Back to Portfolio
KEEL

Keel Infrastructure Corp., formerly Bitfarms/BITF

$2.84 USD 1.71B market cap April 15, 2026
Keel Infrastructure Corp. (formerly Bitfarms Ltd) KEEL BUFFETT / MUNGER / KLARMAN SUMMARY
1 SNAPSHOT
Price$2.84
Market CapUSD 1.71B
EVUSD 1.80B
Net DebtUSD 0.09B
Shares0.603B
2 BUSINESS

Keel Infrastructure (formerly Bitfarms) is a North American energy and digital infrastructure company that completed its U.S. redomiciliation to Delaware and rebrand on April 1, 2026. Revenue comes entirely from Bitcoin mining (selling hashrate to mining pools under FPPS model) with ZERO HPC/AI revenue. The company holds a 2.2 GW energy pipeline across Pennsylvania, Washington, and Quebec. Moses Lake (WA) 18MW HPC conversion underway with $128M supply agreement. Panther Creek (PA) 350MW HPC campus in planning stage. No signed customer leases.

Revenue: USD 0.229B Organic Growth: 19% (72% incl. acquisitions)
3 MOAT NONE

Bitcoin mining is a pure commodity business with zero pricing power, no switching costs, no brand value, and no network effects. The low-cost Quebec hydro advantage is eroding as LATAM assets are divested. The 2.2 GW energy pipeline is potential value but 80%+ is unenergized. No HPC customer has validated the infrastructure proposition with a signed contract. Competitors CORZ, WULF, HUT all have multi-billion-dollar HPC contracts; KEEL has zero.

4 MANAGEMENT
CEO: Ben Gagnon (since July 2024, ~21 months)

Aggressive serial diluter: shares nearly tripled from 252M (2022) to 603M (2026). ATM programs raised $314M+ in 2024. $588M convertible notes raised Oct 2025. Repaid $100M Macquarie facility (Feb 2026). Sold BTC from 1,827 to 1,140 BTC. Recently hired HPC leadership (James Bond SVP HPC, Craig Hibbard SVP Infra). No dividends, no meaningful buybacks. Capital allocation grade: D (perpetual dilution without FCF generation).

5 ECONOMICS
-65.2% Op Margin
Deeply Negative (~-34%) ROIC
USD -0.23B (est. FY2025) FCF
3.1x (on $28.9M Adj EBITDA) Debt/EBITDA
6 VALUATION
FCF/ShareUSD -0.38 (est.)
FCF YieldNegative
DCF RangeUSD 0.13 - 4.04 (fully diluted)

Probability-weighted sum-of-parts: 25% bear (mining only, $0.13 FD), 40% base (Moses Lake + partial Panther Creek, $1.62 FD), 25% bull (full HPC execution, $4.04 FD), 10% extreme bull ($8.08+ FD). BTC mining at $8M/EH/s (declining). HPC at $35/kW/month, 50% EBITDA margin, 10-15x EBITDA. Convertible notes and full dilution subtracted. Probability-weighted fair value: $2.06/share FD.

7 MUNGER INVERSION -71.5%
Kill Event Severity P() E[Loss]
Panther Creek construction delays/cost overruns -45% 40% -18.0%
No HPC contracts signed -- peers 18-24 months ahead -50% 35% -17.5%
Further ATM dilution to fund HPC build -20% 65% -13.0%
Cash burn ($73M/month) exhausts liquidity by late 2026 -40% 30% -12.0%
Bitcoin price decline below $50K -55% 20% -11.0%

Tail Risk: If BTC drops below $50K while Moses Lake conversion overruns and no HPC customer signs, Keel burns through $520M liquidity by mid-2027. The $588M convertible notes trigger restructuring. Probability: 12-18%. Impact: -85% to -95%. The company has never generated positive FCF and has no fallback business model. The "government contract" media error (April 14, subsequently corrected) demonstrates the fragile, narrative-driven nature of the investor base.

8 KLARMAN LENS
Downside Case

In the bear case, BTC declines to $40-50K, halving economics destroy mining margins ($48.2K direct cost), Moses Lake conversion runs over budget with no tenant, and Panther Creek stays on paper. Cash burns at $73M/month, exhausting liquidity by late 2026. Macquarie facility ($300M available) buys 4 more months. Then convertible notes ($588M) become restructuring trigger. Equity value: $0.10-0.25/share.

Why Market Wrong

The market may undervalue 2.2 GW of North American energy pipeline. If AI compute demand creates severe power scarcity by 2027-2028, these assets could be worth $500M-2B+. Moses Lake (near Amazon facilities) and Panther Creek (near CoreWeave) are in prime locations. BTC recovery to $75K provides mining margin cushion while HPC ramps. CEO quoted: hosting margins could exceed 60% vs. 20% for mining.

Why Market Right

KEEL has ZERO signed HPC contracts while peers have $30B+ committed. Liquidity declined 36% in 4 months ($814M to $520M). Net loss DEEPENED to $284M in FY2025. Stock rallied 38% in 3 weeks on rebrand hype and a debunked government contract rumor -- with no fundamental improvement. Shares tripled in 3 years. Management is building hardware (Moses Lake $128M) before securing a customer. This is "build it and they will come" infrastructure speculation.

Catalysts

1) Signed HPC customer contract (only catalyst that truly de-risks the thesis) 2) Moses Lake 18MW conversion completion (target Dec 2026, online H1 2027) 3) Panther Creek ground-breaking (no timeline confirmed) 4) BTC price above $100K (provides mining cushion) 5) Institutional re-rating as infrastructure stock (vs. crypto miner)

9 VERDICT REJECT
D Rejected
Strong Buy$1
Buy$1.5
Sell$5

Keel Infrastructure fails every Buffett quality test: deeply negative ROE, never positive FCF, no moat, shares tripled in 3 years, unproven management in target business, ZERO HPC customer contracts. Since our March 27 analysis, the stock has rallied 38% while liquidity declined 36%, net losses deepened to $284M, BTC holdings were sold down, and the company shifted from net cash to net debt. The rebrand is marketing, not execution. At $2.84, KEEL trades 12-38% above probability-weighted fair value ($2.06-2.54). For AGI infrastructure exposure, CORZ and WULF are superior (signed contracts). Monitor for a signed HPC contract as the single catalyst warranting re-evaluation.

🧠 ULTRATHINK Deep Philosophical Analysis

KEEL (formerly BITF) - Ultrathink Analysis (Refresh April 2026)

The Rebrand Does Not Change the Question

Three weeks ago, Bitfarms was a $2.05 Bitcoin miner with no HPC contracts. Today, Keel Infrastructure is a $2.84 "AI data center company" with no HPC contracts. The name changed. The ticker changed. The country of incorporation changed. The stock price went up 38%. But the fundamental question has not changed one iota:

Can a company that has never generated a dollar of free cash flow, that competes in a commodity business with no moat, that has tripled its share count in three years, become valuable infrastructure -- simply because it owns power contracts in an era of AI demand?

The market's answer, expressed in a $1.71B market cap, is: maybe, with about 50-60% confidence. My answer, expressed through Munger-style inversion, is: probably not, and certainly not at this price.

The Liquidity Clock Is Ticking

The most significant development since March is not the rebrand. It is the balance sheet deterioration.

In November 2025, Keel (then Bitfarms) had $814M in liquidity. By March 2026, it had $520M. That is $294M consumed in roughly four months -- a burn rate of approximately $73M per month. At this rate, the company has seven months of runway from March figures, placing the zero line around October 2026.

Yes, the $300M Macquarie facility is available. Yes, BTC could rally. Yes, they could sell more shares. But each of these "solutions" either adds debt, depends on an external variable management cannot control, or further dilutes the equity base.

The convertible notes tell the real story. $588M at 1.375% interest, convertible at $6.86. If the stock never reaches $6.86, these notes become a $588M debt obligation maturing in 2031. If it does reach $6.86, shareholders absorb 14% dilution. Either outcome is punitive for existing equity holders.

This is not a company with the luxury of patience. It is a company in a race: build HPC infrastructure, sign a customer, and generate revenue before the cash runs out. And it is doing this without a single customer commitment.

Build First, Sell Later: The Dangerous Bet

The most telling distinction between Keel and its peers is the sequencing of commitment.

Core Scientific signed a $10.2B, 12-year contract with CoreWeave BEFORE breaking ground on its HPC expansion. The customer came first. The construction followed. The risk was execution, not demand.

Keel is doing the opposite. It committed $128M to equip the Moses Lake site with liquid cooling, GB300-ready racks, and modular infrastructure -- all before any customer has signed a lease. Management believes AI demand is so overwhelming that "if you build it, they will come."

This is a bet on urgency. And it may be correct -- AI training demand shows no signs of slowing, and power-ready sites are genuinely scarce. But the history of infrastructure investment is littered with companies that built capacity ahead of demand and discovered that "scarce" does not mean "sold." Fiber optic companies learned this in 2001. Solar panel manufacturers learned it in 2012. WeWork learned it in 2019.

The question is not whether AI demand exists. It clearly does. The question is whether Keel, specifically, can capture its share of that demand against competitors who are 18-24 months ahead, who have proven HPC operating capabilities, and who have already locked in the highest-quality customers.

The Fake News Test

On April 14, 2026, a media article incorrectly reported that Keel had "secured a landmark multi-billion-dollar government infrastructure contract." The stock spiked 17.66%. The article was later corrected -- the contract belonged to an entirely different company, Keel Holdings LLC.

This episode is revealing, not because of the media error, but because of the market's reaction to it. The stock jumped nearly 18% on a headline. This tells us two things:

First, the market is desperate for the contract catalyst. The absence of signed HPC contracts is the elephant in every room where KEEL is discussed. When a headline appeared to fill this void, the stock instantly repriced. This confirms our thesis that the first real signed contract will be a massive catalyst -- but also that the stock is trading largely on narrative, not fundamentals.

Second, the investor base is retail and momentum-driven. Institutional infrastructure investors do not buy 18% higher on an unconfirmed headline. They wait for SEC filings, contract details, counterparty creditworthiness, and construction timelines. The fact that KEEL trades on rumors rather than fundamentals tells us who owns it and how it will behave.

What $2.84 Buys You

At $2.84 per share, you are buying:

  • A Bitcoin mining operation worth roughly $156M (19.5 EH/s at $8M/EH/s)
  • 1,140 BTC worth $86M (declining from 1,827 six months ago)
  • $250M in deployable cash (declining from $400M+ six months ago)
  • A $128M equipment purchase for an 18MW facility with no tenant
  • A 350MW Panther Creek campus on paper with no ground broken and no customer
  • A 1.7GW "expansion pipeline" that is mostly applications and LOIs for power
  • MINUS $588M in convertible notes
  • MINUS $23M in other debt
  • MINUS 140M shares of dilution overhang

The net tangible value is approximately negative to zero, depending on assumptions. The $1.71B market cap is entirely option premium -- the market's assessment of probability-weighted future HPC value.

On our probability-weighted scenario analysis, fair value is $2.06-$2.54 per share. At $2.84, you are paying a 12-38% premium to fair value for a company that has gotten more expensive (38% price increase) while getting fundamentally weaker (liquidity down 36%, losses deepening, BTC holdings declining).

Buffett Would Say No. Munger Would Laugh. Klarman Would Wait.

Warren Buffett: "I don't invest in things I don't understand, and I certainly don't invest in things that have never made money."

Charlie Munger: "If a company has to constantly issue shares to fund its operations, the equity is not an investment -- it's a depreciating license to participate in future dilution."

Seth Klarman: "The concept of margin of safety requires that the current price be substantially below intrinsic value. When a company has no earnings, no contracts, and declining liquidity, intrinsic value is highly uncertain -- and paying a premium to the best estimate of that uncertain value is the opposite of margin of safety."

Aschenbrenner: "But the power scarcity is real, and the demand curve is exponential." Fair enough. But his position is 0.4% of a $5.5B fund. He can afford to be wrong. Most retail investors cannot.

What Would Change the Verdict

One event: a signed, binding HPC customer contract for 50+ MW at economics comparable to CORZ/WULF deals ($30-40/kW/month, 10+ year term, investment-grade counterparty).

This would:

  1. Validate the infrastructure value proposition
  2. Provide a revenue floor and cash flow timeline
  3. De-risk the Moses Lake proof-of-concept
  4. Attract institutional capital
  5. Create a basis for fundamental (vs. speculative) valuation
  6. Justify the Panther Creek build-out

Until that happens, Keel is a rebrand without a business model. The name is new. The infrastructure is being built. But the customer -- the entity that transforms a power contract into a revenue stream -- is absent.

The soul of this investment remains unchanged from March: it is a bet on urgency. The belief that AI demand is growing so fast that by the time Keel's facilities are ready, customers will be desperate enough to fill them.

At $2.05, that bet was approximately fairly priced.

At $2.84, you are paying a premium for the same unresolved uncertainty. That is not investing. That is narrative-chasing.

Executive Summary

3-Sentence Investment Thesis

Keel Infrastructure (formerly Bitfarms) completed its U.S. redomiciliation and rebrand on April 1, 2026, signaling full commitment to its HPC/AI pivot away from Bitcoin mining -- but FY2025 results reveal a $284M net loss, declining liquidity ($520M vs. $814M six months ago), and still ZERO signed HPC customer contracts despite peers having $30B+ in committed revenue. The $128M binding supply agreement for the Moses Lake 18MW conversion is a concrete step, but it is an equipment purchase, not a customer commitment -- Keel is building infrastructure without a buyer locked in. At $2.84/share ($1.71B market cap), the stock has appreciated 38% since our March 27 analysis, pricing in more optimism than fundamentals justify given the execution gap vs. peers.

Key Metrics Dashboard (Updated April 15, 2026)

Metric Value Change vs. Mar 27 Assessment
Market Cap $1.71B +39% from $1.23B Mid-cap speculative
Stock Price $2.84 +$0.79 from $2.05 Rebrand momentum
Shares Outstanding 602.85M +5M Continued creep
Revenue (FY2025) $229.3M +72% YoY BTC price driven
Net Loss (FY2025) -$284.5M Deepened from -$54M FY2024 Deteriorating
Adjusted EBITDA (FY2025) $28.9M 13% margin Down from 22% (9M)
Cash $359M Down from $637M Burning fast
BTC Holdings 1,140 BTC ($86M) Down from 1,827 BTC Sold/used
Total Liquidity ~$520M Down from $814M -36% in 6 months
Convertible Notes $588M Unchanged 1.375% due 2031
Macquarie Facility $0 (repaid) Down from $100M drawn Positive
Operating Hashrate ~19.5 EH/s Up from 14.8 EH/s Fleet upgrades
Total Energy Pipeline 2.2 GW Up from 2.1 GW Marginal
HPC Contracts Signed ZERO Unchanged Critical gap
Bitcoin Price ~$75,000 +9% from $69K Supportive
Beta ~4.0 Unchanged Extremely volatile

Verdict: REJECT (maintained, strengthened)

The fundamentals have deteriorated since March: liquidity is draining ($520M vs. $814M), BTC holdings were sold down, net losses deepened to $284M, and the single most important catalyst -- a signed HPC customer contract -- has not materialized. The stock has rallied 38% on rebrand hype, making it MORE expensive on a risk-adjusted basis. This is a speculative option on AI infrastructure, not a value investment.

Phase 0: Quick Screen (Buffett Quality Checks)

Criterion Requirement KEEL Result Pass?
Simple business Understandable BTC mining + HPC pivot (two businesses, neither proven) Partial
Profitable 10+ years Consistent Net loss every year except 2021 ($22M) and 2019 ($3M) FAIL
Consistent FCF Positive FCF Negative FCF every single year FAIL
ROE > 15% High returns Deeply negative (FY2025: ~-47%) FAIL
D/E < 0.5 Conservative ~0.97 (incl. convertibles vs. equity) FAIL
Management skin in game Insider ownership ~4.6% insiders Marginal
Identifiable moat Durable advantage None proven FAIL
Dividend history Shareholder returns Never paid dividends FAIL

Result: FAILS 7 of 8 screens. Unchanged from March analysis.

Phase 1: Risk Analysis (Inversion -- "How Does This Fail?")

Top 10 Risks (Updated April 2026)

# Risk Event Severity Likelihood Expected Loss
1 No HPC contracts signed -- peers 18-24 months ahead -50% 35% -17.5%
2 BTC price decline below $50K (mining uneconomic at $48K+ cost) -55% 20% -11.0%
3 Panther Creek construction delays/cost overruns -45% 40% -18.0%
4 Continued cash burn ($294M lost in 6 months) exhausts liquidity -40% 30% -12.0%
5 Convertible note dilution at $6.86 (85.7M new shares) -15% 50% -7.5%
6 Further ATM dilution to fund HPC build (history of serial dilution) -20% 65% -13.0%
7 Moses Lake 18MW conversion fails to attract customer at target economics -30% 25% -7.5%
8 HPC/AI demand slowdown or overcapacity by 2027-2028 -35% 15% -5.3%
9 New management team (avg 1.5yr tenure) lacks HPC execution experience -25% 30% -7.5%
10 BTC sold down (1,140 from 1,827) reduces financial cushion and upside leverage -10% 80% -8.0%

Total Expected Downside: -107.3% (non-additive; tail scenario = equity near zero)

Critical Risk Update: Liquidity Burn Rate

The most alarming development since March is the liquidity trajectory:

Date Cash BTC Value Total Liquidity Change
Nov 2025 (Q3) $637M $177M $814M -
Mar 2026 (FY2025) $359M $161M $520M -$294M in ~4 months
Implied Burn Rate - - - ~$73M/month

At $73M/month burn, $520M liquidity = ~7 months of runway (to ~November 2026). This is BEFORE Panther Creek construction capex ramps. The $300M Macquarie facility is available but undrawn -- it serves as the backstop. Management will likely need to either sell more BTC, draw on Macquarie, issue more equity, or sign a customer contract with advance payments.

Why "Government Contract" Was Fake News

A media article on April 14 incorrectly reported that Keel had "secured a landmark multi-billion-dollar government infrastructure contract," causing KEEL stock to spike 17.66%. The article was subsequently corrected: there is no government contract. The confusion arose from an unrelated company, Keel Holdings LLC, involved in the Navy's ShipOS initiative. This episode underscores the speculative, narrative-driven nature of KEEL's investor base.

Peer Comparison: HPC Contract Status

Company Ticker HPC Contracts Contract Value Status
Core Scientific CORZ CoreWeave 12yr $10.2B Building
Terawulf WULF Multiple $12.8B Building
Hut 8 HUT Fluidstack 15yr $7.0B Building
Cipher Mining CIFR Multiple $2.0B+ Building
Keel Infrastructure KEEL NONE $0 Prospecting

Keel remains the ONLY major BTC-miner-to-HPC pivot without a signed customer contract. This is the defining risk.

Tail Risk Scenario (Updated)

If BTC drops below $50K while Moses Lake conversion runs over budget and no HPC customer signs, Keel burns through remaining liquidity by mid-2027. The $588M convertible notes become a restructuring trigger. Probability: 12-18%. Impact: -85% to -95%.

Phase 2: Financial Analysis

FY2025 Full Year Results

Metric FY2025 FY2024 Change
Revenue $229.3M $192.9M +19%
Gross Loss -$18.9M -$32.4M Improved
Operating Loss -$149.6M -$107.6M Worsened
Net Loss (continuing ops) -$208.5M -$54.0M -286%
Net Loss (total, incl. discontinued) -$284.5M -$54.0M -427%
Adjusted EBITDA $28.9M ~$55M -47%
Depreciation $98.1M ~$141M Reduced
Impairment Charges $28.4M - New

Revenue grew 19% YoY on a full-year basis (the 72% headline includes Stronghold acquisition effects). But losses DEEPENED dramatically, driven by impairments ($28.4M), infrastructure retrofit costs, and BTC price decline from the $126K October 2025 peak. Adjusted EBITDA nearly halved.

Revenue and Profitability History (Updated)

Year Revenue Gross Margin Op Margin Net Margin
2025 $229.3M -8.2% -65.2% -124.1%
2024 $192.9M -16.8% -55.8% -28.0%
2023 $146.4M -14.7% -49.3% -71.1%
2022 $142.4M 7.4% -199.4% -123.3%
2021 $169.5M 65.6% 37.5% 13.1%
2020 $34.7M 8.3% -19.4% -46.9%

Mining Economics (Post-Halving)

Metric Current Breakeven
Direct Cost per BTC ~$48,200 -
All-in Cost per BTC (est.) ~$65,000-70,000 -
BTC Price ~$75,000 $65-70K
Gross Mining Margin ~35-40% 0% at ~$65K
Network Hashrate ~970 EH/s Rising
KEEL Hashrate ~19.5 EH/s 2% of network
Daily BTC Mined ~7-9 BTC -

Mining margins are razor-thin on an all-in basis. The halving economics remain brutal.

Balance Sheet Analysis (March 2026)

Item Amount Change from Nov 2025
Cash $359M -$278M
BTC Holdings (~1,140 BTC) ~$86M -$91M
Total Liquidity ~$520M -$294M
Convertible Notes (1.375%, due 2031) $588M Unchanged
Macquarie Facility $0 (repaid Feb 2026) -$100M (positive)
Other Debt ~$23M ~Unchanged
Net Debt ~$91M From net cash to net debt
Shareholders' Equity ~$608M Eroding from losses

The company shifted from net cash ($103M) in Nov 2025 to approximately net debt ($91M) in just 4 months. The $300M Macquarie facility is available but undrawn as backstop.

Share Dilution Trajectory (Updated)

Date Shares Outstanding vs. Dec 2022
Dec 2022 ~252M Baseline
Dec 2023 ~339M +34%
Dec 2024 ~554M +120%
Apr 2026 ~603M +139%
Convertible (if converted) +85.7M +173%
Options/Warrants/RSUs ~54M +195%
Fully Diluted ~743M +195%

Shares have nearly tripled since 2022. At $2.84, fully diluted market cap = $2.11B.

Free Cash Flow -- Still Never Positive

Year Operating CF CapEx FCF
FY2025 ~-$100M (est.) ~$130M (est.) ~-$230M
FY2024 -$140.6M $339.9M -$480.5M
FY2023 $24.0M $69.5M -$45.5M
FY2022 $42.7M $194.0M -$151.3M
FY2021 -$42.3M $190.4M -$232.7M

The company has NEVER generated positive free cash flow in its entire public history.

Valuation (Updated April 15, 2026)

Sum-of-Parts:

Component Methodology Value
BTC Mining (19.5 EH/s) At ~$8M/EH/s (peer comp, declining) $156M
Bitcoin Holdings (1,140 BTC) At $75K $86M
Cash (net of ops) Conservative: $250M deployable $250M
Moses Lake (18 MW, under conversion) At cost: $128M supply agreement $128M
Panther Creek (350 MW pipeline) Probability-weighted $0 - $1.1B
Other Pipeline (~1.7 GW) Too early stage $0
Less: Convertible Notes Face value -$588M
Less: Other Debt ~$23M -$23M

Scenario Analysis:

Scenario Probability Equity Value Per Share (603M) Per Share (743M FD)
Bear (mining only, no HPC) 25% $100M $0.17 $0.13
Base (Moses Lake + partial Panther) 40% $1.2B $1.99 $1.62
Bull (full HPC execution, 350MW+) 25% $3.0B $4.98 $4.04
Extreme Bull (1GW+ HPC) 10% $6.0B+ $9.95+ $8.08+
Probability-Weighted $1.53B $2.54 $2.06

At $2.84, the stock trades 12% above probability-weighted fair value (basic), and 38% above on a fully diluted basis. The March 27 analysis found the stock approximately fairly valued at $2.05. The 38% rally has pushed it into overvalued territory absent new contract catalysts.

Phase 3: Moat Analysis

Moat Assessment: NONE (Unchanged)

Moat Source Evidence Rating
Brand "Keel Infrastructure" is brand new; zero recognition None
Switching Costs Zero -- BTC mining has no switching costs; HPC TBD None
Network Effects None None
Cost Advantages Had Quebec hydro (low-cost); selling LATAM assets (losing cheap power) Eroding
Scale Mid-tier; CORZ, MARA, RIOT, CLSK all larger None
Regulatory/IP No patents, no regulatory moat None
Energy Infrastructure 2.2 GW pipeline is the thesis -- but 80%+ is unenergized/under-application Potential only

Moat Verdict: NONE

The low-cost hydro power advantage that was part of the original thesis has been partially abandoned through LATAM asset sales (Paraguay 200MW divested). Quebec operations remain but are being evaluated for HPC conversion. The 2.2GW pipeline is potential value, not proven value. Competitors CORZ and WULF remain 18-24 months ahead in HPC execution.

Pricing Power: Zero

BTC mining = zero pricing power (commodity). HPC hosting = depends on supply/demand and contracts not yet signed. Currently, Keel has pricing power over nothing.

Phase 4: Decision Synthesis

What Changed Since March 27 Analysis

Factor March 27 April 15 Better/Worse?
Stock Price $2.05 $2.84 Worse (more expensive)
Market Cap $1.23B $1.71B Worse (paying more)
Cash $637M $359M Worse (burning fast)
BTC Holdings 1,827 BTC ~1,140 BTC Worse (sold down)
Total Liquidity $814M $520M Worse (-36%)
Net Position -$103M (net cash) +$91M (net debt) Worse
HPC Contracts Zero Zero Unchanged (bad)
BTC Price $69K $75K Better
Hashrate 14.8 EH/s 19.5 EH/s Better
Ticker/Rebrand Pending Completed (KEEL) Neutral
Moses Lake Planning $128M supply agreement Slightly better
Macquarie Debt $100M drawn Repaid in full Better
FY2025 Net Loss N/A -$284.5M Bad

Management Assessment (Updated)

Factor Assessment
CEO Ben Gagnon (~21 months). Completed rebrand and redomiciliation. Hired HPC leadership (James Bond SVP HPC, Craig Hibbard SVP Infrastructure).
CFO Jonathan Mir (~8 months). Executed convertible raise. Repaid Macquarie facility.
HPC Team James Bond (SVP HPC) and Craig Hibbard (SVP Infrastructure) recently hired. Too early to judge.
Insider Ownership ~4.6% -- insufficient for a speculative execution bet
Capital Allocation Mixed: smart convertible pricing but serial dilution history. Sold BTC (reducing upside leverage).
Track Record Zero HPC revenue. Zero signed contracts. Rebrand is marketing, not execution.

The Aschenbrenner Position (Context)

Situational Awareness LP's ~$16M position (0.4% of portfolio). At $2.84, position worth ~$22M if unchanged. This remains rounding error in a $5.5B+ fund. His larger bets (CORZ, WULF, BE) are in companies with signed contracts or proven technology.

Moses Lake: A Step, Not a Leap

The $128M binding supply agreement for the Moses Lake 18MW HPC conversion is the most concrete HPC development to date:

  • $128M with an unnamed "large publicly traded American multinational" for IT equipment and building materials
  • 18MW gross capacity, up to 190kW per rack, liquid cooling, GB300 validated
  • Target completion: December 2026, likely online H1 2027
  • CEO: site "could potentially produce more net operating income than ever generated with Bitcoin mining"

However: this is a supply/equipment purchase, NOT a customer contract. Keel is buying the hardware before securing a tenant. This is the "build first, find customer later" approach that distinguishes KEEL from contracted peers.

Bull vs. Bear (Updated)

Bull:

  1. Energy scarcity premium: 2.2 GW pipeline may be worth $500M-$2B as AI demand accelerates
  2. Moses Lake as proof-of-concept (if customer signs by end-2026)
  3. BTC recovery to $75K provides mining margin cushion
  4. Rebranding may attract infrastructure-focused institutional capital

Bear (stronger):

  1. Zero HPC contracts while peers have $30B+ committed
  2. Liquidity draining at $73M/month -- only ~7 months runway before needing capital
  3. Stock 38% more expensive on zero fundamental improvement
  4. $284M net loss in FY2025 -- losses DEEPENING
  5. Serial diluter: shares tripled in 3 years with more likely ahead
  6. "Government contract" fake news shows narrative-driven investor base
  7. BTC sold down from 1,827 to 1,140 -- reducing financial cushion

Entry Prices (Updated)

Level Price Basis
Strong Buy $1.00 Bear case floor + margin of safety
Accumulate $1.50 Probability-weighted FD value with 25% MoS
Fair Value $2.06-2.54 Probability-weighted range
Current $2.84 12-38% premium to fair value
Sell/Trim $5.00+ Approaches convertible conversion ($6.86)

Monitoring Triggers

Metric Threshold Action
Signed HPC customer contract Any binding 50+ MW commitment Re-evaluate from scratch
Moses Lake customer signed Binding lease for 18MW Partial de-risking
BTC price Below $50K sustained Mining collapse; cash burn accelerates
Cash + BTC liquidity Below $300M Dilution imminent
Share dilution >650M basic shares Value destruction
Panther Creek ground-breaking Physical construction Execution proof

Final Decision
Recommendation REJECT (maintained, strengthened)
Quality Grade D
Tier Rejected
Reason No moat, no profitability, no FCF, serial dilution, no HPC contracts, deteriorating liquidity, 38% more expensive on no fundamental improvement
Fair Value Range $0.13 (bear, FD) to $4.04 (bull, FD); probability-weighted $2.06 FD
Strong Buy $1.00
Accumulate $1.50
Current $2.84
Sell Price $5.00+

One-Line Verdict

Keel Infrastructure is a rebranded Bitcoin miner burning $73M/month with zero HPC customer contracts, trading at a 38% premium to probability-weighted fair value on the hope that AI will need its electricity -- a speculation that has gotten more expensive without getting less risky.

Sources: Bitfarms/Keel FY2025 results (March 31, 2026), investor.bitfarms.com press releases (April 1 2026 rebrand, November 13 2025 Washington site, monthly production updates), GlobeNewswire, StocksToTrade.com (government contract correction), TradingKey.com, CoinDesk. No analyst reports used as investment inputs.