Cointelegraph is committed to providing independent, high-quality journalism across the crypto, blockchain, AI, and fintech industries.

All news, reviews, and analyses are produced with full journalistic independence and integrity. For more details on our standards and processes, please read our Editorial Policy.

Ether.fi has captured roughly $223 million in annualized fees and about $51 million in annualized revenue. In the second quarter, it earned $41 million in gross revenue and nearly $10 million in earnings after rewards and other costs, with only $30,000 of value distributed to ETHFI holders through buybacks.

The split lands as Ethereum’s staking economics are under debate.

A group of Ethereum researchers, one from the Ethereum Foundation, proposed this week that the network stop paying people to stake once half of all ether is locked up. Under the current setup the payment never falls to zero no matter how much gets staked, so there is always a reason to stake more, and they argue that concentrates ether with a handful of large custodians.

Their proposed fix destroys a growing share of the rewards until the payment disappears entirely at around 60 million ether. About a third is staked today.

Ether.fi founder Mike Silagadze was among the proposal’s critics, arguing it would push out smaller stakers and weaken the products built on staking rewards, his own among them.

MARA Holdings reported weaker second-quarter 2026 financial results on Aug. 6. Revenue fell 27% year over year to $174.9 million, while the Nasdaq-listed company recorded a $611.3 million net loss and negative adjusted EBITDA of $360.9 million, according to its official presentation.

The company ended June with 35,577 BTC, down 29% from 49,951 BTC a year earlier. However, that headline decline masks a small sequential increase from 35,303 BTC at March 31. MARA’s presentation also showed approximately $2.5 billion in combined cash and Bitcoin holdings at quarter-end. Shares closed Aug. 6 at $10.65, down 5.25%, according to Google Finance data.

Source: Google Finance

MARA’s Bitcoin holdings fell after heavy first-quarter sales

The annual decline in MARA’s Bitcoin treasury largely reflects sales earlier in 2026 rather than falling mining production. Its first-quarter filing showed the company sold 20,880 BTC for about $1.5 billion as it funded operations, repurchased debt and pursued new infrastructure investments. Earlier Q1 coverage detailed how the sales reduced MARA’s position from 53,822 BTC at the end of 2025.

During Q2, MARA sold another 2,213 BTC at an average price of $73,078 while producing 2,422 BTC. That left holdings slightly higher than at the end of March. The company’s treasury policy now permits opportunistic sales of balance-sheet Bitcoin, a change from its earlier emphasis on retaining mined coins.

At June 30, 4,742 BTC were loaned and 4,528 BTC were pledged as collateral, while 26,307 BTC were unrestricted. After the quarter ended, MARA pledged another 18,750 BTC as initial collateral for two Bitcoin-backed credit facilities, increasing the portion of its treasury being used to support financing.

Mining output rose despite weaker Bitcoin economics

Operational performance improved in several areas. Energized hashrate reached 70.3 EH/s, up 22% from 57.4 EH/s a year earlier. Bitcoin production increased 3% to 2,422 BTC, and blocks won rose 1% to 700. Cost per petahash per day improved 4% to $27.70 from $28.70.

Those gains did not prevent revenue from falling because the average Bitcoin price associated with mining revenue dropped sharply from the prior-year period. MARA reported an average price of Bitcoin mined of roughly $71,325, compared with $98,975 in Q2 2025. Purchased energy cost per Bitcoin at owned sites also rose, showing that greater hashrate alone did not remove profitability pressure.

The net loss was also affected by Bitcoin price accounting. The company recorded roughly $343 million of fair-value losses tied to digital assets and related receivables. That contrasts with the large fair-value gains that supported earnings in the year-earlier quarter and helps explain the swing from $808.2 million in net income to the latest loss.

MARA is using its Bitcoin balance sheet to fund an AI pivot

The company’s strategy increasingly links its Bitcoin reserves with expansion into power and computing infrastructure. After quarter-end, the company arranged two credit facilities that provide $600 million of incremental borrowing capacity and pledged 18,750 BTC as initial collateral. Proceeds may support general corporate purposes, including the planned Long Ridge acquisition.

The Long Ridge transaction is central to MARA’s effort to add AI and high-performance computing capacity. Earlier AI pivot coverage reported that the proposed $1.5 billion acquisition includes a 505-megawatt Ohio gas plant and a campus with potential for more than one gigawatt of computing capacity. The deal still requires regulatory approval before closing.

The firm is also pursuing a Texas project. Its Texas expansion plan covers a 1,200-acre powered site expected to provide up to 2 GW of grid capacity over time. Combined with Long Ridge and other assets, management says its potential power portfolio could reach about 4.8 GW.

What happens next for MARA

The main near-term milestones are financing execution, regulatory approval for Long Ridge and progress on the Texas development. Investors will also watch whether MARA continues selling or pledging Bitcoin as it funds infrastructure. The company has made clear that its treasury can serve as both a long-term asset and a source of liquidity.

Chief Executive Fred Thiel said Bitcoin mining provided the company’s foundation and that digital infrastructure and other initiatives “will expand the value we create from that foundation.” The statement is forward-looking, and the Q2 results show the transition remains costly. Mining output improved, but weaker Bitcoin pricing, higher per-coin energy costs and fair-value losses weighed heavily on reported results.

For now, MARA remains a major public Bitcoin miner and one of the largest corporate Bitcoin holders. Its next quarters will test whether expanding AI infrastructure can add steadier revenue while the company preserves enough Bitcoin exposure to benefit from a recovery in mining economics.

Bitcoin traded near $64,350 on Friday, unchanged on the week, as the whole market drifted ahead of the US payrolls report, per CoinDesk data. Ether held at $1,903 and the rest of the majors sat within a point or two, a market waiting on the data rather than moving on anything of its own.

The setup turned slightly less friendly overnight. Brent rose 1.4% to $83.61 after reports Iran will try to restrict US and Israeli ships through the Strait of Hormuz and demand compensation from countries it deems hostile before letting them pass, stalling the deal that had been pulling oil lower. Higher crude revives the inflation worry that keeps the Fed leaning tight, and the 10-year Treasury yield climbed seven basis points on it during the US session.

That macro chain is the one bitcoin has been stuck inside all summer. Oil up feeds inflation, inflation keeps yields and the dollar firm, and firmer financial conditions cap risk assets. The dollar just posted its best day in two weeks, which is the opposite of the easing setup bulls want.

Today’s jobs number is the release that matters. A soft print revives the case for the Fed to loosen and gives bitcoin room above its range. A strong one, stacked on climbing oil, hands the hawks another reason to hold, and the range that has held since May holds again. Watch the reaction in yields, not just the headline number.

Sui has added two NIST-approved post-quantum signature schemes to its blockchain roadmap as it prepares optional quantum-safe accounts and vaults for future network upgrades.

According to Sui’s latest announcement, the blockchain plans to introduce ML-DSA-65 as a native signature scheme for regular accounts and SLH-DSA-SHA2-128s for high-value smart contract vaults, giving users an optional way to protect accounts against future quantum computing risks without replacing their recovery phrases or moving assets.