The throne of the mining world has genuinely changed hands this time.

MARA has long been seen as the established leader in computing power, with the scale to justify its status and a position that looked secure.

But the latest data set off a stir across the sector: Jihan Wu’s Bitdeer now manages 71 EH/s of total hashrate, putting it ahead of MARA’s 61.7 EH/s.

The numbers need a closer read. Of Bitdeer’s 71 EH/s, 55.2 EH/s is self-owned, while the remaining 15.8 EH/s is hosted for third parties. On self-owned computing power alone, MARA still has a slight lead for now.

Put another way, Bitdeer wins on the size of the platform; MARA wins on the depth of its own asset base.

But the overtake itself is the signal. For the first time, the old leader’s position has been shaken.

The more important question is what Bitdeer used to pull ahead.

The Hidden Killer Move

Bitdeer did not overtake MARA by simply spending heavily on mining rigs. It did it with a card others cannot easily copy: Norway.

In Norway, 88% of electricity comes from hydropower and 10% from wind, making the grid almost entirely powered by clean energy.

At a time when ESG and environmental thresholds are tightening globally, the country has become a kind of promised land for mining companies. Norway’s cool climate also provides natural cooling conditions for data centers, along with advantages such as relatively low power costs.

Bitdeer has built a 175 MW hydro-cooled mining farm in Tydal, Norway, and brought it fully online.

One mining executive with operations in the Nordics said, “Everyone is trying to get into Norway now, because electricity is either expensive elsewhere in Europe or the policies are unfriendly.” Bitdeer is not the only company eyeing the region.

Canadian company Bitzero has also put down roots in Norway. On January 27, it disclosed that its computing power had jumped 59% from 1.76 EH/s to 2.8 EH/s. Its target is to reach 10 EH/s by the end of the year.

But Norway’s door is closing. The country’s digitalization minister said, “Crypto mining consumes huge amounts of electricity, but creates very little local employment or income.”

The Labour Party government also has plans to ban new PoW mining data centers.

One analyst summed it up this way: “In the second half of the computing power race, the contest is no longer how many mining rigs you own, but how much compliant clean power you locked in before the policy curtain came down.”

The existing power capacity held by Bitdeer and Bitzero has become a scarce asset that cannot be replicated.

New entrants face a wall of policy restrictions. Bitdeer’s Norway position is worth far more than the computing power figure on paper. What it secured was an access ticket, one that may not be available at any price in the future.

The Industry Is Undergoing a Quiet Shakeout

Being No. 1 in computing power looks impressive, but behind the headline the entire industry is walking a tightrope.

Inefficient miners are shutting down in batches. The reason is simple: every coin mined loses money.

According to MacroMicro’s estimates, the cost of mining one BTC on January 19 was about $101,000, while the market price was only $93,000, implying a net loss of $8,000 per coin.

Total network computing power has fallen 15%, from a peak of 1.1 ZH/s last October to 977 EH/s.

A mid-sized mining farm owner ran the numbers for me: power prices above 5 cents are the breakeven line.

The current measure of daily Bitcoin mining revenue per unit of computing power is around $39, already close to the threshold where turning machines on means burning cash. “Many mining farms have cut utilization below 60%. They would rather leave machines idle than run them at full power.”

The payback period for new mining rigs has already stretched beyond 1,000 days. In the 2017 bull market, by contrast, miners could recover their costs in three to six months. Leading mining companies are under pressure as well.

Winter storm “Fern” at the end of January was a stress test. Marathon’s daily output plunged from 45 BTC to 7 BTC, while Riot fell from 16 BTC to 3 BTC. Total network computing power at one point dropped to a low of 663 EH/s.

What worries industry participants even more is the outlook for electricity prices. The U.S. Energy Information Administration forecasts wholesale power prices will rise 8.5% in 2026, while the ERCOT-North node in Texas could surge 45%.

One institutional investor in the mining sector was deeply concerned: “The computing power figures look strong right now, but everyone is walking a tightrope. For heavy-asset giants, once the stock price swings or power prices rise even slightly, the daily operating costs they burn through could drag down a mid-sized company.” Yet as smaller miners shut down and exit, the relative advantage of the leading players is only reinforced.

For example, Bitdeer self-mined 636 BTC in December 2025, up 339% year on year.

But beyond that relative advantage, the direction of the industry also deserves caution. Bitfarms, a Nasdaq-listed Bitcoin miner founded in 2017, is an established mining company operating mining farms in Canada and the United States.

Its computing power scale ranks in the upper-middle range among listed companies. It is not the largest, but it is still a meaningful player.

Bitfarms CEO Ben Gagnon said bluntly in an interview: “High-performance computing creates far more value per unit of energy than mining, and the revenue is more stable.

That makes it impossible for the company to justify continued investment in Bitcoin mining.” In 2025, Bitfarms announced a full phased shift toward AI infrastructure.

AI data centers are systematically drawing away energy that once went to crypto mining. Mining and AI are now competing for food from the same pot.

Two Paths, Two Bets

That brings the contest back to Bitdeer and MARA.

On the numbers, Bitdeer has won. But the factors that will really decide the outcome do not show up neatly on paper.

Bitdeer’s lead does not come from pure self-mining. It comes from a combined “self-mining plus hosting” model: it earns money from mining on its own account while also collecting service fees by hosting machines for others, blending heavy and lighter asset models.

Investors say this model gives the company steadier cash flow when the market is weak. MARA has taken a different path: heavy exposure to self-mining and holding coins rather than selling them.

As of the end of last year, MARA held more than 50,000 BTC. Its bet is on long-term Bitcoin price appreciation, using its balance sheet to ride out the cycle.

Two models, two bets. One analyst said: “Bitdeer’s model is more flexible, but rapid expansion can easily spin out of control; MARA’s model is steadier, but if Bitcoin prices stay depressed for a long time, the pressure will be heavy.” One mining farm owner joked, “There is no finish line in this business.

If you dare stop for a day, the people behind you will step on you and rush ahead.” But at the end of the race, they may find three walls waiting: policy, cost, and technological iteration.

Is Bitdeer’s lead today a temporary victory, or the start of a new industry order? The answer may not become clear until the next halving cycle.

But one thing is certain: the mining world of 2026 looks more like a war of attrition than at any point in the past.