MARA is reconsidering the status of 53,822 BTC, calling them a ‘liquid asset with the ability for rapid conversion.’
MARA Holdings appears ready to test the resilience of the current strategy of holding bitcoin in corporate reserves. In recent years, major miners have preferred to accumulate BTC as a strategic balance sheet asset rather than using it as working capital. Any departure from this model could have consequences not only for one company, but for the entire sector.
In its March 2 report, the company authorized the sale of its entire reserve of 53,822 BTC. This is effectively a 180-degree reversal from the 2024 policy, when MARA stated its intention to ‘retain all mined and acquired bitcoin for the foreseeable future.’
At the time of publication, bitcoin is trading around $68,000, which is nearly 46% below the late 2025 highs. Meanwhile, market depth has fallen to levels where even moderate sales volumes can trigger disproportionately strong price moves.
In this situation, an inevitable question arises: what happens if one of the industry’s largest holders starts viewing bitcoin not as a strategic bet on the future, but as a tool for operational financing?
From the 2024 HODL Policy to Authorizing BTC Sales in March 2026. Source: MARA Holdings
A Policy They Did Not Plan to Change
In its 2024 10-K report, MARA essentially put itself on par with Strategy, demonstrating commitment to a bitcoin maximalist holding strategy.
The reversal began in late 2025. At that time, the company sold about 4,076 BTC for $413.1 million, at an average price of about $101,000 per coin. The 2026 report directly states that bitcoin is now considered a ‘liquid asset with the ability for rapid conversion,’ and sales from the balance sheet are officially authorized.
Three factors are intensifying the situation.
First, 15,315 BTC are on loan or used as collateral. This is about 28% of total reserves. 38,507 BTC remain freely available, equivalent to about $2.6 billion at current prices or about 60 days of post-halving issuance.
Second, in 2025 MARA recorded a decline in fair value of assets by $422.2 million and a trading loss of $69.1 million.
Third, the company partnered with Starwood Capital to build AI data centers with a capacity of 1 GW and the prospect of expanding to more than 2.5 GW. This is a capital-intensive infrastructure requiring significant liquidity at an early stage.
The logic of the new strategy is simple: finance operating expenses and AI development by selling BTC instead of diluting shareholder equity. However, this approach changes the business model itself. MARA shifts from a kind of ‘bitcoin ETF’ to a capital allocator managing a highly volatile asset.
.6 billion and 15,315 btc pledged or loaned” title=”38,507 BTC Unrestricted Worth >.6 Billion and 15,315 BTC Pledged or Loaned”>
38,507 BTC unrestricted at $2.6 billion and 15,315 BTC pledged or on loan. Source: MARA Holdings
The Timing Is No Accident
If you ask ‘why now,’ three factors converge at once.
First, balance sheet pressure. After the halving, the block reward dropped to 3.125 BTC, while network difficulty and electricity costs continue to rise, squeezing margins. Despite hash rate increasing to 66.4 EH/s, production volume fell by 7% to 8,799 BTC. When the price drops from the $76,000–$126,000 range to $60,000, liquidity becomes a top priority. Additionally, the company has $350 million in convertible bonds maturing in 2027.
See Also: 24/7 Trading Changed Prices Before the Start of the Week on US Exchanges
Second, capital expenditures for AI infrastructure. The partnership with Starwood involves developing sites that can switch between bitcoin mining and AI computing. Starwood is responsible for design and construction, MARA provides the sites and retains up to 50% ownership. The bet is that monetizing computing power will be more profitable than mining in the post-halving environment.
Third, market microstructure. Since late 2025, liquidity has deteriorated and spot trading volumes are 25–30% below year-ago levels. MARA as a discretionary seller does not have to crash the market. The mere fact of potential sales is enough to create an overhang effect at a time when sentiment is already fragile.
In other words, the decision was made not despite weak conditions, but because of them. In the current environment, selling BTC looks like a more realistic source of funding than more expensive alternatives.
Other Miners Also Create a Supply Overhang
The problem of potential market pressure is not limited to one company. Public miners collectively hold 116,697 BTC, down 4.42% from last month.
Of this amount, 53,822 BTC belong to MARA—nearly half the total reserve. The rest is distributed among other major players. Riot Platforms owns 18,005 BTC, CleanSpark holds 13,513 BTC, Hut 8 has 10,278 BTC, and Core Scientific owns 2,537 BTC.
Core Scientific has already announced plans to monetize a significant portion of reserves in 2026. In January, the company sold 1,900 BTC for $175 million at $92,000 per coin. Bitdeer fully liquidated its reserve at the end of February.
In fact, miners are increasingly viewing bitcoin as a commodity asset to be sold if the economics of AI infrastructure are more attractive than expanding hash rate.
Now the question is how quickly and on what scale others will join this strategy. The possible range of developments is roughly divided into three scenarios.
In the conservative scenario, companies sell current production but keep reserves. A 10% reduction in reserves outside MARA would mean selling 6,287 BTC, equivalent to about 14 days of issuance.
In the moderate scenario, miners fund AI capital expenditures by selling 5% to 10% of reserves. For MARA, that’s 2,700 BTC to 5,400 BTC, equal to about 6–12 days of new issuance and $180 million to $361 million at current prices.
If the total reserve reduction reaches 25%, 29,174 BTC could hit the market. That’s about 65 days of issuance.
The aggressive scenario assumes a 50% reduction in reserves. In this case, supply would rise to 58,349 BTC, equivalent to about 130 days of new mining. However, the main risk is more about the market narrative than the absolute volume.
Bitcoin’s daily turnover exceeds $50 billion. Nevertheless, if several major miners become known sellers during a period of macroeconomic stress, the pressure will show up primarily through investor sentiment and derivatives positioning, not through an instant move in spot prices.
MARA’s reporting sets a precedent. Now other companies can follow this example without appearing to be distressed market participants.
What This Reversal Shows
Beyond the scenarios described, three competing interpretations of what is happening are emerging.
The first is the AI pivot. Miners are starting to use energy infrastructure not only for mining but also for data centers, with bitcoin effectively becoming a source of funding for this transition. The MARA partnership with Starwood targets sites that can switch between mining and AI computing. This is not just an operational tweak, but a reallocation of capital from mining to computing infrastructure.
The second is tactical risk management. After a decline in fair value of assets by $422.2 million and trading losses of $69.1 million, the company is starting to treat bitcoin as a managed position. In an environment of low market depth and high macro sensitivity, flexibility in managing liquidity becomes a distinct advantage.
The third is a structural shift. The era of unconditional HODL for miners may be coming to an end. The contrast between the 2024 statement ‘retain all BTC’ and the 2026 position ‘buy or sell as needed’ shows that companies are increasingly acting as capital allocators, balancing mining, grid services, and leasing capacity for AI.
Each of these logics affects supply differently.
If the AI pivot materializes, BTC sales will become a tool for financing transformation. The market pressure will be concentrated in time but limited in scale.
If risk management prevails, sales volumes will start to correlate with volatility, turning miners into countercyclical sellers.
And if it is a full regime change, about 117,000 BTC held in public miners’ reserves could move into active management. This changes the basic assumptions about how the market absorbs supply and finds equilibrium.
The Timing That Matters
The next point where more clarity will emerge is the MARA 10-Q report for the first quarter, expected in mid-May.
Investors will closely watch how much BTC was monetized after the policy change, whether AI development stages are linked to reserve reductions, and whether the company will set targets for minimum holdings or sales pace.
Until the report is published, there will be an information vacuum. This will be filled by macro noise.
Currently, bitcoin is trading in a risk-off mode amid energy shocks and inflation concerns. In such periods, the market is especially sensitive to the question of ‘who might be forced to sell.’
MARA’s reporting does not state an intention to sell most of its reserves. However, the mere authorization to sell is already a price-sensitive factor. With reduced liquidity, the way a deal is executed will determine whether a $1 billion sale is absorbed calmly or accelerates a downward move.
The schedule for the Starwood partnership also creates separate pressure. The company talks about a 1 GW target in the short term with the possibility of expanding to 2.5 GW, but the timeline remains vague.
If MARA accelerates construction to capitalize on AI infrastructure demand, the need for funding will become more concentrated in time. If projects are launched gradually, BTC sales may stretch over years. This will determine whether MARA’s treasury becomes a source of multi-year pressure or a one-off recapitalization tool.
If first-quarter reports show that other miners are expanding sales authorizations or directly linking BTC monetization to AI project funding, the market may start to revalue the entire volume of miners’ reserves as a potential supply overhang rather than a strategic stockpile.
Such a revaluation does not require seeing actual sales. It is enough for investors to stop considering these volumes as ‘locked’ supply.
What Is Really at Stake
MARA’s reversal is important not so much for what it authorizes, but for the signal it sends to the market.
For the past four years, miners have used their BTC treasury as a positioning tool. Their shares effectively became a proxy for bitcoin’s growth. This model worked during rallies, cheap capital, and theoretical post-halving economics.
Now the picture has changed. Bitcoin trades nearly 50% below its highs, capital markets are more willing to fund AI projects, and post-halving margins are tighter than expected.
See Also: Nasdaq Filed to Launch Binary Contracts on the Nasdaq-100 Index
If MARA successfully executes the AI pivot and uses BTC sales as a one-time funding source, the story of reserve reduction could end without long-term consequences. But if projects drag on or bitcoin recovers faster than expected, the company risks selling assets at cyclical lows for investments that may not meet expectations.
For the crypto market, the stakes are clear.
Miners’ reserves have long remained one of the few sources of relatively non-speculative demand for bitcoin. These companies accumulated BTC as part of their operating model, not for short-term trading.
If this segment shifts to active reserve management, bitcoin will lose a structural buyer and gain a structural seller. When the largest miner by reserves formalizes the right to sell its entire stack, it signals that even the most committed supporters are starting to hedge.
