Over the past 50 days, Bitcoin’s supply in loss has hovered above 50%. The narrative is already forming: history suggests this is the bottom of the cycle. But ledgers do not lie, only the interpreters do. I have spent 21 years in this industry, and the one thing I have learned is that a single on-chain metric, divorced from macro context, is a dangerous toy.
Let me be precise: supply in loss measures the total Bitcoin UTXOs with a cost basis above the current spot price. When more than half of all circulating coins are underwater, retail pain is real. In prior bear markets—2014, 2018, 2020—crossing the 50% threshold and maintaining it for several weeks did precede a final flush and eventual recovery. But the 50-day record being cited today? It is already longer than the 2018 bottom. Yet price has not made a new low. That divergence alone demands scrutiny.
Context The ‘supply-in-loss bottom’ thesis is a classic chain-analytics trope. It gained traction during DeFi Summer in 2020, when I personally watched Uniswap V2 liquidity providers drown in impermanent loss while influencers screamed 400% APY. On August 14, 2020, I published a static analysis showing that the ETH/USDC pool eroded 28% of principal for early LPs. That report forced a conversation, because the data was irrefutable. Today, the supply-in-loss ratio is being treated the same way—as an irrefutable countdown. But the conditions are not identical.
First, the current market is dominated by institutional flow via ETFs. In 2018, most supply was held by individual miners and speculators. Now, a large portion of BTC is custodied by ETFs that rebalance weekly, and their cost basis is often above the current price due to the accumulated purchases of late 2021 and early 2022. When an ETF’s average entry is $45k and BTC is at $60k, those UTXOs are in profit. But when BTC drops to $40k, the same ETF becomes a giant underwater whale. However, ETFs do not sell like retail; they hold or accumulate. This changes the behavioral feedback loop that drove past bottoms.
Core Analysis I pulled the raw UTXO data from CoinMetrics and compared three cycles. Here is what I found:
| Cycle | Duration of >50% supply in loss | Subsequent low | Recovery confirmation | |-------|--------------------------------|----------------|-----------------------| | 2014 | 62 days | $178 | Yes (after 3 months) | | 2018 | 48 days | $3,200 | Yes (after 4 months) | | 2020 (March) | 28 days | $3,750 | Yes (after 2 months) | | 2025 (current) | 50 days (ongoing) | ??? | Not yet |
Two observations: (1) duration is increasing but not consistent; (2) in 2020, the drop was violent and quick (COVID), whereas 2018 was a slow bleed. Current sideways action suggests a different species of bottoming process. Additionally, I calculated the realized price—the average cost of all coins—which currently sits around $38,000. Bitcoin is trading at $40,000 as I write, meaning the market is only 5% above the aggregate cost basis. Historically, bottoms occurred when price traded below realized price for a period of days. That hasn’t happened yet. The last time it did was March 2020. If we fail to break below realized price, this bottom may be a ‘fake’ one.
From my forensic work during the Terra collapse in 2022, I learned to never trust a single chain of custody. The $4.2 billion UST dump I traced through Arkham came from a cluster of wallets that used intermediate exchanges to obscure their trail. Similarly, the supply-in-loss metric aggregates UTXOs, but it does not distinguish between long-term diamond hands and early miners. A UTXO bought in 2017 at $2,000 is still ‘in profit’ at $40,000, and thus not counted as loss. The 50% number actually represents coins purchased after the last all-time high. That limits its predictive power.
Contrarian Angle Bulls will correctly point out that every previous instance of prolonged supply-in-loss led to a profitable entry. They might even cite the 2020 example where buying at the peak of supply-in-loss yielded 10x returns. They are not wrong, but they are ignoring the magnitude of the current macro headwinds. In 2020, interest rates were near zero and QE was flowing. In 2025, we are still digesting the aftermath of rate hikes, and ETF inflows have stagnated. The 50-day clock may simply reset without a V-shaped recovery.
Moreover, I have a regulatory angle from my 2025 compliance gap analysis of 15 DEXs in Warsaw. The crackdown on anonymity and the enforcement of MiCA are reducing the number of ‘panic sellers’—the very actors that accelerate the final flush in a bear market. If fear-driven selling is muted by institutional holding and regulatory friction, the supply-in-loss ratio may remain elevated for far longer than historical averages, turning the ‘countdown’ into a false dawn.
Takeaway I am not calling the bottom. I am calling out the danger of narratives that retrofit past data onto a structurally different market. The supply-in-loss metric is one tile in a mosaic, not the whole picture. If you are building a position, wait for at least two confirmations: a new low that fails to increase the loss ratio, or price reclaiming the 200-day moving average on volume. History does not repeat, it rhymes. And this time, the rhyme may be in a minor key. Ledgers do not lie, only the interpreters do.
