Jake

ngl this ones gonna be a banger

昨天在最低点发现利空很难杀跌，就平了空头开始抄底正股。今天低点附近开了杠杆ETF继续买，拉上去了又追了call。 完全如同我剧本的一次v型反转。 谢谢大家！

$IREN is next, and I think we start to see lift off in the next ~2 weeks. I’m targeting a retest of ATH by June. Nuts on the line for the world to see here. Chart is reminding me of $NBIS back when I longed it in August. *Disclaimer I am long.

Nebius Earnings Through My View And What Most People Will Miss On The Call✍️ So guys, $NBIS just dropped Q1 2026. You all see the numbers, which are decent. Revenue $399M, up 684% YoY, beat consensus. Adjusted EBITDA positive to $130M. The headline net income of $621M is mostly a paper gain on their ClickHouse stake strip that out and they’re still running an adjusted net loss of around $100M, which is fine for a company in heavy build mode. So the print is solid. The deeper things are in the call. When I listened to the call and looked for strategic execution and what they’re actually doing, the following things is what caught my eyes the most, and which will probably be underlighted for most people. Starting with Arkady’s opening. He mentioned that they’re building across four dimensions (capacity, product, customers, and capital.) pay attention to the order here. Capacity first, capital last. His commentary strengthened essentially the neocloud case where Nebius is saying: we will build it, then we will fund it, and the funding will come because of what we’ve built. And that gives you a structurally different bargaining position with lenders and equity markets. Because when you walk into a financing conversation with contracted backlog, asset-backed-financeable customer credit, and capacity you have more influence on the terms. It comes with risk also you will see that later. The second thing that stood out for me and you all know I’ve been working on memory for a while, and a lot of people have been stressing how much memory prices will impact cost pressure for downstream for the neoclouds. Nebius secured a lot of their 2026 components in 2025 already. Andrey pointed out on the call that because of securing early, component inflation was only at a low-single-digits level as a percentage of total spend. Low single digits, while most of the industry is staring at mid-to-high single digit cost pressure for the same buildout year. So good execution on that part. Now what does this mean for 2027? Honestly, this was not addressed directly on the call. They talked about 2027 sites, 2027 power contracts, 2027 customer commitments but they did not disclose anything about 2027 component pricing. NVIDIA gave them line of sight to 5 GW of supply allocation through 2030, but that’s a volume commitment, not a price commitment. So we don’t know at what price they’re getting that supply. The risk skew is materially worse for 2027 than 2026, and you have to understand why. As you know HBM4 is a key component in the architecture of Rubin and by the time Nebius is buying Rubin systems for 2027, HBM pricing will be locked at the new contracted-tier levels we saw confirmed on the SK Hynix call, That’s a structural step-up in BOM. On top of that you have 800V DC architecture transition, co-packaged optics ramp, high-voltage power gear, switchgear, transformers with 18-24 month lead times all in constrained phases. These vendors with pricing power are raising prices. Read further 👇