AI-01 — Grinding Through 7 AI Compute Stocks: CRWV / NBIS / IREN / CIFR / WULF / APLD / SLNH. 2026 Compute, Power Capacity, Revenue Explosion

Date: 2025-12-03 (paid tier) Tickers: CRWV, NBIS, IREN, CIFR, WULF, APLD, SLNH Source: YouTube paid tier

Summary

• Big seven AI infrastructure plays compared on every meaningful metric. Took ~2 weeks to compile and verify. CRWV (CoreWave) is the giant ($36.4B mcap), then NBIS ($24B), IREN ($13.5B). Mid-tier: CIFR/APLD/WULF ($6.5-8B). SLNH (Soluna) is the small/speculative one ($1.2B).
• Speaker's top picks (in order): (1) NBIS — strongest software, deepest team (1,370 employees), low debt, 6-15x revenue growth to 2026. (2) IREN — vertically integrated, 100% green, biggest 2026 capacity (2.3 GW), starting to build software, PS only 12 (cheapest). (3) WULF — only 12 employees but Google/Fluid Stack-backed, 400% revenue growth, 14% insiders, 96% institutional (institutions love it). (4) Speculative: SLNH — waiting to get hit with a big order; if it does, multi-bagger; if not, dilution death spiral.
• Risk callouts: CRWV has 8.8x leverage (highest), 70% of OpenAI's $22B order is CRWV's — and OpenAI is in the eye of the storm. APLD has 70% revenue concentration from CRWV (so APLD goes down if CRWV does). CIFR has a revenue gap 2026-2028 — market may de-rate once visible. SLNH has high dilution (7x share count growth in 24-25), highest leverage ratio after CRWV.
• 2026 revenue growth (vs 2025 base): CRWV 700%, NBIS 600-1500%, IREN 300%, CIFR 300%, WULF 400%, APLD 500%, SLNH 300%. PS multiples: SLNH 3.6 (cheap, weak), IREN 12 (cheap, strong), CRWV 23, WULF/CIFR/APLD ~30s, NBIS 43.

Translation

Hello everyone. Today we're comparing the AI-era infrastructure star stocks side by side — their fundamentals, recent growth, big contracts, risks. The spreadsheet I built took nearly two weeks of research and cross-checking.

The seven companies covered: - CRWV (CoreWave) - NBIS (Nebius) - IREN - CIFR (Cipher) - APLD (Applied Digital) - SLNH (Soluna) - WULF (TeraWulf)

Today is December 3, 2025, Wednesday. Yesterday many AI stocks took a hit: CIFR -10%, IREN -15%, WULF -7%. Pretty steep. But for high-growth names like these, 10-15% swings are something I've stopped paying attention to. What matters is multi-year growth.

Caveat: all my videos are personal investment notes. Not investment advice. Do your own thinking, manage your own risk.

Doing this kind of broad comparative work isn't crowd-pleasing — most channels pick two stars and analyze them deeply. But with this many names, I want to subtract rather than add. I don't have the bandwidth to track all seven. So this is the comparison to figure out who's strongest, who's safest, and who fits which risk profile.

Note: today's data is roughly the past few days, doesn't change much from where we are now (~Dec 3, 2025).

Market caps (size hierarchy)

• CRWV: $36.4B — biggest by far
• NBIS: $24B — about 2/3 of CRWV
• IREN: $13.5B — about half of NBIS (after yesterday's drop, slightly less)
• CIFR: $8B
• APLD: $7.6B
• WULF: $6.5B
• SLNH: $1.2B

CIFR/APLD/WULF cluster at the same tier (~$6-8B), all about 1/3 of NBIS. NBIS and IREN are my two favorites; I'll explain why.

Headcount

• NBIS: 1,370 — most. Massive team.
• CRWV: 881
• IREN: 457
• APLD: 205
• CIFR: ~40
• SLNH: ~40
• WULF: just 12

These are tech companies — headcount isn't a burden, it's capacity. NBIS is full-stack with cloud services AND developer tools (think AWS-style infrastructure for AI builders), plus they're now expanding into autonomous driving. NBIS originally ran an internet portal — software was always their strength. So 1,370 engineers is a serious moat.

WULF having only 12 employees is unusual. They outsource everything that can be outsourced. Initially I marked this red (risk), but given Google/Fluid Stack and other top-tier companies have placed massive orders with them, I trust they can execute. Lean headcount = lower cost.

Business mix: - CRWV + NBIS are full-stack (hardware + software + token-selling AI services) - IREN is building full-stack — software is currently 20-30%, targeting 40-50% by 2026-2027 - APLD is hardware-focused, vertically integrated (handles land, power, base build, design) but no software ambition - CIFR, WULF, SLNH were all originally Bitcoin mining cos that pivoted to AI compute

NBIS is the strongest software player. CoreWave is second but more hardware-leaning. The rest are pure-or-mostly hardware compute providers.

P/E

Only NBIS shows a P/E (113) — and it might disappear next quarter once they record more depreciation. Everyone else has zero earnings under GAAP.

113 sounds high — it's not. A P/E of 113 means the company just turned profitable — this number normalizes fast as profit scales. Even hundreds wouldn't be alarming for early-stage profitability. The other six having no P/E means they haven't crossed into profit yet.

Most of these companies use cash gross margin in their pitches (excluding depreciation). NBIS's cash gross margin is 71%, but if you bake in depreciation properly, it shrinks 30-40 percentage points. That's industry-typical. Some depreciate even more aggressively.

P/B (price-to-book)

• CRWV: 9.4
• NBIS: 5 — software-dominant, low PB is fine
• IREN: 6 — heavy hardware capex + software building, fair
• CIFR: 10
• APLD: 7.3
• WULF: 20+ — high (asset-light, plus accounting depreciation eats book value)
• SLNH: -11 — negative (high debt, low net assets — financially the weakest)

WULF's high PB reflects that they're asset-light but their pipeline (capacity backlog) is huge — most of it not yet on the books, so book understates the business.

P/S (price-to-sales)

• NBIS: 43 — highest
• CIFR/APLD/WULF: ~30s
• CRWV: 23 — middle
• IREN: 12 — cheap, strong
• SLNH: 3.6 — cheapest, weakest fundamentals

Most of these PS numbers are elevated because the big revenue from contracted orders hasn't hit yet — the market has priced part of 2026 into the cap already. For this kind of company, low double-digit PS is OK, single-digit is great, "thirties" is high. IREN at 12 stands out.

2025 revenue (size)

• CRWV: $15.8B — top
• NBIS: $5.5B — 1/3 of CRWV
• IREN: $11B — 2/3 of CRWV, 2x NBIS (3:2:1 ratio across the top)
• CIFR: $2.5B
• APLD: $2.2B
• WULF: $1.8B
• SLNH: $33M — tiny

Combining mcap and revenue, the strength tier is: CRWV → NBIS → IREN at the top, then everything else.

Business profile detail

CRWV: Full-stack but hardware-leaning. Aggressive expansion. In the eye of the storm (lots of news). Three founders all from finance — "three musketeers." Outsources construction work to APLD (which is why APLD is risk-correlated to CRWV).

NBIS: Full-stack with strongest software. Cloud + developer tools (AWS-like). New autonomous driving business. Software firm operating heritage. My top pick for the software-builder bet.

IREN: Vertically integrated. Two brothers as founders — one with deep capital markets experience (resource for fundraising and credibility), one with mining experience (gets land, power, environmental permits done). 100% green energy. Lowest power costs. Largest 2026 capacity pipeline at 2.3 GW (none of the others reach this in 2026), 2027 plan 2.9 GW.

CIFR: Recently dropped a big AWS order plus the Google/Fluid Stack order. Took ~7-8 months to build out their data centers using mining-era construction speed. Pioneered 300 MW liquid cooling in 2023 (fast iteration). Risk: capacity additions are concentrated in 2025-2026 — then a gap until 2028 (3.2 GW total target). Market may de-rate as the pipeline gap becomes visible.

WULF: Same Google/Fluid Stack backing as CIFR. Even more lean — 12 employees, outsources everything. Pioneered fast liquid-cooling builds. Capacity scaling more gradually (smoother quarter-over-quarter): 600 MW (2025) → 1 GW (2026) → 1.5 GW (2027) → 2 GW (2028). This even capacity ramp means less de-rating risk than CIFR.

APLD: Vertically integrated hardware. Got financing through senior secured notes plus long-term contract-backed debt. Big revelation in mid-2025: $11B / 15-year contract from CoreWave — currently ~70% of APLD's projected 2026+ ARR. Project Alanda capex ~$5.2B (so half of the $11B contract goes to capex). Risk: CRWV concentration. If CRWV stumbles (e.g., OpenAI scrutiny), APLD takes a hit. Second client is undisclosed, $5B / 15-year, 200 MW.

SLNH: Tiny. 2.8 GW future pipeline (similar to IREN), but only ~89 MW operational today. First step: scale 8x to 773 MW. Current size is too small for big-tech orders to land — partly because their power source (wind + solar, dispersed in small/folded modular installations) is intermittent, while CIFR/WULF/APLD have steadier baseload (hydro, nuclear). Wind/solar reliability may be why SLNH hasn't gotten hit with a big order. Heavy share dilution through 2024-2025 (10M → 68M shares, ~7x) because they have only ATM as a fundraising route. Investment thesis is speculative: if a big order lands, they can use it as collateral for cheaper financing, dilution slows, stock spikes. If not, dilution death spiral continues.

2026 ARR (annual recurring revenue) projections

This is the key forward number. ARR = annualized revenue from existing contracts (10/15-year deal divided by years).

• CRWV: $12B (vs $15.8B 2025 revenue → ~7x growth in revenue)
• NBIS: $3.3B baseline, $7-9B if upside (vs $5.5B → 6x to 15x)
• IREN: $3.4B (vs $1.1B → ~3x)
• CIFR: $0.7-1.2B (vs $0.25B → 3-4x)
• WULF: $0.75-0.9B (vs $0.18B → 4-5x)
• APLD: $1.1B (vs $0.22B → ~5x)
• SLNH: ~$0.1B (vs $0.033B → 3x)

The market buys expectations and sells facts, so 2026 ARR is what matters now — that's what's getting priced into 2025 multiples.

Why I prefer NBIS over CRWV

• NBIS revenue growth: 6x baseline, up to 15x upside
• IREN's growth: 3x
• The 4x mcap-to-revenue gap between NBIS and IREN gets wiped by NBIS's higher growth (if 6x baseline = roughly fair vs IREN; if 15x = NBIS is cheaper)

NBIS is my favorite. High growth + low debt + software moat + already starting to show profit. If NBIS sells off hard, that's an entry opportunity.

CRWV: 700% growth is incredible, but the leverage and OpenAI concentration risks are real. If you trust that big-tech buyers (Microsoft, Meta) wouldn't sign $22B orders with a sinking ship, then CRWV looks fine. PS at 23 isn't egregious. But the 8.8x leverage is the highest in the cohort — by far.

Insider ownership

• CRWV: 26.7% — three founders own a quarter, very motivated by future profit
• CIFR: 3% — CEO/exec chair holds little. Hired-gun CEO, not founder-aligned. Mild risk signal.
• NBIS: 16% — solid
• WULF: 14% — also strong
• APLD: 6% — lower
• IREN: 10%
• SLNH: 10%

Retail vs institutional

• WULF: only 4% retail (96% institutional/insider) — institutions love this name
• CRWV: 25% retail
• CIFR: 24%
• APLD: 25%
• NBIS: 43% retail
• IREN: 37%
• SLNH: 81% retail (worrying — institutions haven't loaded up; partially because no big orders yet)

WULF being 96% institutional/insider is a strong vote of confidence. SLNH being 81% retail is a yellow flag.

Closely-held shares (locked / not freely traded)

• WULF: 37% (high — early-stage signal but also lockup commitment)
• SLNH: 24%
• CIFR: 18%
• CRWV: 17%
• APLD: 17%
• IREN: 12%
• NBIS: 6%

Cash position

Most are well-capitalized recently — fresh raises. SLNH is the weakest ($0.55 ratio, in the danger zone) — they need a big contract to unlock cheaper debt or face dilution.

Leverage (ROE/ROA ratio)

• CRWV: 8.8x — highest, real concern
• SLNH: 7.1x — also high (out of necessity, low cash)
• WULF: 4.6x — moderate
• APLD: ~2.x
• CIFR: 2.5x
• NBIS: 1.8x — basically debt-free, super healthy
• IREN: 1.4x — also light, financially comfortable

NBIS and IREN's low leverage = high resilience. CRWV and SLNH carry blowup risk if conditions tighten.

2025→2026 revenue growth summary

• CRWV: 700% (highest, but most concentrated)
• NBIS: 600% baseline (could be 1500%)
• IREN: 300%
• CIFR: 300% (then gap)
• WULF: 400% (smooth ramp)
• APLD: 500% (with CRWV correlation risk)
• SLNH: 300%

Bottom line and my picks

Personal portfolio plan:

1. Hold long, trade in and out (拿拿放放): NBIS — software moat + low debt + growth runway. Buy aggressively on big drawdowns.
2. Pure power-build infrastructure: IREN — biggest 2026 capacity, vertically integrated, lowest PS in the cohort (12), strong team.
3. Short-term play: WULF > CIFR — better capacity ramp, more institutional support, smoother revenue trajectory.
4. Speculative slice: SLNH — small position, betting on a big order landing.

If you want the biggest ship (most dramatic upside, highest risk): CRWV. Its 8.8x leverage and OpenAI dependency are real, but Microsoft, Meta, and others wouldn't sign multi-billion contracts blindly. If you trust the downstream demand thesis, CRWV is fine — buy on dips. The "three musketeers" being all finance guys may put off some — that's a personal preference call.

Investing in CRWV essentially = a leveraged bet on OpenAI succeeding. OpenAI's $22B contract is CRWV's largest. OpenAI is currently: - Cash uncertainty (people still asking where the money comes from) - Customer attrition risk — ChatGPT was the original product, now diluted by Grok, Claude, Gemini, etc. Users splitting across LLMs means OpenAI has fewer training data sources and fewer paying customers per dollar of investment. Customer base is the future moat, and OpenAI's is being eroded.

So CRWV's risk is real on both sides: - Cost side: outsources to APLD (margin risk, vendor risk) - Customer side: 70% concentration in OpenAI (bull-thesis risk)

Soluna is a small bet. If they get a big contract, they can use it as collateral, slow dilution, and the stock 3-5x's quickly. If not, dilution wipes you out. Asymmetric risk, small position only.

Between CIFR and WULF, I prefer WULF. Better metrics across the board, smoother capacity ramp, less de-rating risk.

There you have it — pick what fits your risk profile. Big = CRWV. Quality + growth = NBIS. Cheap + steady = IREN. Sneaky-good = WULF. Lottery ticket = SLNH.

OK that's it for today. A lot to digest. Thanks for watching. See you next time. Wishing everyone financial freedom soon. Bye-bye.