Rosenblatt reiterates Terawulf stock Buy rating on HPC momentum By Investing.com

Terawulf reported Q4 fiscal 2025 revenue of $35.8M, down from $50.6M sequentially, with digital asset revenue falling to $26.1M from $43.4M after optimizing miner spacing and reallocating power to HPC. The company secured a $500M bridge loan with Morgan Stanley to fund its Hawesville, KY data center, and multiple analysts updated targets (Rosenblatt $23 PT based on 22.7x 2027 Adj. EBITDA, Cantor $24, Compass Point $28, KBW $23), with Rosenblatt calling WULF a top pick into H1 2026; the stock is up ~33% YTD and ~364% over the past year.

Analysis

The market is repricing optionality around small-cap hosts that attempt to pivot from volatile, spot-exposed revenue streams toward multi-year HPC contracts. If the pivot sticks, these companies convert lumpy, utilization-driven cash flows into contractually-backed RFP cycles that favor upstream hardware vendors (GPUs, PDUs, racks) and financing partners; conversely, legacy spot miners and third-party colocators who can’t secure term customers face accelerating asset stranding. Expect a 3–12 month visibility window where order books, PPA terms, and customer LOIs will drive headline momentum more than quarter-to-quarter mining volumes.

A critical second-order constraint is physical power and electrical balance of plant capacity: transformers, switchgear, and utility interconnects have 6–18 month lead times and concentrated suppliers, giving hosts pricing and timing risk if they under-contract infrastructure. Financing cadence is another hinge — bridge-like facilities that bridge construction to contracted cash flows lower execution risk if covenant headroom exists, but they amplify downside if HPC bookings lag. On the tech side, sustained GPU allocation is a gating factor; upstream supply tightness can bottleneck data center monetization even when floor space is available.

The consensus leans bullish on the thematic transition; the contrarian path is an execution cliff if contracts slip or power buildouts run over budget, which would compress implied multiples quickly for high optionality names. For investors, the cleanest way to express the theme is exposure to upstream, durable beneficiaries of AI demand while using compact, defined-risk instruments to play small-cap hosts — defendable via option overlays or pairs to neutralize beta and power-price exposure over the next 6–18 months.

Key Decisions for Investors

• Long NVDA (buy shares or 3–6 month calls) — tactical overweight for upstream GPU exposure. Target +20–30% over 3–6 months if AI spend stays stable; set a protective stop at -12% to guard against macro drawdowns.
• Defined-risk long on WULF (buy 12–18 month call spread, financed by selling 3–6 month calls) — small position (1–2% portfolio) to capture optionality of successful HPC contracting. Reward skew: 2–4x on upside if term bookings materialize; max loss = net premium paid if execution stalls.
• Pair trade: Long SMCI / Short WULF (equal notional) over 6–12 months — expresses preference for hardware vendors capturing recurring demand over a single-site host that faces execution/financing risk. Aim for 8–20% relative return; close if spread narrows by 50% or if SMCI order cadence disappoints.
• Tactical long MS (small overweight, 3–9 months) — bank fees and lending benefit from financing activity around data centers. Expect modest upside (5–15%) with lower volatility versus single-name small-cap hosts; hedge with a 3–6 month index put if credit markets reprice.