Stock Analysis · Bulls vs Bears

Keel Infrastructure KEEL

🐂 Bulls vs Bears 🐻Report #110 Oct 2026
★ Willem owns this stock · 0.9% of the portfolio
63
high risk
🐂 6vs8 🐻
Bears lead
38%
▲ Chance of the biggest tailwind the forecasts miss
55%
▼ Chance of the biggest headwind the forecasts miss
The verdict

Keel is now a bet on power and permits. It owns rare, grid-connected Pennsylvania capacity, and enough cash to wait, but it has no customer, no building experience and an unpriced bill to build. The stock (~$3.05, down ~58% from its 52-week high) prices in serious doubt, so the outcome hinges almost entirely on whether and on what terms a creditworthy tenant signs at Panther Creek or Sharon over the next 6–12 months.

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 KEEL · AI can make mistakes — numbers in brackets point to the sources · Not financial advice

🐂 Bull arguments 6

  1. 1Firm 2027 power inside PJM, the US's tightest big gridStrong
  2. 2Balance sheet buys patienceMedium
  3. 3Peer leases show large value per MW, and the stock prices little of itMedium
  4. 4The team is being upgraded for the new businessMedium
  5. 5Large expansion options in Pennsylvania and QuébecWeak
  6. 6Insiders bought on the dip, and expectations are lowWeak

🐻 Bear arguments 8

  1. 1No tenant, and the timeline to sign is slippingHigh
  2. 2Large, unquantified funding need, and dilutionHigh
  3. 3No building track record, and dates keep slippingMedium
  4. 4Heavy concentration in Pennsylvania, where the rules are changingMedium
  5. 5A shrinking legacy business is burning cashMedium
  6. 6The AI financing cycle could turn against late signersMedium
  7. 7Québec policy is turning hostileLow
  8. 8Legacy governance and litigation baggageLow

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 #1Firm 2027 power inside PJM, the US's tightest big grid

  • The point: AI tenants pay for "time-to-power". Keel already holds utility contracts for 2027 delivery that newcomers can't get quickly.
  • Evidence:
    • Secured capacity: 648 MW (Panther Creek 350 MW under a PPL energy service agreement, Sharon 110 MW, Moses Lake 18 MW, Québec 170 MW). It is $40M into Panther Creek collateral [2].
    • PJM's 2028/29 capacity auction (14 July 2026) cleared at the $325/MW-day cap. It came in 6,831 MW short of the reliability requirement, and only 525 MW of new generation cleared [16].
    • Grid connection queues: power transformers face multi-year waits (average 128 weeks, up to 4 years) [36]. Keel already has Sharon's transformers on site [7].
  • Why it matters: Scarcity of firm power is what lets a landlord charge for speed and pick its tenant.
  • 😈 Devil's advocate: Large loads may soon be throttled in PJM. A 2026 "connect and manage" proposal would let the grid curtail big loads that don't bring their own generation [17]. Pennsylvania's August 2026 executive order adds consent-order steps for projects over 25 MW [13]. The firm-ness of "firm" power could erode, and hyperscalers are also building behind-the-meter gas plants.
  • Verdict: Strong The capacity is contracted and documented. Grid rules are a risk to watch, not a refutation.

🐂 Bull #2Balance sheet buys patience

  • The point: Unlike many small developers, Keel doesn't have to sign the first lease offered.
  • Evidence:
    • ~$698M unrestricted cash plus ~$121M bitcoin at 7 August [3].
    • Converts: $588M at 1.375% due January 2031 and $458M at 1.25% due January 2032. Neither is callable before 2028 and 2029 respectively [2,11,12].
    • Management says cash funds development through lease signing, plus ~$100M a year of cash overhead through 2028 [7].
  • Why it matters: Patience should translate into better credit and terms, and lower-cost project debt.
  • 😈 Devil's advocate: The cash covers overhead and pre-construction, not the build. H1 2026 operating cash outflow was $117.6M [2]. Equity book value is only $328.7M against ~$1.03B of debt [2]. The patience is real but finite, and the convert holders sit ahead of shareholders.
  • Verdict: Medium Strong. It removes near-term survival risk, not funding risk.

🐂 Bull #3Peer leases show large value per MW, and the stock prices little of it

  • The point: If Keel signs leases on terms like its peers', the contracted value would far exceed today's market value.
  • Evidence:
    • Hut 8–Fluidstack: 245 MW, 15 years, $7.0B base, Google-backstopped, with up to 85% loan-to-cost financing [22].
    • Riot: 191 MW, 20 years, ~$9.1B (summary only) [23].
    • Peers' shares jumped about 8–35% on signing days (secondary).
    • Citizens JMP: Outperform, $10 target (24 June). It says the full 2.2 GW pipeline could be worth >$44/share if fully leased to investment-grade tenants [28].
    • Jones Trading: Buy, $7.50 (28 September). It says a signed contract would "likely re-rate" the stock [27].
  • Why it matters: Keel has ~2.6× Hut 8's River Bend MW in secured capacity. Interpretation: a market value of ~$1.9B (≈ $3.05 × 617.6M shares) means investors assign little to the 1.5 GW of expansion capacity.
  • 😈 Devil's advocate: Peers' values sit on signed credit; Keel's sits on hope. The $44/share figure assumes the 1.36 GW Scrubgrass site, which has no utility agreement, no permits and an unbuilt gas plant [2,9]. Later-signing landlords may face tighter terms as AI financing gets more expensive [34].
  • Verdict: Medium Real optionality, but it depends entirely on signing.

🐂 Bull #4The team is being upgraded for the new business

  • The point: The weakest link, no data-center operating know-how, is being addressed.
  • Evidence:
    • Ganesh Aiyer, ~7 years as Digital Realty's Chief Business Officer, became President on 6 July 2026 [2,7].
    • CFO Mir brings power and infrastructure financing experience from BofA and Lazard [1].
    • Execution on the exits was clean: the Latin America sale and the US mining shutdown [2].
  • Why it matters: Tenants and lenders underwrite the developer as well as the site.
  • 😈 Devil's advocate: One senior hire doesn't create a construction track record. The 10-K's "limited or no experience" warning still stands, and every HPC date has slipped [1,7].
  • Verdict: Medium

🐂 Bull #5Large expansion options in Pennsylvania and Québec

  • The point: Beyond the 648 MW, Keel has applications for ~2 GW of extra Pennsylvania power, plus Québec sites.
  • Evidence:
    • Panther Creek could reach 400–430 MW by converting 60 MW of interruptible service, with >500 MW possible long term [6].
    • Scrubgrass has a 750 MW utility load study and a concept for 550 MW of on-site gas owned by an independent power producer [9].
    • Sherbrooke 96 MW: city approval came 15 July 2026, and provincial approval is pending [2].
    • An update on the Pennsylvania power applications is promised for December or January [7].
  • Why it matters: A first lease would prove the model, and expansion would then multiply it on land and interconnections Keel already controls.
  • 😈 Devil's advocate: Expansion MW have no contracts. Pennsylvania's new rules can make large new loads pay "but for" upgrade costs, meaning the grid upgrades that would not be needed without them [19]. Québec is moving the other way: Hydro-Québec has asked the Régie de l'énergie (Québec's energy regulator) for a data-center rate roughly double today's large-power rate [20].
  • Verdict: Weak Medium. Real option value, but distant and policy-dependent.

🐂 Bull #6Insiders bought on the dip, and expectations are low

  • The point: Management bought stock in the open market after the Q2 drop, and the share price already reflects disappointment.
  • Evidence:
    • CEO Gagnon bought 58,888 and 38,888 shares at $3.33. COO Wilson bought 26,472 at $3.78 and 30,769 at $3.25. All were open-market buys between 13 and 21 August 2026 [31].
    • No insider open-market sales were found. A July 2026 Form 4 was an RSU vesting, not a sale [32].
    • The stock is ~$3.05 against a 52-week high of $7.37 [29,41]. Short interest is ~17–21% of the free float per aggregators (low-quality data).
  • Why it matters: It signals some insider conviction, and good news would land on a sceptical base.
  • 😈 Devil's advocate: About $0.55M of buying is tiny next to a ~$1.9B company. Low expectations can stay low for a long time when there is no catalyst.
  • Verdict: Weak

Bear case — all negative points (most important first)

🐻 Bear #1No tenant, and the timeline to sign is slipping

  • The point: The entire equity story waits on a lease that hasn't come. The first wave of hyperscaler and AI-lab demand has already gone to peers.
  • Evidence:
    • No lease, LOI (letter of intent) or named tenant exists [1,2].
    • The Q1 call said "sign 3 leases by year-end" [6]. On the Q2 call the CEO declined to restate it [7,8]. B. Riley reportedly said the priority had been "dropped" (secondary; note not read).
    • Signed peers: IREN–Microsoft, Cipher–AWS and Fluidstack, TeraWulf–Fluidstack, Hut 8–Fluidstack, Core Scientific–CoreWeave, Riot and CleanSpark [22]–[26].
  • Why it matters: Each month without a lease delays revenue, and the pre-lease cost keeps running. The 2027 power slots lose value if the halls can't be ready.
  • 🛡️ Counter-argument: Keel only got its capacity table "secured" in 2026, and the sites were still being permitted. Demand is still running ahead of supply (PJM shortfall, CBRE record-low vacancy). Holding out for investment-grade credit is a defensible choice.
  • Severity: High Without a lease there is no business model, only a power-sales shell.

🐻 Bear #2Large, unquantified funding need, and dilution

  • The point: Turning 478 MW in the US (Panther Creek, Sharon, Moses Lake) into AI halls costs multiples of today's cash. How it will be financed is unknown.
  • Evidence:
    • Committed HPC capex is only $219.6M so far [2]. Cost per MW and total project cost: not found. The CFO told analysts to use "industry-average" costs [9].
    • The 10-K says construction will be funded with project or parent debt and "equity-linked sources", which "could dilute" shareholders [1].
    • Converts: ~147.5M potential shares at $6.86 and $7.41 conversion prices, about 24% of the 617.6M shares outstanding. The capped calls only offset dilution between those prices and ~$11.9 [2].
    • The earlier ATM sold 165.1M shares for $375M gross, completed October 2025 [2]. Management said in May it "may put a credit line and/or an ATM in place this year" [6].
    • The share count rose from 602.85M (27 March) to 617.57M (7 August) [1,2]. The cause of the ~14.7M increase was not verified; the 10-Q equity note shows only ~3.4M from share awards and RSU/PSU settlements in H1.
  • Why it matters: Even a great lease can be worth little per share if the build is funded with cheap stock.
  • 🛡️ Counter-argument:
    • With a Google-style backstop, peers raised 80–85% of build cost as project debt [22].
    • Keel's $819M could cover much of the equity slice for the first site.
    • Convert dilution only happens above ~$6.86, which would itself be a good outcome.
  • Severity: High The counter works only after an investment-grade lease, which loops back to Bear #1.

🐻 Bear #3No building track record, and dates keep slipping

  • The point: Keel has never delivered an AI data center. Tenants impose penalties and termination rights for late delivery.
  • Evidence:
    • The 10-K warns of "limited or no experience in the digital infrastructure space", delayed-delivery termination rights and uptime penalties [1].
    • The slips are listed in section 1 [1,7,10].
    • DEP permits for Panther Creek and Sharon are running "a few months" late, blamed on DEP's eastern-region backlog [7].
  • Why it matters: A lease signed late in 2026 for end-2027 delivery leaves little slack. A missed date could cost penalties or the lease.
  • 🛡️ Counter-argument:
    • Moses Lake uses a Vertiv turnkey modular design, which reduces construction risk [15].
    • Aiyer joined from Digital Realty [2].
    • Management says the permit slip doesn't change the utility's power-delivery schedule [7].
  • Severity: Medium High.

🐻 Bear #4Heavy concentration in Pennsylvania, where the rules are changing

  • The point: Most of the near-term value sits in two Pennsylvania sites under one state's policy and one grid's rules.
  • Evidence:
    • Northwise Project counts three Pennsylvania campuses as 860 of 974 MW in its 2030 base case, ~88% (secondary) [33].
    • Governor Shapiro's Executive Order 2026-05 (18 August 2026) requires a project-specific consent order and local approval for data centers over 25 MW, and removes fast-track permitting. Keel says its sites are "unaffected" (company statement only) [13].
    • The PA PUC's model large-load tariff (13 May 2026) recommends that big new loads pay for the grid upgrades they cause [19].
    • The PJM curtailment and large-load registry proposal is with FERC, the US federal energy regulator. It is pending [17,18].
  • Why it matters: One adverse ruling can delay or raise the cost of most of the near-term value at once.
  • 🛡️ Counter-argument: Keel's 350 MW and 110 MW are already under utility agreements and zoned. New rules mostly target future applicants, though grandfathering is not confirmed: not found.
  • Severity: Medium

🐻 Bear #5A shrinking legacy business is burning cash

  • The point: Until rent arrives, revenue falls and overhead rises.
  • Evidence:
    • Q2 2026: revenue $30.4M (−50% year on year), adjusted EBITDA −$23.7M, and G&A $31.3M against $19.4M a year earlier [3].
    • H1 operating cash outflow was $117.6M [2].
    • Cash overhead is guided at ~$100M a year [7].
    • Bitcoin sales are funding operations [3].
  • Why it matters: At ~$100M+ of burn a year plus pre-construction spend, the ~$819M cushion shrinks by the time a lease arrives, especially if bitcoin falls.
  • 🛡️ Counter-argument: Management's runway claim covers this through 2028 [7]. The 10-Q flags no going-concern doubt [2].
  • Severity: Medium

🐻 Bear #6The AI financing cycle could turn against late signers

  • The point: Keel's tenants will mostly be AI clouds whose own funding depends on debt markets.
  • Evidence:
    • On 8 October 2026 "financed AI build-out" names sold off: CoreWeave −7%, Nebius −6%, Oracle −5%. Long-term US rates were ~4.8% (secondary) [34].
    • Axios reported on 25 September on Oracle's data-center debt and a force-majeure notice on a New Mexico project [35].
    • Keel's CFO says high-yield project debt spreads have widened [9].
  • Why it matters: A later lease may come with weaker credit, a lower rent or tougher financing.
  • 🛡️ Counter-argument: Hyperscaler capex guidance for 2026 is still rising (~$735–750B combined, secondary). Keel targets investment-grade credit.
  • Severity: Medium

🐻 Bear #7Québec policy is turning hostile

  • The point: Québec's 170 MW of sites face sharply higher power prices for crypto, and later for data centers.
  • Evidence:
    • Québec Decree 88-2026 (18 February 2026) directs a punitive increase in the crypto tariff (Tarif CB). Hydro-Québec proposed demand-charge increases of 166–193% and energy-charge increases of up to 199%. The 10-K says that would make Québec mining "economically unviable" [1].
    • Keel sought judicial review on 17 March. The Régie file R-4333-2026 was still open on 31 August [1,21].
    • Separately, Hydro-Québec seeks a data-center rate averaging ~13¢/kWh, about double the current large-power rate. Hearings ran into October 2026 [20].
  • Why it matters: It could end mining cash flow early and weaken the conversion case for Sherbrooke.
  • 🛡️ Counter-argument: Québec is a minority of value. Management argues data-sovereignty demand can carry premium pricing [9].
  • Severity: Low Medium.

🐻 Bear #8Legacy governance and litigation baggage

  • The point: Bitfarms' history includes a fired CEO and a hostile bidder citing governance.
  • Evidence:
    • The 2024 dismissal of CEO Geoff Morphy and his $27M lawsuit (settled) [40].
    • Riot's hostile ~$950M bid, citing governance [40].
    • An open securities class action naming the current CEO [2].
    • Riot has since sold below 5% [39].
  • Why it matters: Tenants and lenders run counterparty diligence, and the class action is unquantified.
  • 🛡️ Counter-argument: The board and management have largely turned over since 2024, and the company is now redomiciled under Delaware governance [1,4].
  • Severity: Low

Tailwinds NOT in the forecasts

  • Investment-grade-backed lease for ≥300 MW at Panther Creek — 62% chance within 36 months (38% within 12 months) · Impact if it happens: High

    • What it is: a 15-year lease with a hyperscaler, or with an AI cloud backed by a hyperscaler, on the 350 MW PPL capacity.
    • Why it's not priced in: Management gives no guidance. The consensus revenue snapshot shows ~flat growth (Simply Wall St, secondary), and the stock trades ~58% below its 52-week high on lease doubts [29].
    • What would trigger it / the sign to watch: final DEP permits; the PPL energy service and construction agreements executed; the $60.8M letter of credit posted by 31 January 2027 [2].
    • Why this probability: firm 2027 PJM power is scarce and most comparable miners did sign within ~12–18 months of being power-ready, but Keel's own goalposts are moving and AI financing has cooled.

  • Pennsylvania power expansion (part of the ~2 GW of applications) converted into signed utility agreements — 35% chance within 36 months · Impact: Medium–High

    • What it is: new energy service agreements lifting secured capacity well above 648 MW, for example Panther Creek to 400–500 MW or Scrubgrass phase 1.
    • Why it's not priced in: this capacity is expansion only, with no contracts, and analysts mostly model the three priority sites.
    • What would trigger it / the sign to watch: the December/January update; Scrubgrass load-study results; an independent power producer committing to on-site gas [7,9].
    • Why this probability: utilities are now pushing grid-upgrade costs onto large loads, so not all applications will become contracts on economic terms.

  • Takeover or strategic investment — 12% chance within 36 months · Impact: High

    • What it is: a larger operator, or a tenant taking equity, as Google did in Cipher and TeraWulf deals [24,25].
    • Why it's not priced in: no process has been disclosed.
    • What would trigger it / the sign to watch: 13D filings; equity attached to a lease.
    • Why this probability: Riot's 2024 bid shows these assets attract bidders, but convertible notes worth ~$1B raise the cost of buying the company.

  • Québec 96 MW Sherbrooke approval plus a data-sovereignty lease — 25% chance within 36 months · Impact: Low–Medium

    • What it is: provincial approval, then a lease to a Canadian-sovereignty AI tenant.
    • Why it's not priced in: analysts focus on the US sites.
    • What would trigger it / the sign to watch: the ministry approval; the Régie's data-center rate decision [2,20].
    • Why this probability: Québec's rate direction is unfavourable.

Headwinds NOT in the forecasts

  • Construction funded with new equity (ATM or follow-on) at today's low price — 55% chance within 18 months · Impact: High

    • What it is: issuing shares to fund the equity slice of the build or pre-lease spending.
    • Why it's not priced in: buy-side targets of $5.50–$10 assume financing mostly via project debt.
    • What would trigger it / the sign to watch: a new ATM filing (an S-3 or prospectus supplement); a lease signed without an investment-grade wrap; cash falling below ~$500M.
    • Why this probability: Keel has used an ATM before ($375M), it has flagged one, and its build need clearly exceeds its free cash.

  • Lease signing slips into the second half of 2027 or later, and the first Pennsylvania hall misses its end-2027 date — 40% chance · Impact: High

    • What it is: a lost 2027 delivery window; power capacity paid for but idle.
    • Why it's not priced in: company and broker timelines still say end-2027.
    • What would trigger it / the sign to watch: no lease at Q3 results (~9 November, unconfirmed) or at Q4 results; the PPL letter of credit not posted [2].
    • Why this probability: every Keel date has slipped so far, but the utility-side dates have held.

  • PJM or Pennsylvania rules curtail or re-price Keel's large loads — 25% chance within 36 months · Impact: Medium

    • What it is: the curtailment framework for large loads without their own generation, or utility tariffs adding upgrade costs or minimum bills.
    • Why it's not priced in: the rules are still at FERC; grandfathering is unclear.
    • What would trigger it / the sign to watch: FERC orders on the PJM large-load filing; PPL and FirstEnergy large-load tariff filings [17,19].
    • Why this probability: regulators lean toward protecting existing customers, but contracted loads are usually grandfathered.

  • Québec tariff ends mining early, with impairments — 45% chance within 18 months · Impact: Low–Medium

    • What it is: Decree 88-2026 increases upheld.
    • Why it's not priced in: the outcome is still pending.
    • What would trigger it / the sign to watch: Régie decision R-4333-2026; the judicial review outcome [1,21].
    • Why this probability: the government decree directs the increase, so the regulator has limited room.

Investment Risk Score: 63/100

Risk area Weight Score (0–100) Why
Business & competition 25% 70 No tenant or HPC revenue. Peers signed first. Its edge (time-to-power) is real but erodes over time.
Financial health & dilution 20% 58 $819M liquidity and cheap converts due 2031–32, but ~$1.05B of debt against $329M of equity, ~$100M+ a year of burn, and a large unpriced build bill.
Execution (delivering the plan) 20% 72 No data-center build history, and all HPC dates and the lease target have slipped. A Digital Realty hire partly mitigates this.
Dependency & concentration 15% 65 ~88% of 2030 MW sits in Pennsylvania, with one utility per site. Any first lease means one tenant per site.
Regulation, policy & geopolitics 10% 55 Pennsylvania executive order, PJM curtailment proposal, PUC tariffs, Québec Decree 88-2026.
Sentiment & expectations 10% 40 The stock has already fallen ~58% from its high, but broker targets still assume a lease.

Overall = 63/100 (high). Weighted: 17.5 + 11.6 + 14.4 + 9.75 + 5.5 + 4.0 = 62.75. Keel is not a near-term solvency risk: it has cash and long-dated, low-coupon debt. The risk is binary on commercial execution: a pre-revenue developer whose value depends on one or two leases that haven't been signed and builds it has never done. This is the first deep dive, so there is no change to report.

Conclusion

  • What kind of company this is: a power-and-land developer that hasn't yet proved it can turn power into rent. Owning KEEL is a bet that 648 MW of 2027 grid power, mostly in Pennsylvania, gets leased to creditworthy AI tenants on peer-like terms, and financed without crushing dilution.
  • The 2–3 things that will decide the outcome:
    1. Whether, when and with whom the first Panther Creek or Sharon lease is signed, and whether the tenant's credit is investment-grade.
    2. How the build is financed, and at what cost per MW.
    3. Whether Pennsylvania permits and PJM rules let 2027 delivery happen.
  • Signposts — bull confirmed:
    • A signed lease with an investment-grade wrap.
    • The PPL agreement executed and the $60.8M letter of credit posted.
    • Project debt at ≥70% loan-to-cost without new equity.
    • New Pennsylvania energy service agreements in the December/January update.
  • Signposts — bull broken:
    • No lease by Q4 2026 results.
    • A new ATM or discounted equity raise before a lease.
    • Panther Creek ready-for-service slips into 2028.
    • A tenant with sub-investment-grade credit and no backstop.
  • Balanced verdict:
    • Strongest bull: Keel has what AI developers most lack: contracted, near-term grid power in PJM, permits largely in hand and the cash to wait. A peer-style lease on 350 MW would be worth far more than the stock implies.
    • Strongest bear: a power contract is not a business. Keel has no tenant, no building experience and a funding plan that only works once a lease exists, and its own timelines keep moving while better-placed peers took the first wave of demand.
    • Where the evidence leans: for now, slightly bearish on timing and neutral on the asset quality. The power is real and documented; the commercial proof is missing.

Open questions & data gaps

  • Lease economics. Rent per kW-month, build cost per MW and expected yield on cost are not found. This matters most of all; it decides whether a lease creates value per share.
  • No verbatim call transcripts. Q1 and Q2 2026 call content came only from summaries (Motley Fool, Zacks, Investing.com), so quotes are secondary.
  • PPL agreement status. Whether the Panther Creek PPL energy service and construction agreements have been executed is not found. The final DEP permit status after August 2026 is also not found.
  • Grid-rule grandfathering. Whether Keel's contracted loads are grandfathered from the PJM curtailment proposal and the Pennsylvania tariffs is not found.
  • Riot's stake. Riot's current stake and any other holder above 5% are not checked for 2026.
  • The 5 October drop. Blogs claim "project delays and cost overruns"; no company source was found, so it is UNCONFIRMED.
  • Next results date. Q3 2026 results reportedly 9 November, which the company has not confirmed.
Sources
  1. Bitfarms Ltd. Form 10-K FY2025 — SEC EDGAR — 2026-03-31 — https://www.sec.gov/Archives/edgar/data/1812477/000121390026037514/ea0282809-10k_bitfarms.htm
  2. Keel Infrastructure Corp. Form 10-Q Q2 2026 — SEC EDGAR — 2026-08-10 — https://www.sec.gov/Archives/edgar/data/1812477/000181247726000023/keel-20260630.htm
  3. Keel Infrastructure Reports Second Quarter 2026 Results — Keel / SEC 8-K Ex.99.1 — 2026-08-10 — https://www.sec.gov/Archives/edgar/data/1812477/000181247726000021/a202689finalprkeelq22026ea.htm
  4. Bitfarms Officially Rebrands as Keel Infrastructure, Completes U.S. Redomiciliation — Keel / SEC 8-K Ex.99.1 — 2026-04-01 — https://www.sec.gov/Archives/edgar/data/1812477/000121390026038080/ea028450001ex99-1.htm
  5. Keel Infrastructure Reports First Quarter 2026 Results — Keel / SEC 8-K — 2026-05-11 — https://www.sec.gov/Archives/edgar/data/0001812477/000121390026054167/ea028991401ex99-1.htm
  6. Keel (KEEL) Q1 2026 earnings call (summary, not verbatim) — Motley Fool — 2026-05-11 — https://www.fool.com/earnings/call-transcripts/2026/05/11/keel-keel-q1-2026-earnings-call-transcript/
  7. Keel (KEEL) Q2 2026 earnings call (summary, not verbatim) — Motley Fool — 2026-08-10 — https://www.fool.com/earnings/call-transcripts/2026/08/10/keel-keel-q2-2026-earnings-call-transcript/
  8. KEEL Q2 Earnings Call Highlights — Zacks via Yahoo Finance — 2026-08 — https://finance.yahoo.com/markets/stocks/articles/keel-q2-earnings-call-highlights-140000957.html
  9. Earnings call transcript: Keel Q2 2026 (summary) — Investing.com — 2026-08 — https://www.investing.com/news/transcripts/earnings-call-transcript-keel-q2-2026-loss-widens-as-ai-data-center-shift-advances-93CH-4849412
  10. Bitfarms Converts Macquarie Debt Facility to $300M Project Financing — Keel IR — 2025-10-10 — https://ir.keelinfra.com/news-releases/news-release-details/bitfarms-converts-macquarie-debt-facility-300m-project-financing
  11. Keel Infrastructure Announces Pricing of Upsized $400M Convertible Senior Notes — SEC 8-K Ex.99.1 — 2026-06-05 — https://www.sec.gov/Archives/edgar/data/0001812477/000121390026065506/ea029374801ex99-1.htm
  12. Keel closes $458M convertible senior notes — SEC 8-K Ex.99.1 — 2026-06-09 — https://www.sec.gov/Archives/edgar/data/0001812477/000121390026067042/ea029411501ex99-1.htm
  13. Keel Infrastructure supports Governor Shapiro's GRID standards — GlobeNewswire via Finviz — 2026-08-20 — https://finviz.com/news/384331/keel-infrastructure-supports-governor-shapiros-grid-standards-for-responsible-data-center-development
  14. PowerSecure executes agreement with Keel Infrastructure to support Moses Lake Data Center Campus — PowerSecure — 2026-09-09 — https://powersecure.com/blog/powersecure-executes-agreement-with-keel-infrastructure-to-support-moses-lake-data-center-campus
  15. Moses Lake Moves From Bitcoin to AI and HPC — Data Center Frontier — 2026-08-24 — https://www.datacenterfrontier.com/machine-learning/article/55400127/moses-lake-moves-from-bitcoin-to-ai-and-hpc
  16. PJM Capacity Auction Procures 138,318 MW of Generation Resources — PJM / PR Newswire — 2026-07-14 — https://www.prnewswire.com/news-releases/pjm-capacity-auction-procures-138-318-mw-of-generation-resources-as-work-continues-to-address-growing-electricity-demand-302825613.html
  17. PJM board seeks FERC approval on data center curtailments — T&D World — 2026-07 — https://www.tdworld.com/policy/news/55394729/pjm-board-seeks-ferc-approval-on-data-center-curtailments-amid-grid-capacity-concerns
  18. FERC Greenlights PJM Co-Located Load Framework in Part — Willkie Farr — 2026-07 — https://www.willkie.com/publications/2026/07/ferc-greenlights-pjm-co-located-load-framework-in-part-calls-for-further-tariff-updates
  19. PUC Releases Final Order Establishing Large Load Model Tariff Framework — Pennsylvania PUC — 2026-05-13 — https://www.puc.pa.gov/press-release/2026/puc-releases-final-order-establishing-first-of-its-kind-large-load-model-tariff-framework-05132026
  20. Hydro-Québec to face off against industry actors like Google in data centre hearing — BNN Bloomberg / Canadian Press — 2026-10-01 — https://www.bnnbloomberg.ca/business/technology/2026/10/01/hydro-quebec-to-face-off-against-industry-actors-like-google-in-data-centre-hearing/
  21. Régie de l'énergie file R-4333-2026, Keel correspondence — Régie de l'énergie du Québec — 2026-08-31 — https://www.regie-energie.qc.ca/fr/participants/dossiers/R-4333-2026/doc/R-4333-2026-C-Keel-0015-Corresp-Autres-2026_08_31.pdf
  22. Hut 8 Form 8-K Ex.99.1 (River Bend lease) — SEC EDGAR — 2026-02-25 — https://www.sec.gov/Archives/edgar/data/1964789/000110465926019393/hut-20260225xex99d1.htm
  23. Riot Platforms agrees 191MW 20-year lease worth $9.1bn — Data Center Dynamics — 2026-08 — https://www.datacenterdynamics.com/en/news/riot-platforms-agrees-191mw-20-year-lease-agreement-with-anthropic-worth-91bn-report/
  24. AWS signs 300MW hosting deal with Cipher Mining — Data Center Dynamics — 2025 — https://www.datacenterdynamics.com/en/news/aws-signs-300mw-hosting-deal-to-lease-capacity-from-cipher-mining/
  25. TeraWulf Expands Strategic Partnership with Fluidstack (168 MW JV) — TeraWulf / Seeking Alpha PR — 2025-10-28 — https://seekingalpha.com/pr/20281887-terawulf-expands-strategic-partnership-with-fluidstack-through-new-168-mw-ai-compute-joint
  26. Core Scientific Announces Second Quarter 2026 Results — Business Wire — 2026-07-28 — https://www.businesswire.com/news/home/20260727468665/en/core-scientific-announces-second-quarter-2026-results/
  27. Keel Infrastructure stock initiated at Buy by Jones Trading — Investing.com — 2026-09-28 — https://www.investing.com/news/analyst-ratings/keel-infrastructure-stock-initiated-at-buy-by-jones-trading-on-ai-demand-93CH-4919592
  28. Keel Aktie: Citizens startet bei 10 Dollar — Börse Express — 2026-06-27 — https://www.boerse-express.com/news/articles/keel-aktie-citizens-startet-bei-10-dollar-922078
  29. KEEL analyst forecast page — MarketBeat — 2026-10-09 — https://www.marketbeat.com/stocks/NASDAQ/KEEL/forecast/
  30. Consensus: KEEL average target $6.25 — MarketBeat — 2026-10-02 — https://www.marketbeat.com/instant-alerts/consensus-keel-infrastructure-nasdaq-keel-stock-has-average-target-price-of-625-according-to-brokerages-2026-10-02/
  31. With AI Deals in Focus, Keel CEO Buys $200,000 of Stock on Dip — The Energy Mag — 2026-08-15 — https://theenergymag.com/news/market-news/keel-ceo-dip-purchase-ai-focus
  32. Form 4, B. Gagnon (RSU vesting) — SEC EDGAR — 2026-07-14 — https://www.sec.gov/Archives/edgar/data/0001812477/000181247726000016/wk-form4_1784046127.xml
  33. KEEL stock heads for third week of declines; analyst flags concentration risk — Stocktwits — 2026-10-08 — https://stocktwits.com/news-articles/markets/equity/keel-stock-heads-for-third-week-of-declines-analyst-flags-concentration-risk/cZDUIewRBZq
  34. CoreWeave Sinks 7% as Financed AI Buildout Names Sell Off — 24/7 Wall St. — 2026-10-08 — https://247wallst.com/investing/2026/10/08/coreweave-sinks-7-as-financed-ai-buildout-names-sell-off-apart-from-cloud-software-nebius-drops-6-oracle-falls-5/
  35. Oracle rekindles the data center debt risks — Axios — 2026-09-25 — https://www.axios.com/2026/09/25/oracle-debt-data-centers
  36. U.S. transformer market faces severe supply constraints as lead times extend to four years — pv magazine USA — 2026-05-11 — https://pv-magazine-usa.com/2026/05/11/u-s-transformer-market-faces-severe-supply-constraints-as-lead-times-extend-to-four-years/
  37. Bitfarms Rebrands To Keel Infrastructure, But Financial Engineering Still Weighs — Seeking Alpha — 2026-04-03 — https://seekingalpha.com/article/4888283-bitfarms-rebrands-to-keel-infrastructure-but-financial-engineering-still-weighs
  38. Bitfarms stock pumps, dumps bitcoin mining business — Decrypt — 2026-02-06 — https://decrypt.co/357222/bitfarms-stock-pumps-dumps-bitcoin-mining-business
  39. Riot Offloads 11M Bitfarms Shares, Slips Below 5% — The Energy Mag — 2025-08-19 — https://www.theenergymag.com/news/2025-08-19/riot-bitfarms-offload
  40. Riot Platforms publicly discloses rejected $950M proposal to acquire Bitfarms — BetaKit — 2024 — https://betakit.com/riot-platforms-publicly-discloses-rejected-950-million-usd-proposal-to-acquire-bitfarms/
  41. KEEL Stock Price, News & Analysis (close $3.08 on 2026-10-08) — Stock Titan — 2026-10-08 — https://www.stocktitan.net/overview/KEEL/
  42. Situational Awareness LP liquidation — TFTC — 2026-07-30 — https://www.tftc.io/situational-awareness-lp-liquidation-bitcoin-miners-surge
  43. Washington Legislature Punts Comprehensive Data Center Legislation — Davis Wright Tremaine — 2026-03 — https://dwt.com/blogs/energy--environmental-law-blog/2026/03/washington-legislature-delays-data-center-bill
  44. PA DEP Panther Creek Title V permit file (renewal issued 2026-08-06) — Pennsylvania DEP — 2026-08-06 — https://files.dep.state.pa.us/RegionalResources/NERO/NEROPortalFiles/CommunityInformation/Panther_Creek/
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