CRWVNASDAQHOLDFree Access

CoreWeave (CRWV): Gangbangcheon C × Geochajesi 13/20 Do Not Enter — Neocloud #1 · $99.4B Backlog · Nasdaq-100 Inclusion · Meta $21B Contract · 1GW+ Active Power vs GAAP Operating Income Profit-to-Loss Reversal · K-PER Base -23.6% / Conservative -64.6% · Microsoft 67% Customer Concentration · Securities Fraud Lawsuit · Repeat Large Insider Sell-Downs — Re-Evaluate After: GAAP Profit Restored + K-PER Conservative ≥ +10% + Microsoft Share Below 50%

AI GPU cloud neocloud largest by market cap ($64.4B). IPO Mar 2025 ($40), FY2025 revenue $5.13B (+168% YoY), revenue backlog $99.4B (Q1 2026 end). Nasdaq-100 inclusion effective 2026.6.22. GAAP operating income reversed: 2024 +$324M → 2025 -$46M. Microsoft 67% customer concentration. GPU-collateralized debt structure (debt $17.3B+ / equity $4.76B). Securities fraud class-action filed Feb 2026. Co-founder management repeated large sell-downs ($1B+ estimated). Gangbangcheon C × Geochajesi 13/20 (Vol 4, Chart 2, Catalyst 4, Market 3). K-PER (Non-GAAP guidance basis): conservative -64.6%, base -23.6%, optimistic +44.4%. Price $117.61 (6.19), stop $104.50, first target $138.25.

June 26, 2026

Core Position

Neocloud #1 by market cap · $99.4B backlog · Nasdaq-100 inclusion · Meta $21B contract expansion · 1GW+ active power — GAAP operating income reversed from profit to loss · K-PER base -23.6% / conservative -64.6% · Microsoft 67% customer concentration · securities fraud lawsuit · repeat large insider sell-downs — Gangbangcheon C × Geochajesi 13/20 = Do Not Enter. Re-evaluate when: GAAP operating profit restored + conservative upside ≥ 10% + Microsoft share confirmed below 50%.

Investment Thesis

CoreWeave (CRWV) is rated 'Do Not Enter' at Gangbangcheon C × Geochajesi 13/20. Steps 1 (AI infrastructure industry growth) and 2 (neocloud #1, B-grade) pass, but Steps 3 (business model/governance warnings) and 4 (GAAP operating income reversal from profit to loss, ROE -146.8%→-50.4%) generate multiple warning signals, and Step 5 (K-PER) clearly fails — conservative scenario upside is -64.6%, base scenario is -23.6%; only under optimistic assumptions (Non-GAAP operating income +55%/yr × 25x multiple) does upside turn positive at +44.4%. Geochajesi 13/20 is in the watch zone, but intersecting with Gangbangcheon C lands in the 'Do Not Enter' matrix cell. Core strengths — $99.4B revenue backlog, Nasdaq-100 inclusion, and major long-term contracts with Meta/OpenAI/Anthropic — support short-term momentum. However, Microsoft single-customer 67% concentration, GPU-collateralized debt structure ($17.3B+ total debt vs $4.76B equity), an ongoing securities fraud class-action lawsuit, and repeated large insider sell-downs (estimated $1B+ cumulative) form a compounded structural risk.

① Non-Financial — Crypto→AI Cloud Dramatic Pivot · Neocloud #1 · Nvidia Strategic Alliance · $99.4B Backlog vs Structural Risks

CoreWeave began as an Ethereum mining company in 2017, pivoted to GPU cloud in 2019, and achieved a Nasdaq IPO in March 2025 ($40/share) — a dramatic growth story. It holds the only 'Platinum' rating in SemiAnalysis ClusterMAX and is Nvidia's first Elite Cloud Service Provider, with industry-leading AI compute operational capability, and is the independent neocloud group's largest by market cap ($64.4B). Key strengths: The $99.4B revenue backlog (Q1 2026 end, up 283% from $25.9B in Q1 2025) provides multi-year revenue visibility. Major long-term contracts with Nvidia ($2B equity investment + priority GPU supply), OpenAI ($11.9B 5-year), Meta ($21B through 2032), Anthropic, and Jane Street are structurally reducing Microsoft dependency. SUNK (Slurm-on-Kubernetes, industry-unique) and 3–5 year take-or-pay contracts create meaningful switching costs. Key weaknesses: Microsoft single-customer concentration at 67% of 2025 revenue is extreme. Dual-class governance (founders retain majority votes) plus large ongoing founder sell-downs throughout 2026 form governance concerns. The February 2026 securities fraud class-action lawsuit is a core risk. → Full 5-layer analysis in the Non-Financial tab.

② Validator — Gangbangcheon C (Steps 1-2 Pass, Steps 3-4 Warning, Step 5 Fail) × Geochajesi 13/20 = Do Not Enter

Gangbangcheon 5 steps: Step 1 (Industry) ✅ AI computing infra TAM CAGR 38%, direct beneficiary of AI datacenter build-out. Step 2 (Market Position) ✅ B-grade — neocloud group #1, only Platinum in SemiAnalysis ClusterMAX. However, within the full market, CoreWeave is a complementary 'overflow capacity' player vs hyperscalers (71% share). Step 3 (Business Model) ⚠️ Scalability criteria met (W&B acquisition, UK expansion, MLOps addition), but multiple management governance warning signals (CEO large sell-downs, Core Scientific acquisition failed twice, securities fraud lawsuit). Step 4 (Financial Quality) ⚠️ Revenue CAGR +373% (2 years) is hyper-growth, but GAAP operating income reversed from +$324M (+17%) in 2024 to -$46M (-1%) in 2025. ROA improving but still negative. FCF negative for 3 consecutive years (growth-stage investment exception applied). Step 5 (K-PER) ❌ Based on FY2026 Non-GAAP operating income guidance midpoint $1.0B: conservative (-64.6%) and base (-23.6%) upsides are negative; only optimistic (+44.4%) is positive → Gangbangcheon Grade C confirmed. Geochajesi 13/20 (Vol 4, Chart 2, Catalyst 4, Market 3) — Watch zone (8–13 range). Gangbangcheon C × Geochajesi 13 matrix = 'Do Not Enter.' → K-PER 3 scenarios and Geochajesi details in the Validator tab.

③ Technical — Near 50-day MA ($117) Inflection Point · Above 200-day MA ($95) · MACD Negative · Box Range ($95–$138) · NDX Inclusion Flow Momentum

Current price $117.61 sits almost exactly at the 50-day MA (~$117) — a trend inflection point. Trading above the 200-day MA (~$95) — long-term uptrend maintained. RSI 57–62, still below the overbought zone (70), leaving room for short-term upside. However, MACD remains negative (~-2.65) and below the signal line, indicating the May high ($138) correction has not fully resolved. 5-day MA (~$109) < 50-day MA (~$117) — partial bearish short-term alignment signals weakening momentum. Key resistance: R1 $138.25 (recent peak), R2 $187 (52-week high). Key support: S1 $104.6–$106.9 (moving average convergence zone), S2 $71.6–$72.0. Nasdaq-100 inclusion (effective 2026.6.22) passive fund inflows plus 6.19 volume 207% above average sustain short-term flow momentum — but technical and fundamental signals point in different directions, creating a mixed-signal environment. Conservative strategy: entry at current price with stop $104.50, first target $138.25 (R:R 1.57:1). → Full chart and scenario details in the Technical tab.

Key Metrics

Price (Analysis Date)

$117.61

2026-06-19 기준 / 시총 ~$64.4B / 52주 -37% 고점 대비

FY2025 Revenue

$5.13B

+168% YoY / 2년 CAGR +373% / GAAP 영업이익 -$46M(적자 역전)

K-PER Conservative Upside

-64.6%

기본 -23.6% / 낙관 +44.4% — Non-GAAP 가이던스 기준

Geochajesi

13 / 20

강방천 C · 진입 금지 (거4·차2·재4·시3)

Revenue Backlog

$99.4B

2026.1Q말 / 2025.1Q $25.9B 대비 +284% / 5개 분기 연속 급증

Next Earnings Date

~2026년 8월 초

FY2026 Q2 / 백로그·가이던스 변화가 가장 큰 트리거

Bull Case

Massive revenue backlog $99.4B — provides multi-year revenue visibility. Q1 2025 $25.9B → Q1 2026 $99.4B (+284% in 5 quarters). Cantor Fitzgerald forecasts a Q2 backlog surprise. 98% of revenue recognized from customer contractual commitments — structural demand visibility.
Nasdaq-100 inclusion (effective 2026.6.22) + volume momentum — passive index funds create structural new buying demand. 6.19 volume 65.55M shares vs daily avg 31.66M (+207%). Nvidia ($2B) and Jane Street ($1B) large strategic equity investments confirm institutional confidence.
Meta $21B, OpenAI $11.9B, Anthropic multi-year, Jane Street $6B compute commitment — major long-term contracts with the top AI ecosystem customers are structurally reducing Microsoft dependency. Company projects Microsoft share below 50% by Q3 2026.
SemiAnalysis ClusterMAX only Platinum + Nvidia Elite CSP — industry-leading AI compute operational capability. SUNK (industry-unique Slurm-on-Kubernetes concurrent operation) proprietary orchestration + MLPerf record (2026.6.16) + 1GW+ active power maintain technical lead.
Direct beneficiary of AI infrastructure TAM hypergrowth — Bloomberg Intelligence projects AI compute infrastructure market CAGR 38% (2023→2028: $79B→$399B). Neocloud segment alone: $23B (2025) → $180B (2030, CAGR +69%). W&B and Monolith AI acquisitions expand from hardware toward full-stack software.

Bear Case

GAAP operating income reversed from profit to loss — 2024 $324M (+17%) → 2025 -$46M (-1%). Capex guidance escalating: 2025 $10.3B → 2026 $30–35B. Revenue exploding but margins deteriorating. ROA improving (-11.9%→-3.5%) but still negative. FCF negative for 3 consecutive years.
Microsoft single-customer 67% concentration + aggressive GPU-collateralized leverage — any Microsoft contract renegotiation/reduction is a direct hit. Q1 2026 end total debt $17.3B+ (still rising), equity $4.76B, debt ratio extreme. Mixed floating (SOFR+2.25%) and fixed (5.9%–9.75%) rate debt simultaneously exposes the company to rate, refinancing, and GPU-generation risk.
Ongoing securities fraud class-action + repeat large insider sell-downs — filed February 2026 (NJ federal court), core allegation: company knew of Core Scientific construction delays months earlier but delayed disclosure. CEO Intrator sold tens of millions per month March–June 2026 (Mar $6.7M, Apr $34.9M, May $39.3M+$30.2M, Jun $37.6M). Multiple founder cumulative sales estimated $1B+. Governance transparency risk may not be fully priced in.
K-PER base and conservative scenarios both negative — using Non-GAAP guidance Y0 ($1.0B): conservative -64.6%, base -23.6%. With GAAP operating income negative, K-PER itself has methodological limitations. On PSR basis: current market cap / FY2025 revenue = 12.6x, vs FY2026 guidance revenue = 5.2x — looks 'cheap' relative to industry peers, but the entire sector's valuations are uncertain benchmarks.
Third-party dependency execution risk already materialized + AI infrastructure bubble debate — Nov 2025 Core Scientific datacenter construction delay triggered guidance cut (-16.3% stock drop) — execution risk has already played out. GPU generation cycle is fast (H100→B200→GB300→Vera Rubin) — net PP&E $38.8B is mostly 3–5 year useful life GPUs. Some analysts compare GPU-collateralized debt structure to the telecom bubble (Level 3, Global Crossing).

Rating:HOLDCRWV

Technical Summary

Price $117.61 sits at the 50-day MA (~$117) inflection point. Long-term uptrend intact above the 200-day MA (~$95), but MACD negative (-2.65) and 5-day MA below 50-day MA (partial bearish short-term alignment) indicate weakening momentum. RSI 57–62 neutral to slightly bullish (below overbought zone). Key resistance $138.25 (recent peak) and $187 (52-week high). Key support $104.6–$106.9 (MA convergence zone). NDX inclusion is the near-term catalyst.

CRWV Technical Analysis — Price Structure, Moving Averages, Support/Resistance, RSI Panel (Key Pivot Simplified)

Support

S1: $104.6~106.9 (다수 이평선 수렴 구간 — 1차 핵심 지지, 손절 기준), S2: $71.6~72.0 (2025.12 저점 인근 주요 지지 — 모델 훼손 손절 기준)

Resistance

R1: $138.25 (최근 1개월 고점, 1차 저항), R2: $187.00 (52주 신고가, 2025.6 전고점 — 2차 목표)

Trend Analysis

Short-term (5-day ~$109): Weak — price above 5-day MA but 5-day below 50-day creates partial bearish short-term alignment. Medium-term (50-day ~$117): Inflection point — price ≈ 50-day MA, pivotal zone. Long-term (200-day ~$95): Bullish — price $117.61 remains above 200-day MA. MA alignment: 200d ($95) < price ($117.61) ≈ 50d ($117) > 5d ($109) — partial bullish alignment, long-term trend solid but short-term momentum weakened. Trend strength: Moderate — long-term indicators favorable, short-term momentum fading.

Momentum & Indicators

RSI (14) ~57–62 — neutral to slightly bullish. Room below the overbought threshold (70) leaves short-term upside potential. MACD: ~-2.65, below signal line — lagging signal reflecting the May peak ($138) correction not yet fully resolved. Direction near the zero line is the next pivot. Bollinger Bands and OBV: data unavailable. Volume: 2026.6.19 65.55M shares (207% above daily avg) surge — concentrated on NDX inclusion catalyst.

Key Technical Points

$117.61 — 50-Day MA Inflection Point (Current Price)

Current price nearly equals the 50-day MA (~$117) — a technical inflection point. Trading above the 200-day MA (~$95). RSI 57–62, below overbought. NDX inclusion flow momentum intact. However, negative MACD creates mixed signals. Choose between split-entry or breakout-confirmation chase strategy.

$104.6–$106.9 — Primary Support Zone (MA Convergence)

Core support zone where multiple moving averages converge. A break below this zone risks extending the decline to secondary support ($71.6–$72.0). In a conservative split-entry strategy, this is the second buy zone (pullback buy on first resistance failure). Stop-loss reference: $104.50 (break of this zone).

$138.25 — First Resistance (Recent 1-Month Peak)

Strong technical resistance as the recent 1-month peak. A volume-confirmed breakout above this level opens the path toward $187 (52-week high). From current price, +17.6% to this resistance. Usable as a short-term target.

NDX Inclusion (2026.6.22) + Volume Surge Signal

Nasdaq-100 inclusion effective 2026.6.22 creates structural passive fund buying demand. 6.16 rally +9.87% on inclusion anticipation, 6.19 volume +207%. Confirmation of actual fund inflow post-inclusion is the key variable for short-term momentum sustainability.

Trading Scenarios

① Current Price Entry — 50-Day MA InflectionR:R 1.57:1 (T1 기준) / 타이트 손절 $109.4 적용 시 2.51:1 ⚠️

Entry

$117.61 (current price, 50-day MA inflection)

Stop

$104.50 (break of primary support, -11.1%)

Target

T1 $138.25 (+17.6%, first resistance) / T2 $187.00 (+59.0%, 52-week high)

R:R 1.57:1 does not meet the standard (2:1). Tight stop (-7%) improves to 2.51:1 but increases noise-stop risk given extreme volatility. Given Gangbangcheon C grade, treat as a small speculative position only.

② Conservative 3-Tranche EntryR:R 1.97:1 (T1 기준) ⚠️

Entry

1/3 at $117.61 + 1/3 at $105.00 (support touch) + 1/3 at $125.00 (near-resistance breakout confirmed) → avg $115.87

Stop

$104.50 (-9.8% vs avg entry)

Target

T1 $138.25 (+19.3%) / T2 $187.00 (+61.4%)

Improved R:R vs single entry. Second tranche ($105) is a support-touch entry, not a break entry — be careful. Third tranche ($125) is momentum-following. The entire position is speculative — recommend limiting to under 1% of portfolio.

③ Aggressive — $138.25 Resistance Breakout ChaseR:R 25.8:1 (기계적 계산) — 단, 타이트 손절로 노이즈 리스크 극단적으로 높음 ⚠️

Entry

~$140 (closing candle confirming resistance breakout)

Stop

Break below $138.25 (prior resistance re-break, -1.3%)

Target

T1 $187.00 (52-week high, +33.6%)

Do not enter without confirmed resistance breakout. Extremely tight stop (-1.3%) creates high noise-stop risk. Given extreme stock volatility, false breakout risk after initial move is significant. Widening stop to $133–$135 gives more realistic R:R.

Bullish Signals

↑

$138.25 breakout on volume → first resistance cleared, momentum toward 52-week high ($187) retake

↑

Post-NDX inclusion (2026.6.22) volume sustaining + institutional net buying confirmed → passive fund inflow materializing

↑

Q2 2026 earnings (early Aug) — backlog further growing (>$110B) + GAAP operating income improvement guidance → catalyst for Gangbangcheon re-evaluation

↑

RSI crosses above 65 + MACD crosses above zero line → short-term momentum fully recovered

↑

Microsoft revenue share confirmed below 60% (10-Q filing) → structural reduction of customer concentration risk

Bearish Risks

↓

$104.50 primary support break + volume surge → execute stop-loss, risk of decline extending to secondary support $71.6–$72.0

↓

Q2 2026 earnings (early Aug) — backlog decrease or guidance re-cut → Nov 2025 replay risk, model damage

↓

MACD further decline + RSI re-falling below 50 → short-term momentum reversal failure, further correction phase

↓

Microsoft revenue share confirmed expanding to 70%+ → customer concentration risk worsening, investment thesis damaged

↓

Securities fraud lawsuit unfavorable ruling / settlement well above market expectations → governance risk materializes, liquidate position immediately

* Technical analysis is based on historical data and does not guarantee future returns. Final investment decisions are your own responsibility.

CoreWeave Scale & Growth Metrics — Infrastructure Expansion (Datacenters · Power) + Quarterly Revenue Backlog Trend

Switching Cost & Moat

Moat Strength by Type

GPU Supply Priority (Nvidia Alliance)

Strong supply access advantage as Nvidia Elite CSP + $2B equity investment + priority GPU allocation. However, entirely dependent on Nvidia policy — a "borrowed moat" that weakens immediately if the Nvidia relationship deteriorates or GPUs become commoditized.

Medium (Borrowed Moat)

Operational Capability (SUNK · Platinum Rating)

SUNK (Slurm-on-Kubernetes concurrent operation, industry-unique), NGC container deployment under 10 seconds, and other measurable performance advantages. Only Platinum in SemiAnalysis ClusterMAX. Repeated MLPerf record-setting validates technical leadership. Core hardware (GPU) fully dependent on Nvidia, but software orchestration advantage is proprietary.

Strong

Switching Costs

3–5 year take-or-pay contracts + SUNK and proprietary orchestration stack create meaningful switching costs. However, GPU compute is inherently commoditizable — customers can leave at contract renewal. W&B (MLOps) acquisition is an attempt to extend lock-in further.

Medium

Cost Structure / Economies of Scale

Nvidia priority supply + large-scale debt financing for GPU cluster construction is more of a capital access advantage than a structural cost edge. No clear structural cost advantage in key expenses (power, cooling, depreciation) — GAAP operating margin of -1% in 2025 reflects this.

Medium

Network Effects

Classic network effects absent in GPU rental. W&B developer community (1M+ users) could create a weak ecosystem effect, but data on lock-in synergy with CoreWeave compute is insufficient.

Weak to None

CoreWeave's most visible moat is the combination of 'GPU supply priority through Nvidia strategic alliance + industry-unique Platinum-rated operational capability.' SUNK (Slurm-on-Kubernetes concurrent operation, industry-unique), CoreWeave Mission Control, and Fleet/Node LifeCycle Controller form a proprietary orchestration stack that creates meaningful switching costs, and 3–5 year take-or-pay contracts enhance revenue visibility. However, the Nvidia preferential supply position — the core element of this moat — is a 'borrowed moat' that depends on Nvidia's policy decisions and could structurally weaken if hyperscalers accelerate their own AI chip development (Microsoft Maia, Amazon Trainium). Network effects are essentially absent, and GPU-as-a-service is inherently commoditizable.

Management & Governance

CEO Michael Intrator — co-founder (since 2017). Background as a commodity trader with no prior cloud/infrastructure experience — a learning-by-doing growth case. Strategically successful in leading the crypto→AI cloud pivot, building the Nvidia partnership, and executing the IPO. However, the consecutive failures of the Core Scientific acquisition ($9B stock-for-stock, rejected by Core Scientific shareholders in June 2024 and October 2025) leaves questions about capital allocation execution capability. Ongoing monthly sell-downs throughout 2026 under 10b5-1 pre-set plans (March–June combined $118M+) are legally fine, but the pattern of large repeated sell-downs near price highs is difficult to view as a long-term conviction signal. External C-suite hires concentrated in 2024 — CFO Nitin Agrawal (ex-Google), COO Sachin Jain (ex-Oracle AI), SVP Engineering Chen Goldberg (ex-Google) — strengthened governance capability ahead of the IPO, but dual-class structure (founders retain majority votes) meaningfully limits the board's effective check-and-balance role. Investor communication weakness: the 'prior knowledge then delayed disclosure' allegation regarding the Nov 2025 guidance cut (core basis of the Feb 2026 class-action) is a clear transparency weakness.

Competitive Landscape

AWS / Azure / GCP / Oracle Cloud

Very High — Indirect/Structural Threat

~71% of global AI computing (Epoch AI). Using CoreWeave as overflow complement while simultaneously building own infrastructure. Microsoft Maia and Amazon Trainium in-house AI chip development threaten long-term Microsoft-directed demand.

Nebius (NBIS)

High — Neocloud #2, Fast Chasing

Closing in on CoreWeave ($64.4B market cap) with 2026 YoY +684% revenue growth and market cap ~$58.6B. On PSR basis Nebius trades higher than CoreWeave — relative valuation comparison looks favorable for CoreWeave, but the entire sector's benchmark reliability is uncertain.

Crusoe / Lambda Labs / Together AI

Medium — Pure Neocloud Competition

Direct competition within the independent neocloud group. Smaller scale than CoreWeave. Positioned to target CoreWeave customers in specific verticals and price-competitive segments. However, significantly inferior in Nvidia Platinum status, contract scale, and capital access.

비트코인 채굴업체 AI 피벗 (Terawulf·IREN·Hut 8)

Rising — Idle Power Asset AI Pivot

Existing power infrastructure and land holdings reduce datacenter conversion costs. Fast entry into the AI computing market. Operationally less mature than CoreWeave, but low-cost strategy could capture certain market segments.

CoreWeave is #1 within the neocloud group, but in the broader AI computing market, hyperscalers (Amazon, Google, Meta, Microsoft, Oracle) hold approximately 71% share — CoreWeave occupies a complementary 'overflow capacity' position. Within the neocloud group, #2 Nebius (NBIS) is closing the gap rapidly in 2026, both in revenue growth rate (YoY +684%) and market cap. Bitcoin miners (Terawulf, Cipher Mining, IREN, Hut 8) are accelerating AI pivots, entering the market using idle power assets. The biggest structural threat is hyperscaler self-developed AI chips (Microsoft Maia, Amazon Trainium, Google TPU), which could long-term erode the overflow capacity demand directed at CoreWeave from Microsoft. CoreWeave's defenses: SUNK orchestration edge, Platinum-rated operational capability, and track record of first commercial deployment of the latest Nvidia GPU generations.

ESG & Summary

Rapid AI datacenter growth makes power consumption and carbon emissions the central ESG challenge. US export controls on GPUs to China paradoxically strengthen domestic AI infrastructure demand — a policy tailwind. Surging power consumption is increasing community and environmental regulatory pressure. Dual-class share structure limits ordinary shareholder governance influence. Class B sunset clause (automatic conversion to Class A on 7th anniversary of IPO or 61 days after Intrator ceases to be CEO) provides some governance protection. The ongoing securities fraud class-action is the largest governance concern.

Key Risks

Microsoft Single-Customer 67% Concentration Risk

67% of 2025 revenue from a single customer Microsoft — more than double the 30% concentration threshold. Direct hit from any contract renegotiation, scale-down, or Microsoft building its own infrastructure. Company projects this will fall below 50% as OpenAI/Meta contracts ramp, but this is an unconfirmed scenario. Microsoft's own AI chip development (Maia) is a structural threat that could long-term erode the external demand directed at CoreWeave from Microsoft itself.

Aggressive GPU-Collateralized Leverage

Q1 2026 end total debt $17.3B+ (still rising), equity $4.76B — extremely high debt ratio. In the GPU-collateralized SPV debt structure (DDTL 3.0/4.0), simultaneous occurrence of rate increases, refinancing difficulty, and GPU-generation obsolescence reducing collateral value creates liquidity crisis risk. Net PP&E $38.8B is mostly 3–5 year useful life GPUs — technology obsolescence speed directly determines asset quality. FY2026 capex guidance $30–35B means continuously piling on aggressive leverage to sustain topline growth.

Securities Fraud Class-Action + Governance Transparency Risk

Filed February 2026 (NJ federal court, class period 2025.3.28–2025.12.15): allegation that the company was aware of Core Scientific construction delays as early as February 2025 but did not disclose until November 2025. A settlement or unfavorable ruling well above market expectations creates direct financial damage. More than the lawsuit itself, the 'prior knowledge then delayed disclosure' transparency issue is a structural weakness in management credibility. Combined with the repeated large insider sell-down pattern, it amplifies governance risk.

Third-Party Datacenter Dependency Execution Risk

Parallel structure of self-built datacenters + third-party datacenter developer partnerships. The Core Scientific-dependent Nov 2025 guidance cut is an already-materialized precedent for execution risk. When third-party construction delays occur, backlog (contracted orders) may not be recognized as revenue — creating a visibility-to-realization gap risk. Expanding self-build to reduce third-party dependency only increases capex burden further — a structural dilemma.

GPU Generation Cycle Speed + AI Infrastructure Bubble Debate

GPU generation cycles are extremely fast (H100→H200→B100/B200→GB200/300→Vera Rubin) — risk of accelerated depreciation and write-downs on non-contracted inventory. Some analysts compare GPU-collateralized debt structure to the telecom bubble (Level 3, Global Crossing), raising the possibility of an AI infrastructure overinvestment scenario. Debate between "near the peak of an AI computing capex super-cycle" vs "still in early stages" remains unresolved.

Gangbangcheon 2/5 passed

Gangbangcheon Grade C — Steps 1–2 pass, Steps 3–4 partial pass with multiple warning signals, Step 5 (K-PER) clearly fails on conservative scenario. Meets 'Steps 1–2↓ pass or insufficient upside' criterion → Grade C confirmed. Gangbangcheon C × Geochajesi 13 = 'Do Not Enter' matrix. Re-evaluation conditions: GAAP operating income restores to profit + conservative K-PER upside recovers to ≥10% + Microsoft share confirmed below 50% + securities fraud lawsuit major risk resolved.

CoreWeave 3-Year Financial Performance — Revenue/Operating Income (Left) · ROA/ROE (Right)

Gangbangcheon 5-Step Checklist

✓

Step 1

Industry Analysis ✅

AI computing infrastructure market TAM CAGR 38% (2023 $79B → 2028 $399B, Bloomberg Intelligence). Neocloud segment alone CAGR +69% (2025 $23B → 2030 $180B, Synergy Research). Direct beneficiary of AI datacenter proliferation (GPU, power, cooling infrastructure). Entry barriers: capex economies of scale + Nvidia priority supply + 3–5 year switching-cost contracts. Note: telecom bubble comparison and overinvestment debate warning noted. Long-term industry growth ✅.

✓

Step 2

Market Position ✅ B-Grade

#1 by market cap within the independent neocloud group ($64.4B). Only Platinum in SemiAnalysis ClusterMAX. Nvidia Elite Cloud Service Provider. However, in the full AI computing market, CoreWeave is a complementary overflow capacity player vs hyperscalers (71% share) — not a market-wide #1, but a segment leader (B-grade). Pricing power: older-gen GPU prices stable/rising, latest GPU largely sold out → 2 of 3 criteria met (margin maintenance criterion not met due to GAAP margin deterioration). B-grade.

✗

Step 3

Business Model ⚠️ Partial Pass

Scalability: platform expansion (W&B adding MLOps), geographic expansion (UK/Europe), vertical expansion (Monolith AI physics simulation) — 3 of 4 criteria met. Leadership warning signals clear: ① 'Prior knowledge then delayed disclosure' allegation re Nov 2025 guidance cut (core basis of securities fraud lawsuit) ② Repeated large co-founder sell-downs including CEO (2026 Mar–Jun $118M+) ③ Core Scientific acquisition failed twice (capital allocation execution questioned). Business model structure itself meets criteria, but management/governance section has clear warning signals → Partial pass.

✗

Step 4

Financial Quality ⚠️ Passes but Trend Clearly Deteriorating

Revenue: $229M ('23) → $1,915M ('24) → $5,131M ('25), CAGR +373% (2 years) — top-tier topline growth. Operating income: '24 $324M (+17%) → '25 -$46M (-1%) — profit-to-loss reversal is the core warning. ROA -11.9%→-3.5% (improving but negative). ROE -146.8%→-50.4% (small equity base distorts ratio). FCF negative for 3 consecutive years — growth-stage investment exception applied (forced D condition excluded). P×Q-C decomposition: topline-growth type (Quadrant 2) — revenue exploding but margins actually deteriorating. Forced-D checks: explainable BM, FCF growth-stage exception, no share-loss trend, no confirmed executive fraud (lawsuit ongoing) → Forced-D not triggered, but warning noted.

✗

Step 5

K-PER Upside ❌ Fail

Basis: FY2026 Non-GAAP Adjusted OI guidance $1.0B (Year 0). Conservative (+15%/yr · 15x): FY2029E $1.52B → target cap $22.8B → upside -64.6%. Base (+35%/yr · 20x): FY2029E $2.46B → $49.2B → -23.6%. Optimistic (+55%/yr · 25x): FY2029E $3.72B → $93.0B → +44.4%. Current market cap $64.4B. Conservative and base upsides negative → Step 5 fail ❌. Only optimistic assumption turns positive. PSR cross-check: current cap/FY2025 revenue = 12.6x, vs FY2026 guidance = 5.2x — looks "cheap" in relative comparison, but sector benchmark reliability is uncertain.

K-PER Scenario Analysis (3-Year Target)

Basis: FY2026 Non-GAAP Adjusted Operating Income guidance midpoint $1.0B (Year 0). 3-year forward FY2026→FY2029. GAAP operating income is negative (-$46M), so Non-GAAP guidance is substituted — methodological limitation. K-PER multiples 15–25x (applying the lower end of the high-growth innovative category 20–35x range to reflect leverage and governance risk). Current market cap ~$64.4B (price $117.61). Conservative and base upsides negative → Grade C confirmed.

Scenario	Annual Growth	Non-GAAP Profit	Applied PER	Target Cap	Upside
Conservative Scenario	+15%/년 (3년)	FY2029E Non-GAAP OI $1.52B	15x	$22.8B	-64.6%
Base Scenario	+35%/년 (3년)	FY2029E Non-GAAP OI $2.46B	20x	$49.2B	-23.6%
Optimistic Scenario	+55%/년 (3년)	FY2029E Non-GAAP OI $3.72B	25x	$93.0B	+44.4%

Geochajesi Score (13/20)

Volume / Flow4/5

2026.6.19 volume 65.55M shares vs estimated 20-day avg 31.66M (+207%). NDX inclusion (effective 2026.6.22, announced 2026.6.11) — passive fund structural new buying demand. Nvidia ($2B, Jan 2026) and Jane Street ($1B, Apr 2026) large strategic equity investments confirm institutional/strategic accumulation. However, intraday execution intensity and exact 20-day avg not directly accessible — not reflected. Score: 4.

Chart / Technical2/5

Extreme 52-week range (-66%~+193%). Short-term rebound from ~10% correction in past month, including +9.87% surge on 6.16 NDX anticipation. MACD negative (-2.65), 5-day below 50-day in partial bearish alignment — low trend consistency. Above 200-day MA (~$95), long-term trend intact. Accurate MA levels and RSI divergence not directly accessible — conservative scoring. Score: 2.

Catalyst / Material4/5

Multiple A-grade catalysts: NDX inclusion confirmed (structural/long-term, +2) / Meta $21B contract expansion (Mar 2026) / Jane Street $6B compute+$1B equity (Apr 2026) / MLPerf record (Jun 16, 2026) / Cantor Fitzgerald Q2 backlog surprise forecast (Jun 15–16, 2026). Long-term catalyst persistence adds points. Simultaneously, ongoing securities fraud lawsuit and large insider sell-downs are negative catalysts. Net catalysts are positive-dominant but significant negative noise → withhold full score. Score: 4.

Market / Sector3/5

S&P500 up ~+7% YTD as of 2026 (CNBC, May 2026), CRWV up ~+80% over same period — significant outperformance. However, same-day and prior-day Nasdaq/S&P500 moves on 2026.6.21 not directly accessible — conservatively assigned neutral value. Overall market uncertainty considered. Score: 3.

Total13/20

Entry Strategy (3 Tranches)

Current Price Entry (Technical / Speculative)$117.61

50-day MA inflection technical entry. Limit to under 1% of portfolio as a small speculative position. Given Gangbangcheon C, do not increase position without fundamental improvement confirmed.

Conservative Split Entry on Primary Support Touch$104.6~106.9

MA convergence zone (primary support). Recommended to enter after reversal candle confirmed on support touch. Stop $104.50 (support break level). Maintain small position given Gangbangcheon C grade.

Exit Triggers

Break below $104.50 (primary support) → execute stop-loss, risk of decline extending to $71.6–$72.0

Q2 earnings: backlog decrease or guidance re-cut → Nov 2025 replay, model damaged → liquidate immediately

Securities fraud ruling / settlement well above expectations → governance risk materializes → liquidate immediately

Microsoft revenue share confirmed at 70%+ → concentration risk worsening → reduce immediately

⚖

Portfolio Weight Recommendation

Gangbangcheon C × Geochajesi 13 = Do Not Enter. For existing holders: consider reducing position unless taking a long-term DCA approach. New entries are strictly speculative — limit to under 1% of portfolio. Minimum conditions for increasing position: GAAP operating profit restored + conservative K-PER upside recovers to ≥10%.

Editor Note

CoreWeave is both the most dramatic beneficiary of the AI infrastructure boom and the most controversial stock within it. The $99.4B backlog, Nasdaq-100 inclusion, and Nvidia alliance provide compelling demand visibility that makes the bull case persuasive. However, the FY2025 GAAP operating income reversal from profit to loss, K-PER base and conservative scenarios both negative, Microsoft 67% customer concentration, ongoing securities fraud lawsuit, and repeat large co-founder sell-downs create structural barriers too high to justify new buying at this price. If Q2 2026 earnings (early August) confirm GAAP operating income improvement guidance and continued backlog growth, Gangbangcheon re-evaluation becomes possible. Until then, "watch and learn" is the most rational position.

Financial Data

CoreWeave uses a calendar fiscal year (Jan 1–Dec 31). FY2025 complete ($5.13B revenue). K-PER methodology note: Because GAAP operating income is negative (FY2025 -$46M), K-PER uses the company's FY2026 Non-GAAP Adjusted Operating Income guidance midpoint $1.0B as the base (Year 0) — methodological limitation explicitly noted. Current quarter in progress: FY2026 Q2 (Apr–Jun). Q1 2026 earnings reported (revenue $2.08B, +111.7% YoY). Next earnings: FY2026 Q2 (~early August 2026).

Period	Revenue	Growth	Op. Income	Op. Margin
FY2023ROA -11.9% (year-end assets basis). ROE incalculable (negative equity: -$131.9M). Active power 70MW, 10 datacenters. Early stage of crypto→AI cloud pivot — before full revenue ramp-up.	$229M	N/A (기준년도)	미공개 (자본잠식 상태)	N/A
FY2024ROA -7.6%, ROE -146.8% (small equity base ~$1.3B distorts ratio). Cross-model Quadrant 2 (scale-growth): revenue surge + GAAP operating profit achieved. Active power 360MW, 32 datacenters. GPU-collateralized debt model established; first year of direct AI boom benefit.	$1,915M	+736% YoY	+$324M (+16.9% 이익률)	+16.9%
FY2025ROA -3.5%, ROE -50.4% (equity slightly grew but still negative due to losses and leverage). GAAP operating income reversal from profit to loss is the core warning signal. Active power 850MW, 43 datacenters. Capex $10.3B. Backlog $66.8B (year-end). Nov 2025: Core Scientific construction delay triggered guidance cut.	$5,131M	+168% YoY	-$46M (-0.9% 이익률 — 흑자→적자 역전)	-0.9%

GAAP vs Non-GAAP Note

The K-PER base value uses the company's FY2026 Non-GAAP Adjusted Operating Income guidance midpoint ($1.0B), not GAAP. This is because FY2025 GAAP operating income (-$46M) is negative, making direct application of the K-PER formula (current operating income × growth rate) produce meaningless results. This entails a methodological limitation as it is a guidance-based estimate, not a GAAP figure. Also note that Non-GAAP excludes non-cash charges (SBC, depreciation, etc.), making profitability appear better than the actual economic burden — keep this in mind. Current market cap ~$64.4B (price $117.61, ~545.6M shares outstanding).

Key Valuation Metrics

FY2026 Non-GAAP Operating Income Guidance

$900M~$1.1B

Midpoint $1.0B used as K-PER anchor. Methodological limitation noted: based on guidance and Non-GAAP, not GAAP. FY2026 Q1 revenue $2.08B (annualized ~$8.3B), FY2026 full-year revenue guidance midpoint ~$12.5B.

Revenue Backlog Trend

$25.9B → $99.4B

5 quarters 25Q1→26Q1 +284% surge. Quarterly: 25Q1 $25.9B → 25Q2 $30.1B → 25Q3 $55.6B → 25Q4 $66.8B → 26Q1 $99.4B. 98% based on customer contractual commitments. Cash-conversion risk if construction delayed.

PSR Cross-Validation

12.6배 / 5.2배

Current market cap ($64.4B) / FY2025 revenue ($5.13B) = 12.6x. Current market cap / FY2026 revenue guidance ($12.5B) = 5.2x. PSR used as a supplement given K-PER's methodological limitations (negative GAAP operating income). However, whether the broader sector valuations are reliable benchmarks is itself uncertain.

Infrastructure Scale Growth

데이터센터 10→43 / 전력 70MW→1GW+

2023→2025 datacenter count 4.3x (10→43). Active power 2023 70MW → 2024 360MW → 2025 850MW → Q1 2026 1GW+ (14x+ in 3 years). $1M+ spending customers +150% YoY in FY2025. Topline scale metrics are impressive, but the gap from margin deterioration is the central challenge.

K-PER Conclusion

기본·보수 모두 마이너스

Conservative (+15%/yr · 15x) -64.6%, Base (+35%/yr · 20x) -23.6%, Optimistic (+55%/yr · 25x) +44.4%. Gangbangcheon Grade C confirmed — clearly fails the conservative-scenario ≥10% upside threshold. Only under optimistic assumptions (Non-GAAP operating income +55%/yr + 25x multiple) does upside turn positive.

FY2026 Q2 Earnings Date

~2026년 8월 초

Key checkpoints: additional backlog growth (>$110B?), whether GAAP operating income begins to improve, whether Microsoft revenue share is actually declining, and whether construction schedule delays recur.

* GAAP basis. All figures are estimates based on public information and are not investment advice.

Same Exchange

← Back to Company Analysis