KULR's Bitcoin Retreat: A Treasury Strategy Unraveled by Debt and Volatility

The 30% drawdown happened in weeks. 333 BTC sold for $21.5 million. A Coinbase loan repaid, collateral released, mining contracts terminated. KULR Technology Group did not exit Bitcoin gradually—it executed a structured retreat. The numbers tell a story of leverage unwinding, not strategic repositioning. Echoes of past bubbles resonate in current code. The same pattern: accumulate, pledge, mine, then sell under pressure. KULR's board now authorizes BTC sales for operational liquidity. The accumulation phase is dead. Context: KULR launched its Bitcoin treasury strategy in late 2024, allowing up to 90% of surplus cash into BTC. During the first half of 2025, it spent $69.9 million to acquire 693.81 BTC. By June 30, 2026, the position stood at 1,091.69 BTC with a cost basis of $109.8 million—but a market value of only $63.92 million. That's a $45.9 million unrealized loss, excluding the $10.59 million fair-value writedown in Q2 alone. Core: The mechanics of the unwind are classic. KULR pledged 565 BTC against a $20 million Coinbase credit facility. In March and May, it drew $5 million and $15 million respectively. After June 30, the company sold 333 BTC for $21.5 million, used $20 million to repay the principal, and freed the collateral. The 565 BTC returned to the treasury, but the net position dropped by roughly 30% from June 30 to ~760 BTC. Mining followed the same path. One contract expired July 30. Another was terminated early for $150,000, eliminating $2.1 million in future commitments. Q2 mining revenue fell to $606,000 from $1.12 million year-over-year, despite a slight increase in BTC produced (8.44 vs 11.25). The average Bitcoin price earned dropped from $96,225 to $73,594. The math is unforgiving: when the asset price declines, the mining margin compresses faster than the hash rate adjusts. Echoes of past bubbles resonate in current code. The 2020-2021 bull run saw similar narratives: corporate treasuries as BTC accumulation vehicles. But the debt structure reveals the fragility. KULR's $20 million loan was collateralized at roughly 1.7x BTC value. A 40% drop would trigger margin calls. The company sold before that happened, but the timing suggests a preemptive move, not a strategic pivot. CFO Mike Kimel cited Bitcoin volatility obscuring the core battery business. Revenue fell 43% to $2.08 million, operating loss widened 19% to $11.2 million. The net loss of $21.97 million is partly attributable to the Bitcoin writedown. When a non-core asset drives a material portion of losses, the board's decision is rational. But the contrarian angle: the bulls were right that BTC provides liquidity optionality. KULR used the loan to raise cash without diluting equity. The $20 million facility was cheaper than issuing shares. The company issued no shares through its ATM program in H1 2026. That's a win for the treasury strategy as a financing tool. The problem was not the asset class but the execution: buying at high prices, pledging too much, and failing to hedge. Echoes of past bubbles resonate in current code. The same mistakes repeat: over-leverage, lack of hedging, and treating volatile assets as stable reserves. KULR's retreat is not a condemnation of Bitcoin as a corporate asset. It is a case study in poor risk management. The protocol-level analysis shows that the Coinbase loan terms were standard—20% overcollateralization, no liquidation grace period. The company's failure to maintain a sufficient buffer was a governance failure, not a market one. Takeaway: The next cycle will see fewer companies mimicking MicroStrategy without understanding the balance sheet mechanics. KULR's retreat is a warning: treasury strategies must account for volatility, not assume it. The board's decision to sell and refocus on core operations is a surrender to reality. But the real lesson is structural: when a company's primary business is not generating enough cash to cover its Bitcoin losses, the treasury becomes a liability, not an asset. The chain does not forgive poor math.