should I buy IREN, NBIS or ALAB at today's price

For your long‑term, high‑risk sleeve, I’d treat **IREN and ALAB as buy‑candidates on weakness and keep NBIS as the smallest / most speculative**. If you want just one new name today, I’d lean **IREN or ALAB (not NBIS)**, depending on whether you prefer data‑center infrastructure (IREN) or AI semis (ALAB). Below is the reasoning, then a concrete way to size and time entries. --- ## 1. Big picture: what each name gives you | Ticker | Role in AI theme | Core risk profile | |--------|------------------|-------------------| | **IREN** | Power + AI compute (BTC + NVIDIA GPU data centers) | High beta, heavy capex, now higher dilution risk but real hard assets | | **NBIS** | Pure AI cloud (CoreWeave‑style, MSFT/Meta contracts) | Very high beta, heavy dilution and execution risk, flows currently weak | | **ALAB** | AI connectivity semi (PCIe/CXL retimers & switches) | High beta “picks & shovels”, rich valuation but best business quality | You’re already loaded with mega‑cap AI (NVDA/MSFT/META/GOOGL), so these are **satellites**, not core. In our earlier chats we treated IREN as the anchor and NBIS as a smaller, leveraged bet inside a ~4–5% “AI infra” sleeve. The new variable is ALAB. --- ## 2. IREN – still your best infra core, but the $6B ATM changes the risk ### Thesis refresher - Has pivoted from pure Bitcoin mining to an **AI data‑center + BTC “dual engine”**: low‑cost renewable power, grid interconnects, and NVIDIA H100/H200/B‑series clusters. - Moat is **power + permits**: gigawatt‑scale interconnection rights and 100% renewable power, which hyperscalers care about. ### New developments (last few days) - **Massive GPU buy:** IREN agreed to purchase **50,000+ NVIDIA B300 GPUs**, lifting its fleet to around **150,000 GPUs**. - **ATM expansion:** at the same time, it **expanded its at‑the‑market equity program from $1B to $6B** and disclosed it has already sold ~66M shares for about $1B. - Management is now talking about **$3.4–3.7B AI cloud ARR by late 2026** if the build‑out works. **Implication:** the upside is bigger (multi‑billion AI revenue), but the **dilution and execution risk just went up materially**. This is the key change versus our prior IREN vs NBIS discussions. ### Quant & flows From the quant tools for IREN: - **Fundamentals:** overall rating **Neutral** – some positives (revenue vs market cap, cash vs market cap, net‑income vs revenue look OK) but also signs that profits and ROA are under pressure. - **Technical profile:** overall **neutral**, with recent signals like a MACD golden cross and long upper shadow – historically mildly bullish but not a screaming signal. - **Analysts:** 3 recent ratings (1 Strong Buy, 2 Buy); market expectations **“optimistic”** and roughly aligned with the recent mild uptrend. - **Flows:** fund‑flow score **neutral**, but extra‑large and large orders have historically been associated with modest positive forward returns in similar setups. ### Price action Last ~3 months (normalized to 100 in early December), IREN ran hard into January and has since round‑tripped much of that gain, now sitting slightly below that prior peak, roughly in line with NBIS and above ALAB in relative performance. ```reference [{"id":19,"type":"plotly-json"}] ``` Option‑wall data points to **support in the low‑40s and resistance up around 50** over the next 1–6 months; that frames your near‑term trading range. ### Takeaway on IREN - **Pros:** Real power assets, clear AI pivot, huge GPU order, and still your best “balanced” infra play versus NBIS. - **Cons (new):** The **$6B ATM** means the stock is now a **levered capital‑raise machine**; if they overshoot, shareholder value can be eaten by dilution. **How to treat it today (for you):** - I’d still make **IREN the primary infra holding over NBIS**, but: - **Keep it as a satellite** (e.g., up to around 60–70% of your AI‑infra sleeve, not your whole tech bet). - **Buy in tranches on big BTC/AI risk‑off days**, not all at once. Today’s dilution‑driven selloff is an example of the type of move where it’s reasonable to start or add a piece – as long as you accept that they *might* raise a lot more equity. --- ## 3. NBIS – huge story, but weakest near‑term setup ### Thesis refresher - Essentially a **public‑market CoreWeave analogue**: pure‑play **GPU cloud IaaS + PaaS** with ex‑Yandex engineering and multi‑billion MSFT/Meta contracts in its story. - Big vision around **European sovereign AI cloud**, with guidance toward multi‑billion revenue and strong long‑term EBITDA margins. ### Recent developments - **Growth:** Q4 revenue up **~547% YoY**, but missed consensus by ~8%; Nebius is now **EBITDA‑positive** at the group level, not just in the AI cloud segment. - **Capex & capacity:** targeting **$3–3.4B 2026 revenue and $7–9B ARR**, with heavy capex needs. - **Missouri “AI factory”:** got approval for a **1.2 GW AI factory campus in Missouri**, stock popped ~12.6% on the news. - **Ownership moves:** at least one institution (Virtus Wealth Solutions) cut its NBIS stake by ~85% last quarter, even as Street ratings remain broadly positive with triple‑digit price targets. ### Quant & flows - **Fundamentals:** overall rating **Outperform**, mostly because leverage/capital structure metrics look favorable; profit metrics are still weak. - **Technical:** evaluation **neutral** with a recent MACD death cross that, in its historical sample, oddly backtests as bullish – the system is picking up on high‑volatility mean‑reversion, not a clean uptrend. - **Analysts:** 2 recent **Strong Buys**, but the price has fallen ~9% over the same window – a classic setup where **analyst optimism and tape disagree**. - **Flows:** fund‑flow score **poor** – extra‑large and large orders’ backtests are negative in similar ranges, suggesting **institutions are not meaningfully accumulating dips here right now.** ### Price action On the 3‑month chart above, NBIS has been **choppy and slightly negative**, with big intraday swings but no clear relative strength edge over IREN. Options data shows a **tight support/resistance band around 90–96–99** over the next month; you’re not being paid much for the volatility at today’s levels. ### Takeaway on NBIS - **Pros:** The purest AI‑cloud upside, huge contracts, and a scarce narrative. - **Cons:** Heavy capex and expected dilution, recent earnings miss, big holder trims, and **weak near‑term flows**. **How to treat it today:** - I would *not* make NBIS the main buy at current prices. - If you want exposure: - Treat it as a **small, leveraged side bet (≤1% of portfolio)**. - Prefer adding **on deeper pullbacks** (closer to or below the lower end of the 90 support area) or after a clean post‑earnings reset, not on news‑driven spikes like the Missouri approval. Relative to IREN and ALAB, **NBIS is the one I’d be most cautious about buying today.** --- ## 4. ALAB – best business quality, but high multiple and volatility ### Thesis refresher - Astera is a **fabless AI connectivity company**: - Market leader in **PCIe/CXL retimers** and CXL switch chips, critical to building large GPU clusters and “breaking the memory wall”. - Moat comes from **interoperability certifications** with NVDA/AMD/INTC and its **COSMOS software** that gives data centers link observability – high switching costs. This is the classic “AI picks & shovels” – it benefits from AI server growth across *all* cloud players, not just one. ### Recent developments - **Q4 2025 earnings:** revenue **$270.6M, +92% YoY**, EPS $0.58 vs $0.51 est – a beat on both. - Despite the beat, the stock **crashed a bit over 30%** on concerns around margins and very high expectations. - Analysts have stayed very bullish: - Loop Capital initiated with **Buy and $250 PT** (roughly 2x from pre‑note levels). - Citi trimmed PT from $275 to $250 but kept **Buy**, still calling it one of the **best AI stocks for explosive growth potential**. - There has been **insider selling** (e.g., the General Counsel sold 10k shares around ~$117), which is worth noting but not thesis‑breaking in isolation. ### Quant & flows - **Fundamentals:** rating **Underperform**, largely because valuation metrics (PB‑ROE, revenue vs market cap, cash vs market cap) are stretched – you are paying up for the story. - **Technical:** evaluation **outperform**, with lots of **long lower shadows** recently – historically associated with positive average forward returns (i.e., dips being bought). - **Analysts:** 1 recent Strong Buy from a reputable semi analyst; market expectation tagged as “neutral,” but that likely underplays how bullish the Street is based on external notes. - **Flows:** fund‑flow score **good** – extra‑large and large orders show **positive backtested outcomes** in similar setups, suggesting institutions are **accumulating the post‑crash dip** rather than dumping. ### Price action In the same 3‑month window, ALAB **ran the hardest and then sold off the hardest**, now sitting well below its highs. That fits the narrative: it was priced for perfection, then reset. Option‑wall data shows **support around 118 and resistance around 140** for the next 1–6 months – your current price zone (roughly mid‑120s to 130s) is **near support, below heavy resistance**, which is not a bad risk/reward starting point if you can tolerate more volatility. ### Takeaway on ALAB - **Pros:** Best business quality of the three, direct leverage to AI server volumes, strong strategic positioning in PCIe/CXL, and evidence that institutions are buying dips. - **Cons:** Very rich multiples, customer concentration risk (notably AWS), and extreme sensitivity to any wobble in AI capex or guidance. **How to treat it today:** - For a long‑term AI‑bull like you, ALAB is a **compelling high‑beta satellite** in the AI semiconductor stack. - I’d be comfortable: - Starting a **small position now** (e.g., 1–2% of portfolio), - Planning to **add on retests of the ~120 area** or on sector‑wide AI semi pullbacks, - Keeping sizing disciplined, as this can move ±30% around earnings. --- ## 5. Concrete answer to “Which should I buy today?” Putting it all together and staying consistent with your prior sleeve structure: 1. **Between IREN vs NBIS (AI infra/data‑center):** - I still favor **IREN over NBIS** as the *core* infra name, even after the ATM expansion, because it has more tangible power assets and a clearer SOTP path. - NBIS stays a **small, speculative add‑on at best**, ideally on deeper dips. 2. **Where does ALAB fit?** - ALAB gives you **AI semi exposure**, not just data centers, and currently has: - Better institutional flow support, - A strong fundamental story, - And a post‑earnings reset that has partially de‑risked the entry. - If you want *one* new AI‑satellite today, **I’d pick between**: - **IREN** – if you want data‑center / BTC‑leveraged infra and accept dilution risk, or - **ALAB** – if you prefer pure “picks & shovels” semi exposure. 3. **Position sizing & risk guardrails (illustrative, not personalized advice):** For your overall portfolio: - Keep the **combined high‑beta AI‑infra/semis sleeve (APLD + IREN + NBIS + ALAB)** in roughly the **5–7%** range. - Within that sleeve, a reasonable structure could be: - **IREN:** up to **3–4%** (anchor infra position, accumulated in tranches on drawdowns) - **ALAB:** **1–3%** (AI semi satellite, add more on sector pullbacks) - **NBIS:** **≤1%** (only if you explicitly want a high‑octane, pure AI cloud bet) This keeps any single blow‑up from endangering the rest of your long‑term AI core (NVDA/MSFT/META/GOOGL, etc.). --- ## 6. How do you want to lean the sleeve? The key choice now is *directional*: - Do you want to **double‑down on AI infrastructure risk** (IREN/NBIS style), or - Do you want to **tilt a bit more toward AI semis/picks & shovels** via ALAB? Tell me which side you’re more interested in, and I can sketch a step‑by‑step entry plan (exact tranche sizes and rough price zones) for the next few weeks.