CORZ -- Core Scientific, Inc.

Bitcoin miner pivoting to AI/HPC data center hosting. Leopold's activist stake (13D, 9.4%). 1.4 GW of power infrastructure at the center of the AI supply chain. | Analysis date: 2026-03-12

Why are we looking at this?

Leopold Aschenbrenner filed a Schedule 13D on Core Scientific -- an activist filing, not a passive one. He holds 28.8M shares (9.4% of the company, ~$419M at cost). He explicitly reserves the right to recommend mergers, changes in control, and to communicate with other shareholders. He killed the CoreWeave acquisition attempt in October 2025 by rallying shareholders to vote it down. This is the only company where Leopold has taken an activist position. He doesn't just think CORZ is undervalued -- he's actively trying to change how the company is managed. The question: is the pivot from Bitcoin mining to AI/HPC hosting real, and what's the bull case market cap if it works? At what price would we want to buy?

$16.24
Stock Price
$5.12B
Market Cap
19x
EV / Revenue
$319M
Revenue (FY2025)
-38%
Revenue Growth YoY
1.4 GW
Gross Power Capacity
-$963M
Stockholders' Equity
-$289M
Net Loss (FY2025)
6.9
Beta (Extreme)
Stock Price — CORZ

1. The Pivot: Bitcoin Miner to AI Infrastructure

Core Scientific is not really one company. It is two businesses sharing the same physical assets, and one is dying while the other is being born.

The Dying Business: Bitcoin Mining

Self-Mining Revenue (FY2025)$229M (72%)
Self-Mining Revenue (FY2024)$409M
Change-44%
Gross Margin5%
Hosted Mining Revenue$24M (8%)
Hosted Mining Margin32%

Mining is cratering. Revenue down 44% YoY. Self-mining gross margin collapsed from 23% to 5%. The April 2024 halving cut block rewards in half. Unless Bitcoin doubles (or more), mining margins stay near zero. The company is explicitly winding this down: "We intend to convert every megawatt in our portfolio to high-density colocation infrastructure over the next three years."

The Growing Business: AI/HPC Hosting

Colocation Revenue (FY2025)$65M (20%)
Colocation Revenue (FY2024)$24M
Change+168%
Gross Margin30%
Contracted Capacity590 MW
Deferred Revenue$556M

HDC is growing fast but still small. Revenue nearly tripled, margins improved from 11% to 30%, and there's $556M of deferred revenue (contracted but not yet recognized). The company expects "meaningful" HDC revenue in 2026 as billable capacity gets delivered. At full deployment, this becomes the entire company.

The Crossover Question: When Does HDC Overtake Mining?

ScenarioMining RevHDC RevTotal RevHDC %When HDC Dominates
FY2024 (actual)$486M$24M$511M5%Mining era
FY2025 (actual)$254M$65M$319M20%Transition begins
FY2026 (projected)$150-200M$200-400M$350-600M50-67%Crossover year
FY2027 (projected)$50-100M$500-800M$550-900M75-90%HDC dominant
FY2028+ (projected)~$0$800M-1.2B$800M-1.2B~100%Pure HDC

2026 is the crossover year. Mining revenue will continue declining (halving, rising difficulty, converting MW to HDC). HDC revenue will ramp as the 590 MW of contracted CoreWeave capacity comes online. By 2027-2028, this should be essentially a pure AI/HPC hosting company. The critical question is whether the stock should be valued as a dying miner (cheap multiples) or a growing infrastructure company (expensive multiples). The market is currently trying to figure this out.

2. The Power Asset: 1.4 GW Across 10 Data Centers

The real asset here is not the mining equipment. It is the power infrastructure.

MetricValueContext
Gross Utility Power~1.4 GWAcross 10 data centers in 7 US states
Leasable Customer Power920 MWAvailable for HDC customers
Contracted (CoreWeave)590 MW64% of leasable capacity sold
Remaining Available330 MW36% still to be sold
PP&E (Net)$1.29B3x from $433M in FY2024 (massive build)

What Is This Power Worth?

Valuation Method$/MWImplied EV (920 MW)Basis
Bitcoin miner valuation$3-5M$2.8-4.6BHow crypto miners trade
AI/HPC infrastructure valuation$10-15M$9.2-13.8BHow data center operators trade
Premium AI infrastructure$15-20M$13.8-18.4BTier-1 hyperscale colo
Current Enterprise Value$6.6M$6.07BMarket price today

The re-rating thesis: At $6.07B EV, the market is valuing CORZ at ~$6.6M per MW -- roughly midway between Bitcoin miner multiples ($3-5M) and AI/HPC multiples ($10-15M). If the pivot succeeds and the market re-rates this as pure AI infrastructure, the same assets could be worth $9-18B. That is a 1.5-3x from current EV without adding a single megawatt.

Competitors for context: CORZ has the largest power capacity among all Bitcoin miners (~1.4 GW gross). For comparison: IREN ~1GW, RIOT ~1GW, CLSK ~0.9GW, CIFR ~0.7GW. CORZ is also the furthest along in the AI/HPC pivot -- already generating real HDC revenue with an actual hyperscaler contract (CoreWeave).

Data Center Locations

StateOwnershipNotes
GeorgiaOwned
KentuckyOwned
North CarolinaOwned
OklahomaOwned
AlabamaLeased
North DakotaLeased
Texas (Austin)LeasedFirst CoreWeave site (16 MW)

3. The CoreWeave Problem: 100% Customer Concentration

This is the single biggest risk in the entire investment thesis.

CoreWeave's share of Colocation revenue100%
Number of other HDC customersZERO
Total contracted capacity590 MW

Every dollar of colocation revenue -- the segment that is supposed to become the entire company -- comes from ONE customer. The 10-K explicitly states: "One customer, CoreWeave, currently accounts for 100% of our Colocation segment revenue."

CoreWeave Contract Timeline

DateEventCapacity
Feb 2024Initial contract at Austin, TX16 MW
Jun 2024Major expansion contract+200 MW
2024-early 2025Multiple option exercises+374 MW
CurrentTotal contracted~590 MW

Why This Is Dangerous

CoreWeave's Own Risks

  • IPO'd in 2025, heavily leveraged (~$7.5B debt at IPO)
  • Revenue concentrated in NVIDIA GPU rentals
  • If AI demand slows or GPU prices fall, CoreWeave's business model breaks
  • They tried to buy Core Scientific (July 2025) -- shareholders said no
  • CoreWeave going bankrupt = 100% of CORZ's HDC revenue goes to zero

Mitigating Factors

  • Contracts are long-term with option exercises built in
  • $556M deferred revenue = prepaid commitment from CoreWeave
  • Even if CoreWeave fails, the physical infrastructure remains
  • Other hyperscalers (Microsoft, Google, Amazon) need the same capacity
  • The power contracts and data centers can be re-leased to other customers

The Merger That Didn't Happen

July 7, 2025CoreWeave proposed all-stock merger to acquire Core Scientific
October 14, 2025Leopold's 13D/A filed showing 9.4% stake (two weeks before vote)
October 30, 2025Stockholders REJECTED the merger at special meeting
Cost to CORZ$21.6M in advisory/legal fees. No termination fees.

Why shareholders rejected it: The merger was all-stock, meaning CORZ shareholders would have received CoreWeave shares. Given CoreWeave's heavy leverage and CORZ's valuable power assets, the consensus (led by Leopold) was that CORZ's assets were worth more independent than what CoreWeave was offering. Leopold's view: keep the assets independent, diversify customers, and let the market re-rate.

4. Leopold Aschenbrenner's Activist Campaign

Position Summary

EntitySituational Awareness LP
Key PersonsLeopold Aschenbrenner (German citizen), Carl Shulman (US citizen)
Filing TypeSchedule 13D (ACTIVIST)
Current Shares28,756,478
Ownership %9.4%
Estimated Cost Basis~$14.00/share avg
Estimated Total Cost~$403M
Current Value~$467M (at $16.24)

13F Position Build (Quarterly)

QuarterSharesChangeApprox. Value
Q1 20254,521,578New position$32.7M
Q2 20257,994,038+77%$136.5M
Q3 202520,180,534+152%$362.0M
Q4 202528,756,478+42%$418.7M

Key Purchases (From 13D Filing)

DateSharesAvg PriceTotal
Jul 18203,578$13.19$2.7M
Jul 21315,005$13.48$4.2M
Jul 22199,461$13.24$2.6M
Jul 31676,250$13.57$9.2M
Aug 51,000,000$13.75$13.8M
Aug 61,000,000$13.99$14.0M
Aug 71,000,000$14.32$14.3M
Aug 8682,666$14.32$9.8M
Aug 111,493,133$14.63$21.8M
Aug 123,064,787$14.90$45.7M
Aug 13615,413$13.85$8.5M

Average purchase price: ~$14.15/share. Leopold bought aggressively through July-August 2025, deploying roughly $147M in disclosed open-market purchases across just 4 weeks. His buying was methodical -- ramping from 200K shares/day to 3M shares/day as he built conviction.

What Does Leopold Want?

The 13D filing explicitly reserves Leopold's right to:

Our read: Leopold wants CORZ to (1) keep its assets independent (hence killing the CoreWeave merger), (2) diversify HDC customers beyond CoreWeave, (3) sign contracts with 2-3 additional hyperscalers, and (4) let the market re-rate from Bitcoin miner multiples to AI infrastructure multiples. His "Situational Awareness" thesis (AGI is coming and needs enormous compute/power infrastructure) makes CORZ a direct play on the physical layer of AGI.

5. The Bankruptcy Hangover

Core Scientific emerged from Chapter 11 bankruptcy on January 23, 2024 -- just over two years ago. The balance sheet still carries the scars.

Balance Sheet (Dec 31, 2025)

ItemFY2025FY2024Notes
ASSETS
Cash & Equivalents$311M$836MDeclining rapidly (-63%)
Digital Assets (Bitcoin)$222M$24MAccumulating BTC (HODL strategy)
Customer Funding Receivable$362M$43MCoreWeave prepayments for buildout
Total Current Assets$896M$904MFlat despite cash decline
PP&E (Net)$1,293M$433M3x increase -- massive build
Total Assets$2,348M$1,476M+59%
LIABILITIES
Current Liabilities$781M$135MAccrued construction costs
Convertible Notes (LT)$1,060M$1,074M$460M @ 3% + $625M @ 0%
Warrant Liabilities$936M$1,097MFrom bankruptcy emergence
Deferred Revenue (LT)$428M$0CoreWeave prepayments
Total Liabilities$3,310M$2,419M
EQUITY
Stockholders' Deficit-$963M-$943MNegative equity

Debt Structure

InstrumentPrincipalCouponMaturityConversion
3.00% Convertible Senior Notes$460M3.0%2029TBD conversion rate
0.00% Convertible Senior Notes$625M0.0%2031~$22.49/share (44.46 shares/$1K)
Total Notes$1,085M
Warrant Liabilities$936MTwo tranches from Chapter 11 emergence

The $625M Zero-Coupon Notes Are Critical

These convert at $22.49/share. The stock is at $16.24 -- 28% below the conversion price. If the stock rises above $22.49, the notes convert to ~27.8M new shares (about 9% dilution). If the stock stays below $22.49, the company must repay $625M in cash by 2031. With only $311M in cash today, that would require either generating significant cash flow or raising capital.

The 3% notes due 2029 are a nearer-term obligation: $460M plus interest. Combined with the zero-coupon notes, CORZ has $1.085B in debt coming due within 3-6 years.

Warrant liabilities ($936M): These are non-cash liabilities carried at fair value -- they represent warrants issued to creditors when CORZ emerged from bankruptcy. The warrants are exercisable at $6.81 and $0.01 per share. They create dilution but do not require cash repayment. The fair value fluctuates with the stock price, which is why net loss is distorted by warrant revaluation (FY2024 had $1.37B of non-cash warrant revaluation losses).

Cash Flow Reality

Cash FlowFY2025FY2024
Operating Cash Flow$278M
Capital Expenditures-$729M
Free Cash Flow-$451M
Financing Activities-$63M
Net Cash Change-$526M

CORZ is burning half a billion dollars per year building out HDC infrastructure. Operating cash flow is healthy ($278M) but capex ($729M) dwarfs it. The company started the year with $836M in cash and ended with $311M. At this burn rate, they run out of cash by mid-2026 unless capex slows, HDC revenue ramps, or they raise capital. The $362M customer funding receivable (CoreWeave prepayments) and $556M deferred revenue suggest CoreWeave is partially funding the buildout.

6. Financial Deep Dive

Income Statement by Segment

MetricFY2023FY2024FY2025YoY Change
Revenue
Colocation (HDC)$0$24M$65M+168%
Self-Mining$310M$409M$229M-44%
Hosted Mining$69M$78M$24M-69%
Total Revenue$379M$511M$319M-38%
Costs & Margins
Colocation Gross Margin11%30%+19pp
Self-Mining Gross Margin43%23%5%-18pp
Hosted Mining Gross Margin38%31%32%+1pp
Total Gross Profit$155M$121M$38M-69%
Operating Loss-$52M-$142M-$246M-73%
Net Loss-$825M-$1,438M-$289MImproved
Adjusted EBITDA$77M$157M$30M-81%

Reading Through the Noise

7. Decision Tree: Is This a 10x?

Question 1: Can It Go to Zero?

Answer: Unlikely but not impossible. This is a financially fragile company.

Cash$311MDeclining rapidly
Bitcoin Holdings$222MLiquidatable but volatile
Total Debt$1,085M$460M due 2029, $625M due 2031
Warrant Liabilities$936MNon-cash, but dilutive
Stockholders' Equity-$963MDeeply negative
PP&E$1,293MReal, hard assets
Deferred Revenue$556MContracted future income
Free Cash Flow-$451MHeavy capex phase

Zero scenario: If CoreWeave goes bankrupt AND Bitcoin crashes AND CORZ can't find new HDC customers AND capital markets freeze up, the company could run out of cash and default on its $1.085B debt. The physical assets (data centers, power contracts) have value even in liquidation, but equity holders would be wiped out if liabilities exceed asset recovery.

Counter: The power infrastructure is real and valuable. Even in distress, these assets would attract buyers -- other hyperscalers, infrastructure funds, utilities. The data center power shortage means there's a ready market for these assets. Bankruptcy is possible but equity wipeout is unlikely unless multiple things go wrong simultaneously.

Result: This is NOT a "can't lose money" situation. Proceed with caution.

Question 2: What Must Go Right for 10x?

The pivot must succeed AND customer concentration must decrease AND the market must re-rate.

Specifically, three things need to happen:

  1. CoreWeave delivers: The 590 MW contract must be fully built out and generating revenue.
  2. Customer diversification: CORZ must sign 2-3 additional HDC customers to reduce the 100% concentration risk.
  3. Market re-rates: Investors must stop pricing CORZ as a Bitcoin miner and start pricing it as AI infrastructure.

Evidence the pivot is working:

Question 3: The 10x Math

What's the bull case market cap? And what's our 10x entry price?

Bull Case ScenarioValueBasis
Total leasable capacity920 MWCurrent portfolio
Utilization (full)100%All MW contracted
Revenue per MW per year$200-300KIndustry hosting rates
HDC revenue at full utilization$184-276MBase hosting revenue only
Plus services/premium/escalation+50-100%Power pass-through, services
Total HDC revenue at scale$300-550M920 MW fully deployed
HDC gross margin at scale40-50%Currently 30%, improving
EBITDA at scale$120-275MRough estimate
EBITDA multiple (AI infra)20-30xPremium infrastructure
Implied EV (920 MW only)$2.4-8.3BCurrent capacity

At 920 MW only, bull case EV is $2.4-8.3B. With expansion beyond 920 MW, the numbers get bigger. The question is: how much can they expand, and how fast?

Expanded Bull Case (Capacity Expansion)

ScenarioCapacityRevenueEBITDAEV (25x)
Current (920 MW deployed)920 MW$300-550M$120-275M$3-6.9B
Moderate expansion2 GW$600M-1.2B$240-600M$6-15B
Aggressive expansion3-4 GW$1-2B$400M-1B$10-25B

Aggressive expansion bull case: $10-25B EV. This requires CORZ expanding to 2-4+ GW, achieving premium AI infrastructure multiples, and diversifying beyond CoreWeave. Our 10x entry = bull case ÷ 10 = $1-2.5B market cap ($3-8/share).

Bull Case and 10x Entry

Bull case market cap: $9-25B (920 MW at AI multiples on the low end, 2-4 GW expansion on the high end). This requires the pivot to succeed, customer diversification, and capacity expansion.

10x entry = $1-2.5B market cap ($3-8/share). The 52-week low was $6.20 — inside this range. At the 52-week low, CORZ was a legitimate 10x candidate if you believed the bull case.

Key assumptions for the bull case: (1) CoreWeave contract delivers as expected, (2) CORZ signs 2+ additional major AI/HPC customers, (3) capacity expands beyond 920 MW, (4) no liquidity crisis before HDC revenue ramps. Our probability: 40%.

Question 4: Where Is the Floor?

Floor estimate: $2-3B enterprise value (but equity floor is much lower due to debt).

Floor ComponentValueNotes
PP&E (Net)$1,293MData centers, power infrastructure
Bitcoin Holdings$222MLiquid
Cash$311MDeclining
Deferred Revenue$556MContracted, partially funds assets
Other Current Assets$362MCustomer funding receivable
Gross Asset Value~$2.7BTangible assets + contracted revenue
Less: Convertible Debt-$1,085M
Less: Current Liabilities-$781M
Residual Equity Value~$850M~$2.80/share

The floor is LOW. Due to $1.085B in convertible debt and negative equity, the floor for equity holders is maybe $2-3/share in a liquidation scenario. The warrant liabilities ($936M) are non-cash and would likely be worth much less in a distress scenario (warrants are worthless if the stock is near zero). But the operating floor -- with the pivot continuing -- is probably $2-3B EV, which translates to maybe $4-7/share equity after debt claims.

Downside Risk Assessment

At $16.24, the downside to the floor is $10-13/share (60-80%). This is a HIGH-RISK investment. The negative equity, heavy debt load, single customer concentration, and ongoing cash burn create a genuine possibility of significant loss. This is not a "buy it and forget it" stock. It requires active monitoring of the CoreWeave relationship, cash burn trajectory, and customer diversification progress.

8. Competitive Landscape: Bitcoin Miners Pivoting to AI

CompanyPowerAI/HPC ProgressMarket CapCORZ Advantage?
CORZ (Core Scientific)~1.4 GW590 MW contracted to CoreWeave. Real revenue ($65M).$5.1B
IREN (Iris Energy)~1 GWBuilding GPU cloud. Smaller scale.~$3BCORZ has more capacity + actual contracts
RIOT (Riot Platforms)~1 GWFocused on mining. Less AI progress.~$4BCORZ is years ahead on pivot
CLSK (CleanSpark)~0.9 GWPure mining focus.~$3BCORZ diversifying away from mining
HUT (Hut 8)~0.8 GWMining + AI cloud, moderate scale.~$2BCORZ has larger infrastructure
CIFR (Cipher Mining)~0.7 GWSome HPC exploration.~$1.5BCORZ has established HDC revenue

CORZ's competitive advantages among miners: (1) Largest power capacity, (2) furthest along in AI/HPC pivot with actual revenue, (3) contracted pipeline ($556M deferred revenue), (4) activist investor pushing for value maximization. The disadvantage: 100% customer concentration vs peers who haven't yet committed to a single customer.

9. Market Positioning

Market MetricValueInterpretation
Short Interest63.1M shares (20%)Heavy short interest = high skepticism
Beta6.9Extremely volatile (7x market moves)
52-Week Range$6.20 - $23.633.8x range, low to high
Analyst ConsensusStrong Buy (1.27/5)17 analysts, very bullish
Price Target (Mean)$26.40+63% upside from current
Price Target (High)$40.00+146% upside
Forward P/E38.7xOn $0.42 forward EPS

20% short interest is significant. The shorts are betting on: CoreWeave risk, cash burn, failed pivot, or general Bitcoin downturn. If the HDC revenue ramp accelerates, the short squeeze potential is real -- 20% short + beta of 6.9 = explosive moves in either direction.

10. Putting It Together: Why Leopold Is Doing This

The Leopold Thesis (Reconstructed)

Leopold published "Situational Awareness" in 2024 arguing that AGI is coming by ~2027 and will require orders of magnitude more compute. His entire portfolio is positioned around this thesis: BE (power supply), CORZ (power infrastructure/data centers), CRWV/CoreWeave (GPU cloud), INTC (chip manufacturing), EQT (natural gas). CORZ is the only one where he went activist.

Why activist on CORZ specifically?

  1. The assets are worth more than the stock price suggests. 1.4 GW of power capacity, valued at Bitcoin miner multiples (~$5M/MW), should be re-rated to AI infrastructure multiples ($10-15M/MW) as the pivot completes.
  2. The CoreWeave merger would have given the assets away. CoreWeave tried to buy CORZ at what Leopold considered a discount to intrinsic value. By blocking the merger, Leopold preserved optionality.
  3. Customer diversification is the catalyst. If CORZ signs even one additional hyperscaler beyond CoreWeave, the 100% concentration risk drops dramatically, and the stock re-rates.
  4. AGI demand will be insatiable. Leopold believes AI compute demand will far outstrip current projections. Every MW of data center capacity will be needed. CORZ's 1.4 GW is not enough -- they'll need to expand, and the market will pay premium multiples for that capacity.

Leopold's implied target: If he believes the power assets are worth $10-15M/MW in an AI infrastructure context, then 920 MW leasable = $9.2-13.8B EV. Minus ~$1B net debt = $8.2-12.8B equity. At ~306M diluted shares = $27-42/share. He bought at ~$14 avg cost. As an activist with 9.4%, he's not just waiting for re-rating — he's actively pushing for customer diversification and value realization.

11. Final Assessment

VERDICT: WATCHLIST

What's Real

  • 1.4 GW of power capacity = genuinely valuable in AI era
  • Pivot from mining to HDC is underway with real revenue ($65M, +168%)
  • 590 MW contracted, $556M deferred revenue
  • Leopold's 9.4% activist stake provides strategic pressure and credibility
  • Largest and furthest-along Bitcoin miner in AI/HPC pivot
  • PP&E tripled to $1.3B -- the build is happening

What's Dangerous

  • 100% HDC revenue from ONE customer (CoreWeave)
  • Negative stockholders' equity (-$963M)
  • $1.085B convertible debt, $936M warrant liabilities
  • Cash burning rapidly ($311M left, -$526M last year)
  • 20% short interest, beta of 6.9
  • Emerged from bankruptcy just 2 years ago
  • Stock may already price in the bull case at $5.1B

The Math

Bull Case EV (920 MW at AI multiples)$9-14B
Bull Case EV (expanded to 2+ GW)$15-25B
10x Entry Market Cap$1-2.5B ($3-8/share)
Current Market Cap$5.12B ($16.24/share)
Floor (equity value)$0.8-2B ($2.80-6.50/share)
52-Week Low$6.20
Key AssumptionsCoreWeave delivers + customer diversification + capacity expansion

This is the most complex company we've analyzed. The thesis is real: Bitcoin mining infrastructure being repurposed for AI/HPC hosting is a genuine transformation, and the power assets are undervalued if the pivot succeeds. Leopold's activist involvement is the strongest smart-money signal we've seen -- he's not just buying, he's trying to reshape the company.

Our 10x entry zone is $3-8/share ($1-2.5B market cap). The stock touched $6.20 at its 52-week low — inside this range. When it does trade in this zone again (market crash, bad quarter, CoreWeave scare), we need to be ready with this analysis already done. The single-customer concentration is the biggest risk — if CoreWeave stumbles, the stock could drop into our entry zone, which would be the time to assess whether the underlying power assets still have value.

What would change our mind:

Leopold's activist position validates the thesis — the power assets are real and the AI pivot is happening. His cost basis (~$14) and 9.4% stake give him influence over the outcome. The question for us is simply: at what price does the risk/reward become asymmetric in our favor? That price is $3-8/share.

Data sources: SEC EDGAR XBRL (CIK 0001839341), 10-K filing (FY2025), yfinance, Situational Awareness LP 13D/13F filings, corescientific.com. Analysis date: 2026-03-12.