Axel Lynch

$NBIS Q4: Execution Validates the Thesis 🎉 ● Key metrics ▪︎ Revenue $227.7M vs ~$245M est 🔴 (miss) ▪︎ ARR $1.25B vs $900M-$1.1B guide 🟢🟢🟢 (+14% beat) ▪︎ Active MW 170 vs 100 baseline Q3 🟢🟢🟢 (+70%) ▪︎ Adj EBITDA 24% margin (vs 15-17% expected) 🟢🟢 ▪︎ Operating Cash Flow +$834M (first positive quarter) 🟢🟢🟢 🎉 ● 2026 guidance ▪︎ $7-9B ARR maintained 🟢 (implies 5.6-7.2x growth) ▪︎ >3GW contracted 🟢🟢 (raised from >2.5GW) ▪︎ >2GW already secured today 🟢🟢🟢 (9mo ahead) ▪︎ 800MW-1GW connected power by YE 🟢 ● Execution signals ▪︎ $MSFT first tranche delivered on time 🟢🟢 ▪︎ $META fully in servicing stage 🟢🟢 ▪︎ $1.6B deferred revenue (96x YoY) 🟢🟢🟢 ▪︎ Capacity sold out Q4 🟢🟢 ▪︎ Deploy speed 40-50% faster than peers 🟢🟢🟢 Stock flat pre-market despite massive execution beat. Revenue ≠ ARR in hyper-growth infra. Contracted ≠ Delivered matters more than narrative momentum. The numbers validate what matters: industrial execution at scale. Deeper analysis post-earnings call.

@f0xh00d Yes, I need to clean up my positions

@Axel_Lynch_Inv Bit over-diversified imo. You have 0.3% that could be $NBIS

Diversification is key 🔑 $NBIS

@AndyVidaa Il l’a perdue tout seul comme un grand

Toute l’aura perdue avec Lewis en milieu de générique

Always nice when you’re net buyer on $NBIS 😋

The market’s irrationality always creates new opportunities to buy $NBIS at an attractive price 😚

Penser que la menace IA pour les SaaS se limite aux LLM grand public, c’est rater 80 % du sujet : le vrai choc vient des modèles intégrés aux produits et des LLM privés/open source branchés aux données des boîtes. Autre erreur : croire qu’on est déjà au max des LLM. Compare juste le “Will Smith qui mange des spaghetti” de 2022 à aujourd’hui : on est passé du glitch immonde à la vidéo quasi réaliste en un prompt. On est encore très tôt dans la courbe. youtube.com/shorts/7zdVCQ5…

🤖 C'est ça qui doit remplacer tous les softwares ? Put*in je vais faire ce que je veux à la main ça ira plus vite... Long $ADBE

@gabrelyanov The $NBIS community has been pretty healthy so far, focused on execution, MW ramp and capital structure. We don’t need “bigger than Amazon” narratives. If the thesis plays out, the upside is already huge. Let’s keep it analytical.

No. It's going to be much bigger. $NBIS

Always a pleasure to see the market give us another opportunity to buy an exceptional company at a discounted price. $NBIS

You’re right. $160 is more a strategic milestone than a valuation ceiling. For me, it represents the first confirmation that “stage one” of the thesis has played out successfully, execution at scale validated, 2026 materially de-risked. When I say I won’t sell before $160, it’s not about anchoring to a number. It’s about not front-running my own thesis. At that point, I’ll reassess based on fundamentals, execution and forward risk/reward. Price is a checkpoint. Thesis evolution is what will drive my allocation.

@Axel_Lynch_Inv Allocating based on risk/reward rather than arbitrary exits is how you actually capture the fat tail. Most people treat price targets like a hard stop and miss the re-rating that happens once the market realizes the scale of the AI infra moat.

I genuinely enjoy researching other companies. But today, I don’t see a better place to allocate my capital than $NBIS. Added again at the open yesterday. No intention of selling a single share before $160. And even at $160, if it’s still the best risk/reward over the coming years, I’m fully comfortable staying 100% allocated. Price is just a milestone. Thesis is everything.

@cryptospeed05 sometimes, less is more

@Axel_Lynch_Inv looks like I'm not the only one lol $NBIS

100% $NBIS. High conviction. Thesis intact. Now comes the hard part: waiting. Capital concentrated. Curiosity, not so much. So I’ve been quietly digging into a few names that caught my eye: $KRKNF $LITE $ASTS $SATS $ONDS $EOSE What’s your highest-conviction idea outside $NBIS?

@JoseGalp86 What's your thesis?

@Axel_Lynch_Inv $OPEN

Le Q4 m'a vraiment convaincu. La seule crainte possible était l'exécution, le fait qu'ils soient en avance me fait vraiment envisager l'année à venir extrêmement sereinement. J'avais ciblé 160$ fin 2026 avant les résultats, je suis un peu plus haut à présent mais c’est surtout que ma conviction qu’ils puissent atteindre 160$ a nettement augmenté.

@Axel_Lynch_Inv J’ai basculé à 94$ j’en avais déjà sur CTO à 89$. Par contre j’ai une grosse pondération. Je ne compte pas renforcer avant un repli important et si liquidité j’ai. Le Q4 m’a aussi rendu optimiste. As-tu une target ?

Il y a quelques semaines, j’ai pris une décision radicale. J’ai vendu l’ensemble de mes positions sur ce compte pour être quasiment à 100 % exposé à $NBIS. Est-ce que c’est judicieux ? Honnêtement, je ne le sais pas encore. C’est un choix assumé de concentration. Un pari sur une thèse précise, un management, une dynamique sectorielle. Performance annuelle actuelle : +7,78 %. Avec un alpha significatif par rapport à mon CTO global, au Nasdaq et au S&P 500. Mais la vraie question n’est pas la performance court terme. C’est la capacité de la thèse à se matérialiser sur 2 à 4 ans. Concentrer, c’est accepter plus de volatilité. Mais c’est aussi accepter que la performance vienne d’une conviction forte. On verra si c’était une erreur… ou un tournant stratégique.

@JoseGalp86 Thanks for answer, I’ll add it to my deep search list

@Loris_Dalleau @0xtechquity Quand je demande aux gens autour de moi quel serait le dernier abonnement qu’ils supprimeraient, à chaque fois, la réponse c’est Netflix

@0xtechquity On va dire que l'échec cuisant sur Paramount l'a peut-être refroidi sur le secteur. D'ailleurs, je ne suis pas sûr qu'ils comprennent l'avantage concurrentiel de Netflix (ils ont sans doute peur que ce ne soit qu'une mode de dix ans)

À la tête de Berkshire, sur quelle valeur seriez-vous prêt à parier 50 milliards de dollars les yeux fermés ?

@DrTilt2 Je serais curieux de savoir ce que tu as acheté. Londres est absolument hors de prix. Bien plus cher que Paris. Constaté le mois dernier et par ailleurs, tous les indicateurs économiques et statistiques l'indiquent également

Ya vraiment un probleme en france. Je suis a Londres depuis samedi, je suis allé 2 fois au supermarché et 80% des produits sont moins cher qu'en france. La vie nest pas bon marché a Londres et cest quand meme moins cher qu'en france...cest hallucinant.

NEBIUS $NBIS - Q4 2025: 🚀 The First Proof of Execution Just Arrived. $NBIS had one question to answer in Q4: execution. They answered it, and I'm updating my valuation. Full analysis below. Not demand. That was already signed, $20B+ in contracted backlog with $MSFT and $META. The question was execution. Q4 answered. Every single metric beat guidance. ● What Q4 Validates Q4 answered the question asked in January: "Can they execute?" ▪ ARR: $1.25B (vs guidance $900M-$1.1B) = +14% to +39% above ▪ Active power: ~170 MW (vs 100 MW target) = +70% above ▪ Microsoft tranche 1: delivered November 2025 on time ▪ Meta deployment: February 2026 on time ▪ Group EBITDA: +7% (vs guidance "slightly positive") On Q4, the factual answer is yes. Margins improved during scale-up: Q2 10% → Q3 19% → Q4 24% (core adjusted EBITDA margin). Core revenue +830% YoY. Cash flow became structurally positive: Q4 2025 +$834M (vs -$73M Q4 2024). FY 2025 +$402M (vs -$270M FY 2024). Source: Microsoft/Meta prepayments. ClickHouse revalued at $15B: January 2026 funding round (source Q4 Shareholder Letter page 7). Nebius holds 20-30% = $3-4B. I apply 50% discount (private round, not mark-to-market). Financing plan clarified: CapEx $16-20B. 60%+ already financed (cash $3.7B, cash flow, prepayments). ATM not used to date. My modeling: 290-310M shares by end-2026 (15-23% dilution from current 253M base). ● What Q4 Does NOT Validate The x5-6 scale-up to 1 GW remains to be proven. Q4: ~170 MW active. 2026 requires 800 MW-1 GW connected. Why this is industrially complex: ▪ 9 new sites announced Q4 (16 total) ▪ Multi-jurisdiction (US + EMEA) ▪ Permitting, GPU supply chain, network interconnections ▪ CoreWeave missed Q3 2025 due to delays from a single provider Q4 proves execution at ~170 MW. Not yet at 1 GW. Different order of magnitude. Net ROIC remains theoretical. Core margin 24% is consistent with my theoretical model of potentially very high ROIC. But actual net ROIC won't be measurable until fleet is mature (2027-2028), with visibility on actual maintenance + GPU refresh costs. Current positive cash flow comes largely from prepayments, not just recurring operations. Client concentration 80%+ persists. Microsoft + Meta = 80%+ of backlog. Q1 pipeline >$4B confirmed, but no new mega deal announced. ● Why the Market Remains Cautious Current multiple: EV $25.2B (Market Cap $24.8B + Net Debt $0.4B) / 2026 Revenue $3.0-3.4B = 7.4-8.4x EV/Revenue forward The market values on revenue (7-8x) rather than ARR because it awaits: ▪ Scale 170 MW → 1 GW proven (x5-6 in 12 months) ▪ Positive cash flow ex-prepayments (mature operations) ▪ Visible net ROIC post-maintenance This is rational. Once these conditions are validated, the market could re-rate to premium ARR multiple. ● My Valuation Model Current price: $98.01 Estimated fair value: $145-175 Core: ARR $8B × 5-6x = $40-48B 5-6x multiple on ARR = ~14-18x revenue. This is premium and requires: ▪ 800 MW-1 GW connected end-2026 ▪ Positive cash flow ex-prepayments H2 2026 ▪ No major Microsoft/Meta delays ▪ Net ROIC confirmed once fleet mature Subsidiaries: $6-8.5B × 50% discount = $3-4.3B Dilution: 290-310M shares by end-2026 (15-23% dilution from current 253M base). Stress 25-35% possible if cash flow disappoints or markets close. Calculation: EV $43-52B / 290-310M = $139-179/share My range $145-175 assumes dilution 290-300M. If dilution 320-340M, fair value $125-160. Important: This valuation assumes successful 2026 execution, not a neutral case. ● Asymmetry and Scenarios The asymmetry: ▪ Upside (if execution succeeds): $145-175 = +48-79% ▪ Severe downside (if scale-up fails): $75-95 = -23% ▪ Asymmetry: ~2.5:1 upside vs downside I assign 60-70% probability to successful execution scenario. This implies: ▪ Expected upside: 65% × +60% = +39% ▪ Expected downside: 35% × -23% = -8% ▪ Expected value: +31% This probability-weighted return justifies concentrated positioning. Scenarios: Base (successful execution): $145-175 ▪ 800 MW-1 GW delivered ▪ Positive cash flow ex-prepayments ▪ Dilution 15-23% Stagnation (slowed execution): $85-100 ▪ MW 400-600, ARR $5-6B ▪ Multiple stays 3.5-4x revenue ▪ Dilution 20-25% Severe (scale-up failure): $75-95 ▪ MW <400 ▪ Client restructure or dilution >30% ▪ Multiple compresses to 3x ● Key Risks Industrial scale-up (critical): Threshold: <600 MW connected AND ARR <$6B end-2026. Pricing compression (critical): If MW pricing drops from $8-12M to $5-7M (sector overcapacity), entire model compresses. 5-6x ARR multiple becomes unsustainable. Threshold: Pricing <$7M/MW on new contracts = major bearish signal. Dilution (critical): Stress 25-35% if cash flow disappoints or markets close. CapEx $16-20B requires massive financing. "60% non-dilutive" announced assumes: ▪ Cash flow stays positive ex-prepayments, ▪ Credit markets accessible, ▪ Asset-backed financing at expected terms. If one fails, equity becomes necessary. Threshold: >25% dilution in single event. Client concentration: Threshold (full exit trigger): Announcement of complete cancellation from Microsoft or Meta = immediate 40-50% reduction. Threshold (partial restructure): Meta delays 6-12 months OR Microsoft reduces scope 20-30% = reassess positioning but not mechanical exit. Monitor for pattern vs one-off adjustment. Sector risks: ▪ Hardware generation changes (accelerated obsolescence) ▪ AI demand normalization (end of "supply-constrained market") ▪ Macro: 3+ hyperscalers cut CapEx 30%+ = sell 50% ● My Positioning 🟢 Position maintained | 100% of portfolio Buy zone: $90-110 12-month target: $145-175 (+48-79%) Why I maintain: - Q4 reduced small-scale execution risk. Metrics exceeded targets. x5-6 scale-up remains to be proven. - Favorable asymmetry. 2.5:1 upside/downside if execution holds. Severe downside $75-95 if failure. - Strong fundamentals. Core 24%, positive cash flow, >50% of 2026 ARR contracted. - Subsidiaries = optionality. ClickHouse $15B (50% discount) = +$12-18/share if IPO 2027-2028. - Active monitoring. Clear alert thresholds. Mechanical reduction to 60-70% if triggered. This concentration is a bet on 2026 execution. Not a certainty. ● Conclusion What Q4 proves: Execution possible at ~170 MW. ARR beat, 24% margins, positive cash flow, Microsoft/Meta on time. What remains to prove: x5-6 scale-up to 1 GW. Cash flow ex-prepayments. Mature net ROIC. Client diversification. My positioning: Maintain 100% assuming execution bet. Fair value $145-175 vs $98 = +48-79% if success. Severe downside $75-95 if failure. 2.5:1 asymmetry. Q1 2026 monitoring. Mechanical reduction if thresholds triggered. Disclaimer: I hold NBIS as my only position (100% of portfolio). This concentration reflects my conviction that post-Q4 fundamentals justify this bet in my personal case. This note does not constitute investment advice. Each investor must determine their allocation according to their risk tolerance.