Due Diligence Capital Notes

CleanSpark just reported a $757 million net loss for the first half of its fiscal year 2026. In the same six months, management spent $460 million buying back its own shares. That is not a typo, and it is not the behavior of a company circling the drain — or at least, that is the argument management is implicitly making with every capital allocation decision in its latest quarterly filing. Whether they are right is a genuinely interesting question.

I spent time with CleanSpark's 10-Q for the quarter ended March 31, 2026, filed May 11, 2026, and what I found is a company in the middle of a high-stakes bet — one where the headline loss figures are almost entirely driven by accounting rules around Bitcoin pricing, yet the underlying operational and financial stress is very real. Let me walk through what is actually happening.

Why the Loss Number is Misleading — and Why It Still Matters

Before we get into strategy and capital allocation, we have to talk about accounting. CleanSpark holds Bitcoin on its balance sheet using fair-value accounting — meaning that at the end of every quarter, the company marks its Bitcoin holdings to the prevailing market price, and any change in that price flows directly through the income statement as a gain or loss. This is distinct from older accounting rules that only allowed companies to record losses (never gains) until Bitcoin was sold. The newer fair-value treatment creates wild swings in reported earnings that have nothing to do with how much cash the business is generating or burning.

During the six months ended March 31, 2026, Bitcoin's price fell sharply — from roughly $114,000 per coin at the start of the period to approximately $68,222 at quarter-end. CleanSpark held 11,920 BTC at March 31, 2026, and that price decline produced a non-cash fair-value loss of $470.9 million — meaning a paper loss that required no cash to leave the building. Strip that out and the operating picture is still pressured, but the catastrophic headline number dissolves considerably.

That said, it would be intellectually lazy to declare "it's just accounting" and move on. The cost basis problem is real: CleanSpark paid an average of $97,309 per coin across its treasury, and with fair value sitting at $68,222 per coin, the company holds 11,920 BTC worth $813.2 million against a cost of roughly $1.16 billion. That is a real economic loss in any meaningful sense, even if no check has been written.

What a Halving Actually Does to a Bitcoin Miner

To understand why CleanSpark's revenue is falling even as it holds more Bitcoin, you need to understand the Bitcoin halving — an event baked into Bitcoin's code that occurs roughly every four years and cuts the block reward (the new Bitcoin paid to miners for validating transactions) in half. The most recent halving occurred in April 2024, reducing per-block rewards from 6.25 BTC to 3.125 BTC. Combined with rising network difficulty — a measure of how much computing power is competing globally to win those block rewards — the economics of mining became materially harder.

The numbers in the 10-Q reflect this directly: Bitcoin mining revenue fell to $317.6 million in the first half of FY2026, down from $344.0 million in the same prior-year period — a 7.7% decline year-over-year, despite CleanSpark growing its Bitcoin holdings from 10,428 BTC to 11,920 BTC. You can be mining more efficiently and still earn less. That is the post-halving trap.

How CleanSpark Is Responding: Three Moves Worth Examining

1. Raising $1.13 Billion in Convertible Debt

During the first half of FY2026, CleanSpark raised $1.133 billion in gross proceeds from two new series of convertible notes — bonds that pay a fixed interest rate but can be exchanged for company shares at a predetermined price if the stock rises enough. Investors accept terms that they might not accept on ordinary bonds because the conversion option gives them exposure to equity upside without the full downside of owning stock outright.

The result is that CleanSpark's long-term debt has ballooned from $644.6 million at the end of September 2025 to $1.788 billion at March 31, 2026. That is a significant liability for a company whose stockholders' equity stands at $986.2 million. The debt matures in 2030 and 2032, which buys time — but also means CleanSpark needs Bitcoin prices and its operating business to recover meaningfully before those obligations come due or need refinancing.

2. Buying Back $460 Million in Stock

Simultaneously, CleanSpark repurchased $460 million worth of its own common shares in the same six-month period, bringing total treasury shares (shares bought back and held) to 42.4 million shares at a cumulative cost of $608.2 million. As of May 7, 2026, 256,608,606 shares remain outstanding.

Why would a company with a $757 million net loss and rapidly rising debt aggressively retire its own shares? The logic — and this is management's implicit argument — is that if Bitcoin's price recovers, fewer shares outstanding means each remaining share captures more of that recovery. It is a leveraged conviction play: we believe the stock is cheap relative to our Bitcoin holdings and future capacity, so we are concentrating ownership for those who stay. The counterargument is obvious: this used $460 million of capital that could have served as a cushion if conditions worsen further.

One number that catches my eye is the accumulated deficit, which is an accounting term for the running total of all net losses a company has ever recorded, net of any prior profits. CleanSpark's accumulated deficit jumped from $125.9 million at the end of fiscal year 2025 to $912.9 million at March 31, 2026 — a $787 million deterioration in six months. Much of that is the fair-value accounting effect, but the pace of that widening is striking nonetheless.

3. Acquiring Texas Land and Hinting at HPC/AI

The operational expansion story involves three Texas land acquisitions completed during the period: a Brazoria County site for $28.7 million, an Austin County site for $65.7 million, and additional locations for $13.3 million. These sit alongside expansion into South Dakota as CleanSpark builds toward larger data center footprint.

More interesting to me is the AI/HPC whisper in the filing. The company's forward-looking disclosures explicitly acknowledge "our ability to diversify and expand into the market for high-performance computing (HPC) and artificial intelligence (AI) solutions and data centers" alongside an admission of "our limited experience with respect to new markets we are entering." That parenthetical is doing a lot of work. HPC and AI workloads require different infrastructure — lower latency, different power density requirements, hyperscaler customer relationships — than Bitcoin mining. CleanSpark is signaling the ambition but the 10-Q offers no contracts, no revenue, and no timeline. It is aspirational at this stage.

4. The Derivative Hedging Program

One mechanism I found genuinely interesting: CleanSpark uses Bitcoin-linked derivative contracts — financial instruments whose value is tied to Bitcoin's price — to generate liquidity without selling coins outright. The 10-Q describes these as serving "as a strategic alternative to selling bitcoin directly and are intended to monetize the Company's bitcoin holdings while managing exposure to adverse price movements." In practical terms, this lets the company access cash tied to its Bitcoin treasury while preserving the underlying holdings. It is sophisticated treasury management, though it also introduces complexity and counterparty risk that simple coin-holding does not.

Notably, 96.6% of CleanSpark's Bitcoin is held in cold storage — meaning offline, air-gapped wallets — with no Bitcoin pledged as collateral under any credit line. That is a meaningful data point: the company is not running the classic DeFi risk of having its collateral liquidated if the price drops further.

What the Cash Flow Picture Actually Looks Like

Cash and equivalents at March 31, 2026 stood at $260.3 million, up dramatically from $43.0 million at September 30, 2025. That improvement was funded primarily by the convertible note issuances and Bitcoin-related derivative activity — not by the operating business throwing off free cash. The company has $137.4 million in outstanding miner deposits (prepayments made to hardware manufacturers for next-generation ASIC mining rigs not yet delivered), and recorded $4.0 million in impairment on unused transformer deposits and $1.4 million on idle mining equipment. Those small impairments signal some operational slack — capacity that is paid for but not yet producing.

Depreciation and amortization ran at $222.2 million for the half-year, against net property and equipment of $1.334 billion. This is an asset-heavy business. The machines depreciate fast, require continuous replacement to stay competitive, and must be sourced from a limited set of manufacturers — mostly overseas.

What Could Break This Thesis

Bitcoin price risk is the dominant variable. BTC fell from roughly $114,000 to $68,222 during the six-month period covered by this filing, generating that $470.9 million non-cash loss and leaving the treasury below cost basis. If Bitcoin continues to decline or stagnates for an extended period, CleanSpark's equity erodes, its debt load becomes harder to refinance, and the entire capital structure comes under stress. The company has $1.16 billion of economic exposure in its Bitcoin holdings and needs the asset to appreciate meaningfully just to return to break-even on treasury value.

Leverage and refinancing risk. $1.788 billion in long-term convertible debt against $986 million in stockholders' equity is a 1.8x debt-to-equity ratio for a company in a cyclical, capital-intensive business. The 2030 and 2032 maturities provide runway, but if Bitcoin is not materially higher by the mid-2020s, CleanSpark will face a refinancing market with limited goodwill and a deteriorated balance sheet.

Mining revenue compression. Post-halving difficulty increases drove a 7.7% revenue decline year-over-year even as the company expanded its holdings. Without a sustained Bitcoin price recovery, every new miner deployed earns less than the one it replaced — a treadmill that requires continuous capital expenditure just to stay in place.

Tariff and supply-chain risk. The 10-Q explicitly flags "uncertainty as to whether the Company will face materially increased tariff liability in respect of miners purchased since 2024 and in the future." Bitcoin mining hardware is a globally supplied, concentrated market. If tariffs materially raise the cost of next-generation ASICs (the specialized chips that mine Bitcoin), CleanSpark's capital expenditure projections become unreliable and its competitive position deteriorates.

The Question the Filing Leaves Open

CleanSpark is not a company in obvious distress, but it is a company that has made a very large bet on a very specific outcome: Bitcoin recovers and appreciates past $97,309 per coin (the average acquisition cost of its treasury), mining economics stabilize post-halving, its convertible debt converts rather than comes due in cash, and the HPC/AI opportunity materializes into real revenue before the capital cycle catches up.

That is a lot of dominoes that need to fall in sequence. What I find credible is management's operational discipline on Bitcoin custody — no leverage against the holdings, cold storage custody, derivative-based liquidity generation rather than forced selling. What I find harder to assess is the $460 million buyback in an environment where the company simultaneously raised $1.13 billion in debt. Concentrating ownership is a compelling argument when you are confident in the recovery; it is a costly mistake if the timeline extends further than the balance sheet can absorb.

The data center land acquisitions in Texas and South Dakota, combined with the HPC/AI language in the filing, suggest management is trying to evolve the story from "pure Bitcoin miner" toward something more like "digital infrastructure operator with multiple workload revenue streams." That transition, if it happens, would change the risk profile meaningfully. But the 10-Q is clear-eyed enough to admit the company has limited experience in those new markets and no hyperscaler customers yet. The ambition is there; the evidence is not.

For investors already holding CLSK, the next twelve months will likely be determined almost entirely by where Bitcoin prices. For those evaluating a new position, the honest framing is this: you are not buying a cash-generative business at a reasonable multiple; you are buying a leveraged, operationally intensive bet on Bitcoin recovery, wrapped in a management team that is making aggressive capital allocation decisions with real conviction and real consequences.