW of large load requests? — Power Engineering — 2026-06-29 — https://www.power-eng.com/business/policy-and-regulation/can-ercots-new-batch-zero-process-alleviate-the-burden-of-438000-mw-of-large-load-requests/
  32. Bitcoin Mining Report Q2 2026 — CoinShares — 2026-09-15 — https://coinshares.com/corp/insights/research-data/bitcoin-mining-report-q2-2026/
  33. Cipher Digital business update (8-K Ex. 99.2) — SEC EDGAR — 2026-08-04 — https://www.sec.gov/Archives/edgar/data/0001819989/000181998926000038/cipherdigital_businessup.htm
  34. Applied Digital Reports Fiscal Fourth Quarter and Full Year 2026 Results (read via summary) — Applied Digital — 2026-07-27 — https://ir.applieddigital.com/_assets/_304b68447a593477100d9f303faa80fa/appliedblockchaininc/news/2026-07-27_Applied_Digital_Reports_Fiscal_Fourth_Quarter_and__159.pdf
  35. Texas adopts large load grid rules with a $100,000 study fee and $50,000 per MW in security (secondary) — Colocation Scout — 2026-09/10 — https://colocationscout.com/news-texas-puc-large-load-interconnection-rule.html
  36. Texas SB6's Real Test: What ERCOT's New Curtailment Order Means for Co-Located Data Centers (secondary; White & Case alert not readable) — Verse — 2026 — https://verse.inc/blog/texas-sb6-curtailment-rules-ercot-data-centers
  37. ERCOT curtailments persist as load rises — FactSet Insight — 2025-02-25 — https://insight.factset.com/ercot-curtailments-persist-as-load-rises
  38. Texas PUC approves first 765-kV lines despite opposition (date not confirmed) — Texas Policy Research — 2025/2026 — https://www.texaspolicyresearch.com/texas-puc-approves-first-765-kv-lines-despite-opposition/
  39. Navigating safe harbor rules for solar and wind Sec. 48E facilities — The Tax Adviser — 2026-02 — https://www.thetaxadviser.com/issues/2026/feb/navigating-safe-harbor-rules-for-solar-and-wind-sec-48e-facilities/
  40. Soluna Q2 2026 earnings call summary (secondary) — Stock Taper — 2026-08 — https://www.stocktaper.com/earningsCallSummary/SLNH/2026/Q2
  41. ERCOT LLWG March 2026: Large Load Queue Surges to 380 GW, Batch Zero & CLR/BYOG Frameworks (secondary) — Zero Emission Grid — 2026-03-18 — https://www.zeroemissiongrid.com/iso-rto/ercot-llwg-3-13/
  42. Miners beat bitcoin by 70% in 2026 as TeraWulf locks $12.8B in AI contracts (secondary) — Bitcoin.com — 2026 — https://news.bitcoin.com/miners-beat-bitcoin-by-70-in-2026-as-terawulf-locks-12-8b-in-ai-contracts/
Disclaimer. AI-generated research and podcast (Claude by Anthropic, voices by Google Gemini) for educational purposes — not financial advice and not a recommendation to buy or sell. Willem (MR.INVEST) owns this stock. Check the company’s own filings before making any decision.
← All Bulls vs Bears reports