TeraWulf: Real Power Advantages, but a Changed Financial Bet
October 22, 2026 PWT :)>
Not financial advice. This article is for information and discussion, not individualized investment guidance or a recommendation to buy, sell, or hold any security. Investing can result in loss.
TeraWulf deserves its own investment case. Its ability to secure power, redevelop industrial sites, and lease infrastructure can distinguish it from companies primarily buying GPUs and reselling computing capacity.
But two parts of the familiar bullish story need updating: the records reviewed do not establish ownership of a hydroelectric station, and the company’s 2024 debt-free milestone no longer describes its finances.
The central investment question is whether WULF’s infrastructure advantages will produce enough cash to reward shareholders after construction spending, financing costs, and dilution.
Unless otherwise indicated, financial balances below reflect June 30, 2026 disclosures. They should not be mistaken for a live October balance sheet. Construction timelines and future capacity figures are company expectations, not completed results.
The debt payoff was real. So was the subsequent borrowing.
On July 9, 2024, TeraWulf announced a final payment of $77.5 million, including fees and interest, that eliminated its outstanding debt. That was a genuine achievement.
However, the company subsequently took on substantial financing to support its expansion into AI infrastructure.
TeraWulf also repaid $100 million of bridge-loan principal in April 2026. Nevertheless, its June filing reported these principal balances:
- $3.200 billion of Wulf Compute secured notes due 2030.
- $500 million of convertible notes due 2030.
- $1.000 billion of convertible notes due 2031.
- $1.025 billion of convertible notes due 2032.
Together, those instruments represent approximately $5.725 billion in principal, excluding leases and other liabilities. Principal differs from accounting carrying values after discounts and issuance costs.
Repaying a bridge loan can reduce near-term financing risk while leaving a much larger debt burden outstanding. Both facts can be true simultaneously.
Sources: July 2024 debt payoff; June 2026 Form 10-Q.
Share buybacks and debt repayment are different
The share repurchases were real, too. In February 2025, management reported more than $150 million of buybacks covering more than 24 million shares.
But buying back stock and repaying loans are different uses of cash. Neither permanently prevents future borrowing or share issuance.
In April 2026, TeraWulf priced an equity offering at $19 per share, initially targeting approximately $900 million before the underwriter option. Its stated uses included Kentucky construction and repayment of the bridge facility.
That strengthened funding, but it also issued additional shares.
The right question for existing shareholders is whether the return earned on newly raised capital will more than compensate for the additional ownership claims.
Sources: Full-year 2024 results; April 2026 offering announcement.
What is WULF’s actual power advantage?
The reviewed records do not establish that TeraWulf owns a hydroelectric dam.
Lake Mariner occupies a former coal plant site in western New York. An earlier SEC filing describes electricity purchased from the regional grid, including a 90 MW agreement with the New York Power Authority.
That supports a valuable power-access thesis. It does not establish ownership of the hydroelectric generation supplying the grid.
The distinction matters because ownership of generation, purchased electricity, transmission access, and utility service agreements create different economic rights and risks.
There is a separate generation acquisition involving Morgantown in Maryland. February’s announcement described approximately 210 MW of operating generation and redevelopment potential, while also stating that closing conditions remained. The August update still described the acquisition as pending following FERC authorization.
A subsequent closing has not been verified for this article, so the acquisition is treated as pending.
Sources: Lake Mariner power-sourcing disclosure; Maryland acquisition announcement; August operating update.
Power access can still be a meaningful competitive advantage
Existing grid connections, industrial infrastructure, and power agreements may shorten the route to delivering a usable campus. They can also make a site more attractive to prospective tenants.
Those advantages deserve recognition. However, each campus needs its own economic assessment.
A favorable New York power arrangement does not automatically determine Kentucky economics. TeraWulf’s announcement of the Justified power approval says the company bears its attributable market, transmission, delivery, infrastructure, and credit-support costs.
The resulting shareholder return depends on how those obligations interact with tenant leases, construction costs, and financing.
The useful question is therefore broader than whether the company has access to electricity:
How much does each site cost to build and operate, and how much cash remains after all obligations are paid?
Source: Justified power approval.
Why the bullish case still has substance
TeraWulf’s transition is already producing revenue.
The company reported approximately $44.8 million of second-quarter revenue, including $31.9 million from HPC leases—about 71% of the total.
Its August update reported 102 MW of revenue-generating critical IT capacity after an early-July delivery, with another 336 MW under construction.
That is evidence of an operating transition, although the scale of future development remains much larger than the completed platform.
Source: Second-quarter 2026 results.
The Anthropic lease is another meaningful distinction. The Justified agreement covers approximately 401 MW of critical IT capacity and represents about $19 billion of contracted revenue over an initial 20 years.
Initial service is expected in the second half of 2027, with full capacity targeted for early 2028.
That gives construction a clearer commercial destination than a campus developed entirely on speculation. However, contracted revenue arrives over time and depends on delivery and contractual performance.
For scale, $19 billion divided by 20 years is $950 million annually on average. That simple calculation is not the actual annual rent schedule, a profit forecast, or cash available today.
Source: Anthropic lease announcement.
Muskie provides longer-term potential. The October 5 announcement expanded contracted utility capacity to 1 GW. The first 500 MW is targeted to begin ramping in 2028; the second is planned for 2029, subject to regulatory approval and utility construction.
Those are future milestones. Utility megawatts, critical IT megawatts, completed buildings, signed tenant commitments, and revenue-producing capacity are different measures. They should not be added together as though they represent the same thing.
Source: Muskie power expansion.
Where the bearish case becomes stronger
The strongest concern is the time between spending construction money and collecting sufficient rent.
At June 30, TeraWulf reported approximately $2.619 billion of cash and equivalents, plus roughly $409 million of restricted cash. First-half operating cash outflow was $154.3 million, alongside $1.379 billion of equipment purchases and deposits.
The liquidity provides a cushion. It is also capital being deployed into an expensive buildout.
Some cash is restricted. The same dollars cannot simultaneously repay debt, finish construction, and remain available as a reserve.
Source: Second-quarter financial statements.
Debt service is meaningful as well. At the June principal balance, the $3.2 billion secured notes at 7.75% imply approximately $248 million of annual contractual coupon interest.
That is an illustrative calculation, not reported total interest expense. It excludes other instruments, fees, and accounting adjustments.
Existing fixed-rate coupons do not automatically rise when market yields increase. Future financing and refinancing can still become more expensive, while investors may lower the valuation they assign to distant cash flows.
Source: June 2026 Form 10-Q.
Credit protection helps, but its scope matters
Google’s backstop covers specified Fluidstack lease obligations. It does not protect every WULF project or the share price.
The support also came with warrants for 73.58 million shares at a penny each. That credit protection has value, but it carries an equity cost.
TeraWulf has also provided completion guarantees for the relevant Lake Mariner buildout. Project financing should therefore not be assumed completely detached from the parent company.
For shareholders, the important distinction is between protection that supports a particular contractual obligation and protection against every source of loss.
Sources: Google agreements and warrants; Completion guarantees.
The headline loss needs context
Second-quarter net loss attributable to TeraWulf was approximately $939.9 million, including a $755.7 million warrant fair-value charge.
That charge does not represent an equivalent amount of cash leaving the company during the quarter. Equally, removing noncash charges does not establish positive free cash flow.
A useful review needs to distinguish accounting losses, operating cash generation, construction spending, and financing proceeds. Each tells a different part of the story.
Source: Second-quarter financial statements.
Why sector fear can still hit WULF
Reuters reported an October selloff across AI infrastructure companies driven by concerns about debt-funded expansion. AP also documented the broad AI stock decline on October 22.
Those reports support a sector explanation for selling pressure. They do not prove a single cause for every WULF price move.
The connection is financing. Even a company with stronger power access can lose market value when lenders demand higher returns, prospective tenants face pressure, or investors become less willing to fund construction.
Comparisons should focus on business models. Owning and leasing powered infrastructure can involve a different mix of risks from owning rapidly aging GPUs.
The relevant questions are who funds the equipment, who pays for power, when rent starts, and who absorbs overruns and defaults.
Sources: Reuters on AI financing fears; AP market coverage.
What would strengthen—or weaken—the investment case?
The most useful evidence will come from execution and cash flow:
- Delivery: Are completed facilities beginning rent on schedule?
- Funding: What construction spending remains, and how is it financed?
- Liquidity: How much unrestricted cash is available after commitments?
- Debt: Are obligations declining, or is expansion requiring additional borrowing?
- Dilution: Is cash flow improving relative to the fully diluted share count?
- Power: Are utility delivery dates, approvals, and costs tracking expectations?
- Customer protection: What exactly do guarantees cover, and when do they become effective?
- Transactions: Has Morgantown closed? Has the Abernathy sale closed and delivered cash?
An agreed asset sale should not be treated as money already received. A contracted power allocation should not be treated as an operating, rent-producing data center.
TeraWulf lists its next earnings call for November 6, 2026, at 8 a.m. Eastern. That is the next scheduled opportunity to replace older financial balances and construction guidance with fresher evidence.
Source: TeraWulf investor calendar.
A more cautious thesis
WULF’s infrastructure case remains credible. Power access, industrial redevelopment, and long-term tenant commitments can create substantial value.
But a bullish thesis based on being debt-free and owning hydroelectric generation needs revision.
The company must turn its advantages into completed facilities, collect rent, service debt, and produce returns that justify the capital raised along the way. Growth in company size does not automatically mean growth in value per share.
Becoming more cautious is reasonable when financial commitments are large and important revenue milestones remain ahead. The falling share price alone does not establish that the business thesis has failed. A lower price also does not establish that the stock is undervalued.
That judgment requires a closer look at project cash flows, remaining construction costs, financing terms, and dilution.
AI disclosure: This article was researched and drafted using an OpenAI AI agent to review public filings, company announcements, and news reporting. AI can make mistakes or miss developments. Readers should verify material claims against original sources, particularly transaction status, financing terms, and updated financial results. This is not an independently audited investment report.
Not financial advice. This article provides general research and commentary and does not account for any reader’s financial circumstances, investment objectives, or risk tolerance.