One recent shift in the crypto market may be more noteworthy than Bitcoin’s short-term price fluctuations: mining companies that once focused on accumulating BTC as their primary asset strategy are gradually changing how they operate.

On one hand, U.S. spot Bitcoin ETFs continue to see outflows, temporarily weakening institutional demand for BTC through ETFs. On the other, some publicly listed mining companies have begun selling their Bitcoin holdings to repay debt and raise funds for construction, and even converting power infrastructure originally used for mining into AI data centers.

Meanwhile, the minutes of the Fed’s October meeting sent hawkish signals, further heightening market concerns about future financing costs.

These developments may seem unrelated, but they all point to the same issue: the crypto market is undergoing a reallocation of capital.

In the past, mining companies earned BTC through mining and accumulated assets by holding it. Now, some are considering how to turn their Bitcoin reserves, power resources, and infrastructure into new sources of operating revenue.

This not only changes mining companies' own business models, but could also affect the future supply and demand dynamics of the Bitcoin market.

I. What Signal Did the Fed Minutes Send? The High-Interest-Rate Environment Could Last Longer

Let's start with the macroeconomic environment.

On November 18, the Federal Reserve released the minutes of its October 27–28 monetary policy meeting.

The minutes conveyed a more cautious signal than the market had previously expected.

The minutes showed that almost all participants believed it could be appropriate to raise the federal funds rate further in the near future to ensure inflation returned to the 2% policy target in a timely manner.

Some officials believed that if upcoming inflation data were disappointing, a rate hike of more than 25 basis points could even be considered in December.

This means that the Fed remains highly concerned about inflation.

Although the October meeting ultimately ended with a 10–2 vote to leave rates unchanged, two members supported a 25-basis-point hike at the time.

Market expectations for a December rate hike are also elevated, with interest rate futures pricing at one point implying a probability close to 95%.

There's an economic rationale worth understanding here.

When interest rates remain high, the cost of using capital increases.

For companies, borrowing to expand, buying equipment, and building infrastructure all entail higher financing costs.

For investors, the opportunity cost of holding volatile assets also changes.

For example, investors who were previously willing to allocate funds to BTC may now place greater importance on cash flow, liquidity, and returns on dollar-denominated assets.

If the market believes interest rates will remain relatively high for some time, some investors may reduce their allocations to high-risk assets.

This is also why Bitcoin, tech stocks, and other growth assets often come under pressure at the same time when rate hike expectations rise.

But I think there is another easily overlooked point in these meeting minutes.

In its discussion of financial stability, the Fed staff specifically mentioned debt risks associated with AI infrastructure development, including off-balance-sheet lease commitments and circular financing arrangements between AI companies.

If returns on AI investments fall short of expectations, these financing structures could magnify losses.

This shows that even as the AI industry continues to grow rapidly, capital markets are paying closer attention to the relationship between investment scale, financing structure, and ultimate returns.

This issue happens to be connected to Bitcoin mining companies that are undergoing a transition.

II. ETFs See Consecutive Outflows as Institutions Reassess BTC Allocations

Aside from Fed policy, another data point worth watching is the flow of funds into and out of U.S. spot Bitcoin ETFs.

According to the latest data as of November 18, U.S. spot Bitcoin ETFs saw net outflows of about $306.4 million in a single day.

Of that amount, BlackRock's IBIT saw net outflows of about $148.7 million, Fidelity's FBTC about $69.3 million, and Bitwise's BITB about $36.6 million.

On November 17, the ETF market had already seen net outflows of about $239.9 million.

This means that U.S. spot Bitcoin ETFs saw fairly significant outflows on two consecutive trading days.

Looking at the cumulative figures for November, net outflows have also reached the billions of dollars.

These figures are worth taking seriously because spot ETFs have become an important channel for traditional financial markets to gain BTC exposure.

Through ETFs, some institutions and retail investors can gain exposure to Bitcoin's price without having to manage digital wallets directly.

ETF fund flows therefore offer some indication of how demand for BTC through traditional financial channels is changing.

But it's also important to avoid a common misconception here.

ETF net outflows do not mean that all institutional investors are selling Bitcoin.

Investors in different funds have different profiles, holding periods, and trading objectives. ETF outflows also cannot fully represent the flow of funds across the entire on-chain market.

A more reasonable interpretation is that net demand entering the BTC market through U.S. spot ETFs is currently weakening.

When institutional demand for allocations declines while the market faces expectations of higher interest rates, Bitcoin is more likely to be affected by changes in capital flows in the short term.

For retail market participants, rather than focusing only on price movements, it's more useful at this point to watch whether ETFs return to sustained net inflows and whether spot-market trading volumes gradually recover.

If prices rebound while ETFs continue to see outflows, market demand may not have fully recovered.

If prices stabilize as ETF inflows gradually increase, that may indicate improving demand for allocations through traditional financial channels.

III. Why Are Mining Companies Starting to Sell BTC? The Old HODLing Strategy Is Changing

I think the most interesting development to examine in this market is Bitcoin mining companies.

In the past, many publicly listed mining companies followed a relatively simple asset-accumulation model:

They participated in the Bitcoin network with mining machines, earned block rewards and transaction fees, and kept some of the BTC they mined on their balance sheets.

When the price of BTC rose, the value of the digital assets held by these companies rose as well.

This model can be appealing during a bull market.

The problem, however, is that mining is a business that requires ongoing capital investment.

Mining companies have to pay for electricity, buy mining machines, maintain equipment, and build mining facilities. They also have to manage revenue volatility caused by changes in network difficulty.

When the price of BTC falls but electricity and equipment costs do not fall at the same time, mining companies' profit margins can shrink.

If a company also has debt, it needs to account for interest expenses and repayment schedules.

In that situation, continuing to hold large amounts of BTC may not be the most efficient way to manage assets.

Selling part of their reserves for cash to repay debt or invest in new business projects may be a more practical option.

MARA Holdings is a fairly typical example.

According to the company's recent disclosures, MARA sold 13,597 BTC over six weeks, for a total value of about $1.1 billion.

About $750 million of that was used to repay loans collateralized by Bitcoin.

After this adjustment, the share of the company's remaining BTC that was loaned out or used as collateral fell from about 57% to less than 5%.

The company also said it may continue to sell Bitcoin in the future to fund AI infrastructure and other business opportunities.

This means that mining companies' former business model of accumulating assets by holding BTC is gradually shifting toward flexible management of digital assets and physical infrastructure.

It's worth noting that this does not mean all mining companies are preparing to exit Bitcoin mining.

Some companies still view mining as their core business, but are reallocating power and capital among different projects.

What is clear, however, is that for some publicly listed mining companies, BTC is no longer just a reserve asset to hold for the long term. It has also become an important source of funding for their operations and transition.

IV. Why Have Mining Facilities Suddenly Become an Important Resource for AI Companies?

Selling BTC is only the surface-level change for mining companies. The deeper shift is happening at the level of physical infrastructure.

Why are more and more mining companies turning their attention to AI data centers?

The reason is simple: power resources and the conditions needed to build data centers are becoming important competitive factors in the AI industry.

Training and running inference on large AI models requires a lot of computing equipment to operate continuously.

This equipment requires not only GPUs, but also a stable power supply, cooling systems, network connectivity, and a suitable data center environment.

Building a large data center is not as simple as just buying servers.

Site selection, land-use approvals, grid connections, transformers, power capacity, and cooling systems all require substantial time and investment.

Companies that already operate large mining facilities have some of these important resources in place.

For example:

First, access to power.

Bitcoin mining facilities typically require a large, continuous power supply, which overlaps to some extent with the basic needs of AI data centers.

Second, land and infrastructure.

Mining facilities already have some land, buildings, and supporting power infrastructure that can provide a foundation for future conversion.

Third, experience in energy management.

Mining companies have long dealt with fluctuations in electricity prices and need to adjust their equipment operations based on power costs, network difficulty, and BTC prices.

This experience can also help them understand how to operate large-scale energy infrastructure.

As a result, some mining companies are converting their existing facilities into high-performance computing and AI data centers.

Riot Platforms is a noteworthy example.

According to the company's third-quarter disclosures, Riot has signed lease agreements for a total of about 1,151 MW of critical IT load capacity, representing roughly $38 billion in long-term contracted revenue.

Since the start of this year, the company has sold 20,264 BTC, raising about $1.5 billion, primarily to fund data center development and related projects.

This shows that some mining companies are no longer relying solely on mining revenue. Instead, they are trying to use their existing power and land resources to establish more long-term sources of infrastructure revenue.

This represents a clear shift in their business models.

In the past, mining companies' revenue was driven mainly by BTC prices, network difficulty, and electricity costs.

In the future, if AI data center leasing becomes established, companies may be able to generate relatively stable infrastructure revenue through long-term contracts.

But that doesn't mean the transition is guaranteed to succeed.

AI data centers require more reliable power, more complex cooling systems, and substantial investment in equipment.

Having access to power at a mining facility does not mean it can immediately meet all the technical requirements of AI customers.

Conversion costs, construction timelines, and financing terms will all affect whether a project can ultimately be completed.

V. Transitioning to AI Is Not a Risk-Free Shortcut

Mining companies' entry into AI infrastructure may look like a highly attractive business upgrade.

But I think there is an issue that the market may be overlooking: AI data center construction also requires significant capital.

Although Bitcoin mining's business model is affected by BTC prices and network difficulty, its equipment investment, energy consumption, and revenue cycles are relatively easy to understand.

AI data centers, by contrast, have to accommodate different customer needs, equipment standards, and long-term contract arrangements.

For example, a mining company planning to convert an existing facility into an AI computing center may need to invest in additional power equipment, cooling systems, server infrastructure, and network infrastructure.

If a project is delayed, the company will have to bear additional construction costs and financing expenses.

If a customer postpones deployment, the expected lease revenue may also be delayed.

More importantly, AI customers typically have very high data center reliability requirements.

Power interruptions, inadequate cooling, and substandard network quality could all affect customers' ability to use the facility.

Mining companies therefore cannot assume that having access to power means they already have the capabilities needed to fully operate an AI data center.

A cautionary example has recently emerged in the market.

Hyperscale Data had previously sold all of its BTC, hoping to use the funds for an AI data center project in Michigan.

But on November 18, the company disclosed that it still needed to raise an additional $74.8 million to complete the related infrastructure.

As of the end of September, the company had only about $17 million in unrestricted cash, while its current liabilities amounted to about $231 million.

Management also expressed substantial doubt about the company's ability to continue operating over the next 12 months.

This case shows that selling BTC does not mean a company has solved its financing problems.

If a new business requires more funding than a company can provide through its existing cash and financing capacity, it may face liquidity pressures even if it has long-term customer contracts.

So I think the competition among mining companies transitioning to AI may ultimately be less about who has the most power resources and more about who has a better financing structure, more reliable construction capabilities, and customer contracts they can actually deliver on.

VI. Three Forces Are Changing the Operating Environment for Bitcoin Mining

Taken together, Fed policy, ETF fund flows, and mining companies' transitions reveal three forces at work simultaneously.

The first force is the cost of financing.

The hawkish signal from the Fed minutes suggests that the cost of capital may remain high for some time.

For companies that need to buy mining machines, expand mining facilities, or build data centers, financing costs directly affect project returns.

This is especially true for companies that rely on debt financing: higher interest expenses could lengthen the time it takes for projects to become profitable.

The second force is institutional demand for allocations.

Consecutive net outflows from U.S. spot Bitcoin ETFs suggest that demand for BTC through traditional financial channels is temporarily weak.

If mining companies also begin selling their BTC reserves, the market will need to reassess the potential supply coming from corporate holdings.

However, BTC sold by mining companies does not necessarily all go directly to exchanges, and the value of corporate asset sales should not simply be equated with immediate selling pressure in the market.

The third force is the changing business model of mining companies.

Some mining companies are shifting resources away from traditional mining and toward AI data centers and high-performance computing.

This may diversify their sources of revenue in the future, but it could also change their long-standing practice of holding BTC.

In the past, the market sometimes viewed publicly listed mining companies as an indirect way to hold BTC.

If more mining companies sell the BTC they mine to fund other businesses, this indirect-holding thesis will need to be reassessed.

I think these three changes together show that the Bitcoin market is moving beyond a narrow focus on price and block rewards, and entering a phase that places greater importance on institutional capital, corporate balance sheets, and the allocation of energy resources.

VII. How Should Retail Participants Understand This Shift in Capital Flows?

For people who follow BTC, I think it's important to develop three new perspectives to monitor the market.

First, don't look only at daily ETF inflows and outflows.

ETF fund flows can reflect allocation demand in traditional financial markets, but they do not represent the entire Bitcoin market.

You should also consider exchange balances, long-term holder behavior, on-chain transfers, and spot trading volume.

Second, pay attention to the real reasons mining companies are selling BTC.

If a company is selling BTC mainly to repay high-cost debt, its balance sheet may be improving.

If the proceeds are being used for new AI projects, it's important to monitor the projects' financing and construction progress.

If a company is forced to sell its reserves because of operational difficulties, the market may pay more attention to its financial risks.

The same action—selling BTC—can have very different business implications depending on the circumstances.

Finally, distinguish between Bitcoin itself and the stocks of publicly listed mining companies.

BTC prices are driven mainly by market supply and demand, capital flows, and macroeconomic conditions.

Mining stocks require additional consideration of electricity costs, equipment depreciation, corporate debt, operating cash flow, and progress on AI projects.

Even if BTC prices rise, that doesn't mean all mining stocks will rise in tandem.

Conversely, a mining company signing an AI data center contract does not necessarily have a direct positive effect on BTC.

This distinction is especially important for users who follow both digital assets and traditional stock markets.

For example, BiyaPay offers digital asset spot trading and also connects users to the U.S. and Hong Kong stock markets.

If you usually use limit orders to trade BTC on the spot market, you can learn about the platform's zero-maker-fee offer and calculate your trading costs based on the actual execution type, bid-ask spread, and any other applicable fees.

If you're researching publicly listed mining companies such as MARA and Riot, you need to analyze their corporate finances, debt structures, and AI project progress separately. You can't use BTC price movements as a substitute for company research.

VIII. Why Is Cross-Market Capital Management Becoming More Important?

This shift also brings to mind another practical issue.

As the links between Bitcoin, AI infrastructure, and traditional financial markets grow closer, users' capital management needs will also become more complex.

For example, some users hold a certain amount of U while also following BTC spot markets, U.S. tech stocks, and Hong Kong-listed companies.

As mining companies' transition to AI becomes a new market trend, they may want to research both digital assets and U.S.-listed companies in energy infrastructure and data centers.

At that point, in addition to assessing the value of the assets themselves, they also need to consider how to allocate funds across different markets.

BiyaPay offers U-to-USD conversion and supports trading in U.S. and Hong Kong stocks.

Users who hold U and also follow traditional stock markets can learn in advance about conversion quotes, the fund conversion process, and applicable fees.

For U.S. stock trading, the platform offers commission-free trading, which can be one cost factor to consider when comparing different trading channels.

For users who also need to manage different types of digital assets, the platform supports exchanges between more than 30 fiat currencies and more than 200 cryptocurrencies. Check the actual availability of the relevant conversion services.

However, zero-commission stock trading does not mean the entire trading process is cost-free. The actual exchange rate, bid-ask spread, and other applicable fees still need to be considered.

I think what the market needs most right now is not frequent switching between assets, but ensuring that every portion of capital has a clear purpose.

For spot digital asset trading, pay attention to market depth and price volatility.

When researching publicly listed mining companies, understand their operations and financing.

When investing in traditional stocks, consider company valuations, earnings, and industry prospects.

Cross-market capital management only makes sense once you understand the sources of risk for different assets.

Final Thoughts

The Fed minutes, consecutive outflows from Bitcoin ETFs, and mining companies' shift toward AI infrastructure may seem like three separate developments, but in fact they reflect the same trend:

Global capital is reassessing how efficiently funds are being used.

A high-interest-rate environment makes companies more cautious about financing. ETF outflows reflect temporarily weaker demand for BTC through traditional financial channels, while mining companies selling BTC to build AI data centers shows that some businesses are rethinking their future sources of revenue.

In the future, Bitcoin mining may no longer be about competing solely on the number of mining machines or the scale of computing power.

Power resources, infrastructure, long-term customer contracts, financing capacity, and energy management efficiency could all become new competitive factors.

But the transition also comes at a cost.

Selling BTC gives mining companies access to capital, but may also reduce some of their long-term reserves. AI data centers create new revenue opportunities, but require greater capital expenditure and carry construction risks.

For market participants, the real question is whether companies can turn their resource advantages into sustainable operating capabilities.

BiyaPay's features—including U-to-USD conversion, zero-maker-fee digital asset spot trading, and commission-free U.S. stock trading—can provide users following multiple markets with a choice of capital management and trading tools.

Whether you're researching the BTC market or tracking the long-term changes as mining companies transition to AI, understanding capital flows, corporate finances, and actual trading costs matters more than simply chasing short-term market moves.

This article is for informational purposes only. Please conduct your own analysis and make your own judgments. It is not investment advice. Avoid unnecessary losses!