My bullish case for TeraWulf starts with what an AI customer needs before its first server can operate: a suitable site, dependable electricity and a facility that can be delivered on time.

The commercial value lies in bringing those requirements together. A large power allocation has limited use to a customer if the building is unfinished. A completed building has limited use if the electrical connection cannot support the equipment. The developer earns its return by delivering a working facility under terms that justify the capital invested.

TeraWulf is building that business through long-term data center leases. Lake Mariner is producing revenue, while the Anthropic agreement in Kentucky adds a substantial development commitment. The opportunity is becoming specific enough to assess through contracts, construction milestones and financing decisions.

I am bullish because successful delivery could leave TeraWulf with a much larger base of income-producing infrastructure and stronger prospects for winning subsequent projects. The investment case depends on how much value those assets create for each share. That is the standard I apply throughout this article.

Power access has commercial value

Electricity demand provides the broader context. In its 2025 Energy and AI report, the International Energy Agency projected that global data center electricity consumption would more than double to roughly 945 TWh by 2030 in its base case. It also identified grid constraints as a potential cause of project delays. These are scenario estimates, not guaranteed outcomes.

For an infrastructure investor, the local details matter more than the global total. Power must be available at the right site, in the required quantity and within the customer's timetable. Transmission access, utility agreements and construction readiness can therefore influence which developer wins a contract.

I see TeraWulf's opportunity in assembling sites where those requirements can be met economically. Reusing industrial infrastructure may reduce some development work, but the advantage has to survive engineering studies, permits and the actual cost of serving a new load.

The most convincing evidence is a customer willing to commit for years. That converts an attractive location into something investors can evaluate against a construction budget and an expected stream of rent.

The business model concentrates on infrastructure

WULF Compute develops facilities for demanding AI and high-performance computing workloads. TeraWulf describes the offering as high-density infrastructure with resilient electricity systems and diverse network connectivity.

The distinction between providing the facility and operating a cloud service matters. A landlord's economics depend on contracted rent, delivery costs, operating obligations and financing. A cloud operator must also manage the utilization and commercial pricing of its computing equipment. Investors should evaluate TeraWulf primarily through the obligations it assumes under its leases.

This model appeals to me because the customer can commit to a facility for a much longer period than a single generation of computing hardware. A suitable campus may accommodate successive equipment upgrades. That possibility gives the underlying location and electrical infrastructure enduring utility.

Buildings still require reinvestment. Changes in rack density, cooling requirements or redundancy standards can make upgrades necessary. The quality of the business depends partly on who pays for those changes and whether the lease compensates TeraWulf adequately for the services it provides.

Lake Mariner makes the thesis measurable

The August update reported 102 MW of revenue-generating critical IT capacity at Lake Mariner following CB-3's delivery in early July. Another 336 MW was under construction. CB-4's phased delivery was expected in the second half of 2026, with CB-5 beginning in early 2027.

That schedule is central to the near-term investment case. A completed phase can begin contributing rent while later phases remain under construction. Investors can follow whether additional capacity reaches service and whether the resulting income matches the cost and timing assumptions used to finance it.

Critical IT capacity means capacity available to computing equipment. Gross electrical capacity also supports cooling and other facility systems. Mixing the two can overstate what a campus can deliver to tenants. Comparisons between projects need the same measurement basis.

There is already a financial record to examine. HPC lease revenue was $9.7 million in the fourth quarter of 2025. In the second quarter of 2026, it reached $31.9 million, approximately 71% of total revenue of $44.8 million.

The direction is encouraging. My focus now is on sustained collections, service performance and the economics of each new phase. A construction completion announcement is valuable because it brings the business closer to those outcomes; the operating record must then confirm them.

Google improves the contract credit profile

Google's role is defined in TeraWulf's filings. For specified Fluidstack leases, the support becomes effective at lease commencement. Following a qualifying payment default or insolvency, Google can pay the termination fee or pay overdue rent and assume the lease. TeraWulf issued Google warrants covering 73.58 million shares at a $0.01 exercise price in exchange for that support.

I consider this relationship meaningful because lenders assess who ultimately supports contractual payments. Stronger protection can help make a large construction project financeable and improve the confidence investors place in the associated income.

The agreement should be understood on its own terms. It does not guarantee every TeraWulf project or eliminate construction risk. The warrant consideration also carries a real economic cost for existing shareholders, even though it differs from paying a fee in cash.

The relevant judgment is whether the financing access and credit protection create more value than the ownership surrendered. I see a credible argument that they can, especially when a project would otherwise require more expensive capital. Actual borrowing terms and subsequent operating performance determine the result.

Anthropic creates a substantial second source of growth

In July, TeraWulf announced a 20-year lease with Anthropic for approximately 401 MW of critical IT capacity at Justified Data in Hawesville, Kentucky. The initial term represents approximately $19 billion of contracted revenue. Initial service is expected in the second half of 2027, with full delivery in early 2028.

The significance extends beyond the headline amount. A direct agreement with a major AI developer gives TeraWulf another customer relationship around which to plan a large campus. It also provides a practical test of whether the company can reproduce its development capabilities beyond Lake Mariner.

The contract total needs careful interpretation. Dividing $19 billion by 20 produces roughly $950 million per year as a simple arithmetic average. That is not guidance for a particular year, annual free cash flow or the present value of the lease. Delivery timing, contractual escalators and expenses affect the financial outcome.

For my thesis, the attraction is the duration of the commitment relative to the development effort. If the project is financed sensibly and delivered successfully, it could contribute to TeraWulf's business for many years. The scale also makes construction discipline especially consequential.

Kentucky shows how the power advantage must be earned

TeraWulf's August 24 announcement reported regulatory approval for up to 482 MW of electric service at Justified. The former aluminum site retains substantial transmission infrastructure. Under the approved arrangement, TeraWulf bears the costs attributable to its load, customer-specific infrastructure costs and significant credit-support obligations.

The approval is useful evidence because it advances a particular project. It also explains why access to power should never be confused with inexpensive, unrestricted electricity. The commercial agreement has to compensate the developer for the obligations attached to that supply.

The same release estimated approximately $4.0 billion to $4.5 billion for site development and the initial data halls, excluding customers' computing equipment and related infrastructure. That scale of spending makes financing terms, contingency reserves and the sequencing of construction essential to the equity case.

I view the reuse of existing electrical infrastructure as an advantage when it improves a project's delivery schedule or total cost. Its value must be established project by project. A regulated structure that assigns costs clearly may also make future development more predictable for utilities, local communities and investors.

Selling Abernathy tests capital allocation discipline

TeraWulf agreed to sell its entire 50.1% interest in the Abernathy joint venture to a Fluidstack-led group. The July announcement identified approximately $450 million of invested capital; the August update disclosed approximately $530 million of aggregate cash consideration. These disclosures describe an agreed sale rather than establish that closing has occurred.

This decision interests me because adding capacity is only one way to grow shareholder value. Selling a project can be sensible if the proceeds fund opportunities with better expected returns or more favorable ownership terms.

Greater control also brings greater responsibility. A wholly owned development leaves TeraWulf with more of the potential income, but it may also require more capital and leave the company bearing more risk. The benefit cannot be assessed from ownership percentages alone.

I would judge the transaction through its final proceeds, the obligations released and the returns on the replacement investment. A disciplined developer should be willing to sell when doing so improves the economics of the remaining business.

Additional sites provide future opportunities

The Muskie acquisition adds another Kentucky development opportunity. The May announcement described utility infrastructure involving a planned 345 kV substation connected to an existing 765 kV transmission network, with power and infrastructure agreements executed alongside the acquisition. The August update put initial electric service in the fourth quarter of 2028.

The commercial attraction is having another location to offer customers whose deployment schedules extend beyond the current construction program. Different sites can support different customer requirements and reduce dependence on securing every future expansion at one campus.

I assign less certainty to this opportunity than to capacity already collecting rent. Electric-service arrangements, tenant commitments and completed data halls represent different stages of development. An unleased future phase should carry a different valuation from an operating facility.

This distinction improves the bullish case by keeping expectations realistic. TeraWulf can create additional value if it signs attractive leases and funds them on acceptable terms. There is no need to assume that every potential megawatt already deserves the valuation of a finished asset.

Financing determines the return for shareholders

At June 30, cash and restricted cash totaled approximately $3.03 billion, including about $409 million of restricted cash. Reported debt and convertible-note carrying amounts totaled roughly $5.21 billion. First-half operating cash outflow was $154.3 million, with $1.38 billion spent on plant and equipment purchases and deposits.

The capital base supports construction, but it must be assessed alongside the remaining spending commitments and restrictions on its use. Cash raised for development cannot simply be treated as surplus value available for distribution.

Equity issuance is already part of the funding model. In April, TeraWulf closed an offering of 54.51 million shares at $19 each, with proceeds intended partly for the Kentucky campus and bridge-facility repayment. New capital can create value when the assets it finances earn sufficient returns. The relevant outcome is future cash generation per diluted share.

Reported earnings also require interpretation. The second-quarter loss attributable to TeraWulf was approximately $939.9 million and included a $755.7 million noncash warrant remeasurement charge. That accounting movement differs from operating cash spending, while the associated warrants remain economically relevant.

I would therefore assess construction spending and cash obligations directly, alongside the rent being collected. Neither a large accounting loss nor an adjusted earnings measure, considered alone, provides a satisfactory view of what shareholders own.

The evidence that would strengthen the investment case

The next material evidence should come from the conversion of contracted projects into functioning facilities. At Lake Mariner, I want to see rent commence as additional phases are delivered, followed by reliable service and collections. In Kentucky, progress should become visible through financing commitments, construction milestones and a credible path to tenant acceptance.

I would also watch changes in cost estimates. A project can achieve its planned capacity and still produce a disappointing return if construction costs or financing expenses rise too far. Conversely, a repeatable delivery process could improve the economics of later phases through better procurement and fewer avoidable mistakes.

A convincing valuation would estimate the cash each contracted site can distribute after operating costs, debt service and necessary reinvestment. It would discount those amounts for time and risk, then account for corporate obligations, remaining equity funding and dilution. Future uncontracted developments deserve a separate, probability-weighted assessment.

That approach leaves room for substantial upside without assuming it. Successful commissioning can make future cash flows more credible. New contracts can add value when their expected returns exceed their capital costs. A higher share price is justified only to the extent that those improvements outpace what the market already expects.

What could weaken the thesis

Delivery delays are the clearest immediate risk. They can postpone rent while the company continues paying for construction and financing. Equipment lead times, commissioning problems or changes requested by tenants can affect both cost and schedule.

Customer concentration is another consideration. A long lease provides visibility, but the identity and creditworthiness of the counterparty still matter. The remedy available after a default can differ materially from receiving the originally expected rent on time.

There is also a longer-term risk that demand or facility requirements change. More efficient computing, new competing campuses or revised customer investment plans could weaken future leasing economics. Existing contracts provide a measure of protection; they do not guarantee attractive returns on every expansion.

For me, the most damaging combination would be repeated delays, rising capital requirements and dilution that outpaces the increase in asset value. A bullish view has to remain responsive to that evidence.

Why I remain bullish on TeraWulf

TeraWulf offers a concrete way to invest in the facilities needed for AI computing. Its appeal comes from identifiable sites, customer commitments and development work that can be measured against an operating result.

I am particularly interested in what a successful delivery record could do for subsequent business. A customer choosing a development partner has to assess whether that partner can coordinate power, construction and service obligations. Demonstrated performance can make the next contract easier to win and finance, although favorable terms must still be negotiated.

The opportunity is to build a larger portfolio of useful, income-producing infrastructure while retaining enough of the economics for common shareholders. That outcome would give the long-term investment case substance well beyond an increase in announced capacity.

I remain bullish on $WULF because the company has meaningful commercial commitments against which to execute. My conviction will be reinforced by delivered facilities, collected rent and improving cash generation per share. Those are the results that would make TeraWulf a more valuable business to own.