The $18.07 million loan from Arch Lending matured Friday afternoon with no public repayment notice, and the company has not filed the collateral terms.
PowerCompute borrowed roughly $18.07 million from Arch Lending under two bridge notes signed on July 27, and the notes matured at 5 p.m. Eastern on August 1 without a public repayment notice. The bridge equalled 97 per cent of the miner's disclosed bitcoin treasury value. The company has not filed an 8-K describing the outcome. Arch has not commented.
The mechanics are unusual by any standard. A 96-hour bridge on nearly the entire treasury is not corporate financing; it is a payment-timing device. The stated purpose was to repay outstanding debt to Galaxy Digital and Liebel Capital while PowerCompute finalised what it described as an anticipated bitcoin-secured term facility, also with Arch. Neither the size of that follow-on facility nor its loan-to-value ratio or liquidation triggers has been disclosed.
That disclosure gap is what matters. Public bitcoin treasury companies have spent 2026 telling equity investors that BTC-collateralised debt is safer than sceptics allow and that liquidation risk is remote. The absence of LTV ratios, cure periods, and margin-call thresholds in PowerCompute's filings makes it impossible to test either claim from the outside. Investors are being asked to underwrite a capital structure whose most sensitive terms are private.
The context matters because 2026 has already delivered two documented collateral calls in the sector, both disclosed only after the fact. In a separate incident in June, two public companies liquidated a combined 511 bitcoin over 24 hours to retire $31.7 million of debt after a margin call triggered. That is now the pattern: the debt exists, the leverage is real, and the disclosure arrives after the market has already moved on.
Arch Lending has become the most active BTC-collateralised lender to public companies in 2026, largely by offering shorter tenors and higher advance rates than incumbent lenders will underwrite. Short-dated paper of the kind PowerCompute took on is priced for a specific outcome: the borrower closes a larger facility within the maturity window, and the bridge either rolls into that facility or is repaid from its proceeds. If the closing slips, the collateral is already at the lender.
PowerCompute has not published the terms of the follow-on facility, and it has not confirmed whether the July 27 bridge was repaid, extended, or defaulted. What is public is the ratio, and the ratio is the tell. Committing 97 per cent of a treasury to a four-day loan is what a borrower does when it has to. Companies that expect the term facility to close on time do not put the entire balance sheet on a stopwatch.
The wider read is that the bitcoin treasury model has begun to fracture along the lines analysts warned about eighteen months ago. The largest treasuries, notably Strategy's 846,000 bitcoin position and BitMine's roughly 4.8 per cent of circulating ether, can absorb price shocks and roll debt on public terms. Smaller entrants that copied the playbook without matching the balance sheet cannot. Zhibao's PIPE last week traded board control for 2,380 bitcoin, which is the ceiling of what an underfunded treasury has to give up to get on the balance sheet. PowerCompute's bridge shows the floor.
None of this requires PowerCompute to have failed. The bridge may have been extended or refinanced quietly, and the term facility with Arch may still close this week. But the pattern the miner has established — bridge financing on nearly the entire treasury, disclosed only after the deadline passes, with no accompanying detail on collateral terms — is not one investors can price. The trade is to assume the worst until PowerCompute files.
