Is AI eating away at crypto liquidity? Dismantling the underlying logic of 2026 money flowing from Bitcoin to semiconductors

Entering the second half of 2026, crypto investors are repeatedly asking the same question: whether the money that may have flowed into Bitcoin has been sucked away by artificial intelligence once again. This question is not an emotional guess; it is supported by real money flow data. According to data quoted by AMBCrypto, since April, the combined net outflow of US gold and Bitcoin ETFs was about $12 billion, while US semiconductor ETFs attracted more than $20 billion in net inflows during the same period — capital did not leave the market, but simply changed the track.

What really alerted the market was the nature of this round of capital transfers. According to Investing.com's analysis, in the past, crypto retracements were often accompanied by full risk aversion, and almost all assets fell simultaneously; this time, capital was shifting from one highly volatile topic to another. This difference determines how the bottom of Bitcoin will be formed.

Core points

Since April, the combined net outflow of US gold and Bitcoin ETFs is about $12 billion, and the net inflow of semiconductor ETFs has exceeded $20 billion.

The US spot Bitcoin ETF had a net outflow of around $4.5 billion in June, the worst month since its launch.

The five tech giants are expected to spend around $600 billion to $725 billion on AI infrastructure in 2026.

The market is clearly divided over “structural shift” or “cyclical rotation”.

Early signs of a possible return of funds to crypto appeared in early July, and Bitcoin once regained its position at $63,000.

Bitcoin exchange reserves have fallen to their lowest level in about seven years, and long-term holders are attracting funds at an accelerated pace.

Where is the money flowing: a war for marginal dollars

What actually happened

That same sense of caution pushes money away from crypto while pulling it towards AI infrastructure, and the latter's spending scale is hard to ignore. According to Investing.com, the five largest cloud vendors in the US are expected to spend about 725 billion US dollars on AI infrastructure in 2026, of which about 70%, close to 450 billion US dollars, will go directly to chips, servers, networks, and data centers. Nvidia is at the center of this construction wave. Its revenue guidance for the quarter was approximately US$91 billion, an increase of about 85% over the previous year.

Why is this time different

According to Tech Times, Samir Kerbage, chief investment officer of crypto asset management company Hashdex, stated bluntly in a report in early July that the weakness of crypto is more a reflection of investors allocating funds elsewhere, rather than problems with the digital asset ecosystem itself. The logic is straightforward: when a new story is compelling enough, money will flock to it and “cut off” other asset classes for a period of time. Generative AI is just one of the strongest narratives of recent years.

Key Data: The Funding Sector of Crypto and AI

Diverting at the ETF level

According to Tech Times, the combined capital expenditure of US technology companies such as Microsoft, Amazon, Alphabet, and Meta is expected to exceed 650 billion US dollars in 2026, most of which will be invested in AI; SpaceX's June 12 listing has also absorbed a new wave of risk capital. Meanwhile, the net outflow of US spot Bitcoin ETFs in June was about 4.5 billion US dollars, the worst monthly performance since the launch of spot funds, and the cumulative flow turned negative for the first time throughout the year.

The miner's turn explains the problem best

According to Crypto Economy's analysis, the trend of Bitcoin miners is probably the most revealing indicator. Companies that transformed data centers into computing power services for AI customers, such as TeraWulf, recorded positive returns of around 73% in 2026, while mining companies that still focus on pure Bitcoin mining had negative returns over the same period. According to its estimates, by the end of the year, up to 70% of the revenue of listed mining companies may come from AI contracts — this is not so much an opportunistic diversification as a survival response to declining mining profits.

Structural transfer or cyclical rotation

Two tit-for-tat interpretations

The core of the disagreement is whether the funds that left the market will be returned. According to Crypto Economy, Strategy Executive Chairman Michael Saylor characterized it as a “cyclical capital rotation” and believes that in the face of Bitcoin's trillion-dollar market capitalization, the outflow of multi-billion dollars of ETFs is manageable. However, the analysis also reminds that the key to determining whether this round of migration is only cyclical is the nature of exit capital — if it moves towards an AI capital cycle spanning several years, then this money will tend to be locked in for a long time rather than a short-term return.

Why the difference is so important

According to Investing.com, when capital flows from crypto to treasury bonds or money markets, it can return quickly once sentiment improves; but when it flows to a capital cycle supported by multi-year contracts and construction cycles, such as AI infrastructure, the return schedule will be significantly lengthened. This is also the most realistic meaning of “structural” and “cyclical” disputes for investors.

July's reversal signal: whether money is flowing back

Early signs

Entering July, there was a slight change in the balance. According to InvestorIdeas, citing Bitfire Group Research, after a half-year rise, AI assets are facing the double structural pressure of overvalued valuations and overcrowded transactions, and Bitcoin after a deep correction is viewed as a “value range”; last week, spot Bitcoin ETF finally broke the net outflow for many consecutive days, and Bitcoin recovered $63,000. The agency believes that the rotation of funds from AI to crypto is still in its early stages.

Structural conditions to support reflux

According to Tech Times, several supply-side data are rare in history: Bitcoin exchange reserves have fallen to their lowest level in about seven years, long-term holders are attracting funds at the fastest rate in many years, and Bitcoin's volatility is declining cycle by cycle. Hashdex and Carson Wealth Management believe that once AI transactions cool down, macroeconomic policies shift, or regulation progress, crypto can be expected to accept the return of funds; however, Jiaxin also warned that summer is usually the season when Bitcoin institutional purchases are weak.

What it means for investors and potential risks

For investors, a more sensible approach would be to view crypto and AI as two sides of the same risk appetite rather than separate stories unrelated to each other. According to Investing.com, before the flow direction is clearly reversed, trying to accurately decipher the bottom is often in opposition to capital flow data; investors who are willing to allocate encryption can open positions in batches within an established support range and control positions in a way that can withstand continuous fluctuations.

The risks are just as clear. First, if the stickiness of the AI capital cycle exceeds expectations, crypto may not receive marginal incremental funding for a longer period of time. Second, the macro environment is unfriendly — according to Tech Times, Deutsche Bank expects the Federal Reserve to raise interest rates twice in 2026. If growth stocks are under pressure, speculative capital will need new sources, but the direction may not immediately point to crypto. Third, whether the July rebound can continue depends on whether the ETF capital flow actually changes from negative to positive, rather than just a technical recovery.

Exclusive views from the MEXC Crypto Pulse research team

What's really important about this theme is not the conclusion that “AI stole crypto money” itself, but rather that it reveals a deeper fact: Bitcoin is increasingly like a high-beta asset competing with AI for marginal dollars in global risk budgets, rather than an alternative narrative independent of traditional finance. When the funding curves of semiconductor ETFs and Bitcoin ETFs show a mirror relationship, the pricing logic of cryptography is already deeply tied to the entire world of risky assets.

The easiest way to misinterpret the market is to treat “structural” and “cyclical” as an either/or choice. A more accurate understanding is: in the short term, AI and cryptography are indeed competing for the same amount of capital; in the long run, the two may not necessarily have a zero-sum relationship — if AI agentic commerce (agentic commerce) actually takes hold, programmable, borderless financial infrastructure may become just what is needed, and this is where blockchain is used. In other words, every dollar flowing into AI today isn't necessarily an enemy of crypto.

For investors, the next thing to focus on is not the price, but the resonance of three clues: whether ETF capital flows continue to change from negative to positive, whether the AI sector's valuation and congestion levels have peaked and declined, and whether regulatory developments (such as stablecoin legislation and related frameworks) can be implemented. Only when the three improve in the same direction can the return be sustainable; if they contradict each other, the market is more likely to fluctuate over and over again in competition.

From a cross-asset perspective, the implications of this theme are simple: in an environment with limited liquidity, the narrative is a gravitational field for capital. Understanding the flow of funds between AI and cryptography has to some extent become a prerequisite for understanding Bitcoin's next step.

FAQs

Is AI Really Draining Money from the Crypto Market

Judging from capital flow data, this phenomenon did exist in the first half of 2026. According to market data, the combined net outflow of US gold and Bitcoin ETFs since April is about US$12 billion, and the net inflow of semiconductor ETFs during the same period has exceeded US$20 billion. Instead of leaving the market, capital has moved from crypto and gold to the AI and chip sectors. However, early signs of a possible return of funds to crypto appeared in early July, and it remains to be seen if the trend reverses.

Why is the AI sector attracting so much money

The core reason is massive and visible demand. According to market data, the five major US tech giants are expected to spend between $600 billion and $725 billion on AI infrastructure in 2026, with most of it going to chips, servers, and data centers. For investors, a sector with years of contract support and clearly visible demand is clearly more attractive than inflows of shrinking volatile assets, which is the direct reason why capital continues to pour into AI.

Is this round of capital rotation structural or cyclical

The market is clearly divided. One party, represented by Michael Saylor, believes that this is a “cyclical rotation”, and that short-term outflows do not change the long-term value of Bitcoin; the other side believes that since exit capital is moving towards an AI capital cycle spanning several years, the return time will be significantly lengthened, so it is closer to a structural transfer. The key to judging is the nature of exit funding and when AI valuations will peak.

Why are Bitcoin miners turning to AI

Mainly a survival option under profit pressure. According to market analysis, as mining profits decline and network difficulty rises, mining companies that transform data centers into AI to provide computing power services have significantly higher returns. For example, TeraWolf recorded a positive return of about 73% in 2026, while pure Bitcoin mining companies had negative returns. According to estimates, up to 70% of the revenue of listed mining companies may come from AI contracts by the end of the year. This shift itself also forms part of the outflow of capital from pure cryptographic businesses.

Will funds flow back into the crypto market

There is such a possibility, but it has not been confirmed. According to market opinion, the Bitcoin ETF broke the continuous net outflow in early July, and Bitcoin regained its position at $63,000, which is regarded by some institutions as an early sign of capital returning from the overvalued AI sector. Supporting factors include exchange reserves at a low of about seven years, accelerated fund-raising by long-term holders, and potential regulatory developments. However, summer is usually the season when Bitcoin institutional purchases are weak, and whether the return flow can continue depends on whether the capital flow can actually be corrected.

How should ordinary investors deal with this round of competition for capital

The key is to view cryptography and AI as two sides of the same risk appetite to avoid positions that are “nominally scattered and essentially in the same direction”. According to market analysis, before the flow of capital is clearly reversed, trying to accurately cut the bottom is often more effective; a safer approach is to open positions in batches within an established support range and control positions in a way that can withstand continuous fluctuations. At the same time, it closely tracks the three main lines of ETF capital flow, AI sector valuation, and regulatory progress.