TeraWulf (WULF): what it does, how it makes money, and why a bitcoin miner became an AI landlord

The standard $WULF story is "bitcoin miner", and on that story the stock is a bet on the price of bitcoin. That framing is out of date. What TeraWulf actually owns that the AI buildout wants is not its mining fleet. It is its power: connected, permitted sites that take years to replicate.

The more accurate frame: TeraWulf is converting powered mining sites into data center capacity leased to AI tenants on long contracts, and using mining cash flow to fund the conversion. The stock trades on how fast and how cleanly that conversion happens. This piece walks through what TeraWulf owns, how it makes money, where it sits in the AI infrastructure stack, and the execution risk underneath. Figures are as of 2026-09-18 unless noted.

Why it matters now

The stock is up about 52% over a year but down about 40% over the three months to 2026-09-18. There is no trailing P/E to anchor on (the company is not reporting positive trailing earnings), and it trades above 50x trailing sales, so the price is almost entirely a bet on future lease income. That makes the evergreen question sharper than usual: what does TeraWulf actually control, and how much of the AI story is signed?

What does TeraWulf do?

TeraWulf started as a bitcoin miner, and it still mines. But its growth story is power-backed AI compute capacity: taking sites that already have a grid interconnection, permits and cheap zero-carbon electricity, and building data center capacity on them for AI tenants on long-term leases.

Per QA's desk thesis, TeraWulf holds about 770 MW of contracted or permitted capacity across two New York sites:

• Lake Mariner, powered by hydro and grid energy, where Core42 (a subsidiary of Abu Dhabi's G42) anchors 200+ MW.
• Cayuga, with a nuclear power purchase agreement, which is rare among miners and fits hyperscaler net-zero targets.

A Fluidstack lease, backstopped by $GOOGL, adds roughly 200 MW more.

How TeraWulf makes money

TeraWulf has two engines, and the investment case is the hand-off between them.

• Bitcoin mining still generates cash flow. That cash flow helps fund the AI conversion without the heavy share dilution that pure-play neoclouds lean on.
• AI data center leases are the growth engine: fixed-tenor lease income from AI tenants, the kind of predictable revenue that data center landlords are valued on.

The desk's bull thesis puts the prize in one line: a re-rating from a miner's multiple (the desk uses about 8x) toward a data center landlord's multiple (about 20x for names like Digital Realty and Equinix). Treat that as a thesis framing, not a forecast.

The headline risk sits right under the leases: concentration. Core42 and Fluidstack together represent the majority of TeraWulf's AI capacity. A handful of tenants carry the growth story. That is the number to hold onto through the rest of this piece.

Where it sits in the AI infrastructure bubble

TeraWulf sits in QA's Cooling / DC Infra bubble, the physical plant of AI, alongside equipment names like Vertiv. It maps onto the Compute Capacity and Neoclouds themes.

The useful way to place TeraWulf is by what it provides. It is not a GPU cloud selling compute by the hour, like CoreWeave. It is closer to a landlord: it provides the powered, cooled building and leases it. For how the GPU-cloud side of this trade compares on risk, see neoclouds ranked by risk-adjusted value.

Its peers are the rest of the miner-to-AI pivot. WULF's tightest correlations in the QA universe (252 trading days, market beta stripped out) run to Cipher Mining ($CIFR, 0.78), Core Scientific ($CORZ, 0.77) and Hut 8 ($HUT, 0.74), with IREN ($IREN) and Applied Digital ($APLD) close behind. They trade as one group, so holding several of them is less diversified than it looks. The full peer table is on /stocks/wulf.

The numbers

The Street mean target is an estimate, an average of analyst opinions, not a promise or a QA view. With no trailing earnings and a sales multiple above 50, the stock is priced on lease income that is still being built.

The bull case

• Power is the bottleneck, and TeraWulf has it: grid interconnection and permits take roughly 4 to 7 years to replicate.
• Long leases (Core42, Fluidstack with a Google backstop) convert volatile bitcoin cash flow into fixed-tenor lease income.
• The Cayuga nuclear PPA is rare among miners and matches hyperscaler net-zero targets.
• Mining still produces cash flow that helps fund construction.

The bear case

• Execution risk. Building liquid-cooled GPU capacity is new work for TeraWulf.
• Tenant concentration. Core42 and Fluidstack are the majority of the AI capacity.
• Bitcoin cash flow halves with each halving cycle, narrowing the funding bridge.
• AI-ready buildout costs roughly 3 to 5 times more per MW than a mining shell.
• Competition from the same pivot: IREN, Cipher, Core Scientific and Applied Digital are all heading the same way, so scarce power under contract is the main moat.

How to access

TeraWulf trades on the Nasdaq as WULF, a clean US listing. To trade it from a US-retail account alongside the rest of the AI-infrastructure names, see /stack/ibkr. The live ETF holdings breakdown is on /stocks/wulf.

Bubble-correlation shifts and rule-based alerts on $WULF are part of /pro.

What to watch

• Delivery of the leased capacity at Lake Mariner on schedule, the proof that the conversion works.
• Any new tenant beyond Core42 and Fluidstack, which would widen the base.
• Capex per MW against plan, where the execution risk shows first.
• Mining cash flow against the funding needs of the buildout.
• The desk has no curated level on this name yet; for the chart, use the live levels on /stocks/wulf.

Live data on this ticker: /stocks/wulf. Price, ETF holdings, bubble correlation, curated levels, bot positions.

Bubble context: /bubbles/cooling-dc-infra. The cluster this name belongs to and how it's moving.

QuantAbundance is educational research. Nothing here is investment advice. See /disclosures.