Bitcoin mining can no longer depend on the old playbook of surviving a halving and waiting for a major price rally to restore margins, according to a March 12 report from Wintermute. The firm says that model has started to fail as bitcoin matures into an institutional asset, pushing miners toward high-performance computing and AI-related infrastructure.
Cycle returns have fallen sharply in the current epoch
Wintermute’s argument rests on a simple comparison. In Epoch 3, covering 2016 to 2020, bitcoin returned more than 20x. In Epoch 4, from 2020 to 2024, it returned 10x. Current figures for Epoch 5 show only a 1.15x return. For miners, the report says, that is not a weak stretch inside a normal cycle. It points to a structural break in the economics of the business.
Block rewards have already been cut in half, yet price action has not delivered the kind of upside that once repaired miner profitability. That leaves revenue under direct pressure. Wintermute argues that the industry’s celebrated milestones, including U.S. Securities and Exchange Commission approval of exchange-traded funds and corporate treasury adoption by firms such as Strategy, are also part of what is tightening miner margins.
Institutional ownership changes how bitcoin trades
The report says a more liquid asset held more heavily by institutions is less likely to generate the kind of four-year returns that defined earlier mining eras. As bitcoin trades more like a macro risk asset, similar to tech stocks, volatility has compressed. Stability may suit long-term holders. It is much harder on mining operations built around steep price appreciation.
Wintermute frames this as a regime change rather than a temporary setback. The infrastructure behind mining is still valuable, but its use case is shifting. What miners control is not only hashpower. They also sit on stabilized, high-density power resources that remain in demand beyond bitcoin production.
Fee revenue has not become a durable backstop
The report also challenges a long-standing assumption in mining economics: that transaction fees would eventually replace declining block subsidies. Wintermute calls that idea intuitive, but wrong. Fee revenue, in its view, remains episodic instead of structural.
Bursts linked to Ordinals activity or broader network congestion can lift revenue for a period, but they rarely make up more than a low single-digit share of total mining income. Wintermute’s conclusion is blunt: a business model cannot be built on recurring congestion.
Mining firms may need to repurpose what they have built
That leaves miners with a narrower set of options. The report says the sector still owns assets with real market value, especially dense and stable power infrastructure, yet the profitable application of those assets is changing. In that context, a pivot toward AI data centers is no longer a side business. Wintermute describes it as the most realistic path for staying afloat as bitcoin’s older hyper-growth pattern fades.
Its broader message is that bitcoin has not stopped evolving. The halving mechanism remains, but the assumption that each halving will be followed by explosive price gains strong enough to carry miners through the cycle no longer holds in the same way.