Is Trillion-Dollar Capital Flowing from Crypto to AI? The Restructuring Logic Behind Ongoing BTC ETF Outflows and AI Infrastructure’s “Siphon” Effect

On June 29, 2026, according to Gate market data, Bitcoin was reported at $59,612, down over 50% from its all-time high of $126,080 in October 2025. The total cryptocurrency market cap fell to approximately $2.07 trillion. Meanwhile, NVIDIA (NVDA) shares remained near $193, with a market cap of about $4.66 trillion; AI cloud infrastructure company CoreWeave (CRWV) saw its market cap climb from approximately $35.7 billion in 2025 to $64.4 billion. The shifting balance between these two sets of numbers points to a thesis being repeatedly validated by the market: a significant portion of institutional capital is flowing from crypto assets to AI infrastructure. But how much exactly? Is this flow a temporary narrative rotation or a structural capital reset? A quantitative deduction can be made from verifiable data dimensions.

Bitcoin in June 2026: A Picture Painted by Data

Let's first look at the condition of the crypto market itself. On June 16, Bitcoin was still at $67,203, then slid all the way down, hitting a low of $58,188 on June 25. On the morning of June 29, it briefly fell to $58,888. The Fear and Greed Index dropped to 12, in the "extreme fear" zone, having been mired in that territory for several consecutive weeks.

ETF flows are the most direct window into institutional behavior. For the week ending June 26, US spot Bitcoin ETFs saw net outflows of $1.79 billion, the second-highest single-week outflow since listing. On June 26 alone, daily outflows reached $445 million, with BlackRock's IBIT accounting for $444.5 million, the largest single-day withdrawal for that fund since its launch in January 2024. Bitcoin ETFs have recorded net outflows for 13 consecutive days, cumulatively losing $4.33 billion. On a monthly basis, net outflows in May totaled $2.43 billion, and June has seen another outflow of over $2.2 billion so far. Two consecutive months of bleeding have pulled the full-year 2026 fund flow into negative territory.

Since its inception, BlackRock's IBIT has attracted cumulative inflows of $60.77 billion, but its net assets now stand at only $44.42 billion. This implies an average paper loss of about 40% for IBIT investors. The shift from "rushing in" to "lining up to exit" for institutional funds has occurred in less than half a year.

The Scale of AI Capital Expenditure: Understanding the Magnitude of Fund Diversion

To quantify the scale of fund diversion, one first needs to understand the magnitude of capital absorption by the AI industry.

According to the Bank for International Settlements (BIS) Annual Economic Report for June 2026, the world's five largest hyperscale cloud providers are expected to invest a total of over $1 trillion in AI-related projects from 2025 to the end of 2026. Other reports predict that by the end of 2026, cumulative AI-related capital expenditure by global hyperscale cloud providers and emerging AI cloud platforms will reach approximately $2 trillion. Morgan Stanley estimates that current annual AI investment worldwide has increased to about $1 trillion to $1.5 trillion.

What does the magnitude of these numbers mean? The annual AI-related capital expenditure alone is equivalent to 50% to 72% of the entire cryptocurrency market cap. When two asset classes have such a huge disparity in size, one-way flow from the smaller market to the larger market is almost inevitable—this is not just "diversion" but a "siphon."

Barclays data provides another piece of evidence: the investment-grade bond issuance of hyperscale cloud providers in 2026 has already exceeded $200 billion, and the full year could rise to around $240 billion. This scale even surpasses the annual bond issuance of major US banks. AI infrastructure financing is not only massive in size but is also consuming credit lines and liquidity in traditional financial markets that might otherwise have been allocated to other assets.

NVIDIA and CoreWeave: Micro Verification of Two Capital "Reservoirs"

Breaking down macro data into specific targets allows for a clearer view of capital flows.

NVIDIA (NVDA) is the most direct beneficiary of AI capital expenditure. In Q4 2025, revenue reached $68.1 billion, up 20% sequentially and 73% year-over-year, with gross margins holding at 75%. Its market cap is about $4.66 trillion. Although the stock price has pulled back from recent highs, institutions' long-term expectations for AI demand have not fundamentally wavered.

CoreWeave (CRWV) is an even more symbolic case. This AI cloud infrastructure provider reported Q1 2026 revenue of $20.7k, up 112% year-over-year. As of the end of Q1, remaining performance obligations (backlog) stood at $98.8 billion, with 36% expected to be recognized as revenue in the next 24 months. The company has already secured 90% of its target of reaching $30 billion in annual recurring revenue by the end of 2027. Analysts expect CoreWeave's median 2026 revenue to be $12.59 billion, surging to $46.6k in 2027.

CoreWeave's growth trajectory stands in stark contrast to crypto mining companies. According to Robbie Mitchnick, Head of Digital Assets at BlackRock, some Bitcoin mining companies have shifted from pure mining operations to providing power and infrastructure support for AI data centers, signing contracts worth billions of dollars, including Core Scientific, IREN, HIVE Digital, and TeraWulf. The capital these miners previously used to buy mining rigs and expand mining farms is being reallocated to AI data center construction—a direct industry-level reflection of fund diversion.

Quantitative Deduction: How Much Capital Has Shifted from Crypto to AI?

Based on available data, deductions can be made from three dimensions:

Dimension One: ETF Outflows. Spot Bitcoin ETFs saw cumulative net outflows of approximately $4.66 billion between May and June (May $2.43B + June so far $2.23B). During the same period, BlackRock's IBIT net assets fell from about $60.7 billion to $44.4 billion, a contraction of $16.3 billion. While ETF outflows are not entirely equivalent to "shifting to AI," the timing shows a high correlation between the scale of ETF redemptions in Q2 2026 and capital inflows into AI-related assets.

Dimension Two: Bernstein's Overall Estimate. A Bernstein report indicates that in 2026, Bitcoin treasury companies and ETFs together attracted inflows of about $12 billion, a sharp decline from $60 billion in 2025. A significant portion of the roughly $48 billion reduction in incremental capital is judged to have flowed to AI-related assets.

Dimension Three: The "Crowding Out Effect" of AI Financing. Over $200 billion in AI-related bond issuance and cumulative capital expenditure exceeding $1 trillion are absorbing substantial liquidity from global risk asset allocations. An HTX Research report points out that from late 2024 to mid-2026, a large portion of new dollar liquidity was absorbed by the AI supply chain: stock investors buying AI equity, bond investors purchasing AI-related credit assets, private equity funds participating in data center financing, and banks and non-bank institutions lending to tech giants and data center projects.

Integrating the above dimensions, a reasonable estimate range is $40 billion to $80 billion in institutional capital shifting from crypto asset allocations to AI infrastructure-related assets between 2025 and 2026. This estimate is based on cross-validation of three layers of evidence: ETF outflow data (around $10 billion level), incremental capital reduction (around $48 billion level), and the macro crowding-out effect of AI financing (hundreds of billions level). Note that this estimate carries significant uncertainty—capital flows are not a zero-sum game; part of the capital may come from new liquidity rather than existing transfers, and the crypto market's own cyclical factors are simultaneously at play.

Consensus Judgments from Industry Leaders

This capital diversion trend has been confirmed by several key industry figures.

Binance founder Changpeng Zhao (CZ), in a June 2026 interview, explicitly stated that the sharp decline in the crypto market in the first half of 2026 had no single cause. Geopolitical tensions, investors shifting funds to AI, and the typical four-year crypto cycle together led to the sustained decline of Bitcoin and other crypto assets. Bitcoin fell about 50% from its all-time high of $126,080 in October 2025. CZ also noted that emerging industries like AI are absorbing "hot money" from the crypto sector, but this could be a positive factor in the long term.

BlackRock's Head of Digital Assets, Robbie Mitchnick, was more direct. He said that AI-related assets have been continuously siphoning market capital and attention, with a large amount of capital flowing into AI themes since 2025, and AI stocks have outperformed Bitcoin in 2026. Mitchnick believes the current market is more focused on which assets can directly participate in AI infrastructure, computing power, and application building, while assets less connected to AI face stricter capital screening.

Bernstein analysts attributed the main reason for the slowdown to retail investors chasing AI-related opportunities. These judgments, coming from a crypto exchange founder, top asset management executives, and leading Wall Street research institutions, are highly consistent in direction.

The Four-Year Cycle Is Still at Play

Capital diversion is not the only factor pressuring the crypto market. Galaxy Research shows that Bitcoin's four-year cycle remains intact, although the influence of the "halving rule" is diminishing. Historical analogies indicate that the base-case bottom of the current drawdown is between $40,000 and $46,000, roughly occurring between now and Q4 2026.

This means the market decline in 2026 is a combined effect of "natural cyclical downturn" and "AI capital siphoning." Even without AI interference, Bitcoin faced cyclical pullback pressure in the 12-18 month window after the April 2024 halving. The rise of AI has added an extra layer of liquidity extraction on top of the existing cycle.

The Fracture of BTC-Nasdaq Correlation

A notable derivative phenomenon is the drastic change in the correlation between Bitcoin and the Nasdaq. In April 2026, the 30-day rolling correlation coefficient between Bitcoin and the Nasdaq 100 hit a record high of 0.96. However, by early June, this coefficient had dropped to near zero. From extreme correlation to near decoupling took less than two months.

This fracture suggests a deeper structural change: Bitcoin is breaking away from the "tech stock beta" narrative, but its new pricing anchor has not yet been established. As momentum traders shift their attention from the crypto market to AI and semiconductor sectors, Bitcoin has lost its linkage support with the Nasdaq while failing to establish an independent upward logic. This state of "narrative vacuum" may have more long-term impact than mere price declines.

Conclusion

Based on the above analysis, several relatively certain conclusions can be drawn:

First, capital flowing from the crypto market to AI infrastructure is a real and quantifiable fact. The reasonable estimate is in the tens of billions of dollars, a magnitude sufficient to have a significant impact on the marginal pricing of the crypto market.

Second, this flow is not merely a narrative rotation but has structural industry-level support. AI is experiencing a capital spending frenzy similar to the dot-com bubble era around 2000, while the crypto industry happens to be in the natural downturn phase of its four-year cycle—the overlap in timing amplifies the downside for the crypto market.

Third, the sustainability of this trend depends on the return on AI capital expenditure. The BIS warned in its June 2026 report that if AI investment returns fall short of expectations, it could trigger a rapid tightening of financing. If signs of an AI bubble burst appear, capital may flow back into the crypto market—but this takes time, and the prerequisite is that the crypto industry itself can offer a more attractive risk-return profile than currently.

For market participants, understanding the structural changes in capital flows is more valuable than focusing solely on price points. When trillion-dollar capital expenditures are reshaping the allocation landscape of global risk assets, crypto assets need to find new footing in both narrative and fundamentals simultaneously.

FAQ

Q1: What are the core reasons for Bitcoin's price decline in 2026?

Bitcoin fell about 50% from its all-time high of $126,080 in October 2025 due to a combination of factors: AI-related assets absorbing significant institutional capital, the Fed maintaining a hawkish interest rate stance (June FOMC held at 3.50-3.75%), spot Bitcoin ETFs experiencing 13 consecutive days of outflows totaling $4.33 billion, and the crypto market's own four-year cyclical downturn.

Q2: How much capital has AI "siphoned" from the crypto market?

Although there is no precise official statistic, based on ETF outflow data (about $4.66 billion in May-June), Bernstein's estimate of reduced incremental capital (about $48 billion), and the macro crowding-out effect of AI's trillion-dollar capital expenditure, a reasonable estimate range is $40 billion to $80 billion. This estimate carries significant uncertainty, but the directional trend of capital diversion is clear.

Q3: Is Bitcoin's four-year cycle still valid in 2026?

Galaxy Research indicates that Bitcoin's four-year cycle remains intact, but the marginal influence of the "halving rule" is diminishing. The base-case bottom of the current drawdown is expected between $40,000 and $46,000, roughly occurring in Q4 2026. AI capital diversion adds an extra layer of downward pressure on top of the existing cycle.

Q4: Why did the correlation between Bitcoin and the Nasdaq drop from 0.96 to near zero?

In April 2026, the correlation coefficient hit a record 0.96, but by early June it fell to near zero. The core reason is that momentum traders shifted attention from the crypto market to AI and semiconductor sectors. Bitcoin lost its "tech stock beta" linkage support, but a new pricing anchor has not yet been established.

Q5: Is the long-term impact of AI capital diversion on the crypto industry positive or negative?

CZ believes that AI absorbing "hot money" is a positive factor in the long run—it forces the crypto industry to move away from reliance on speculative liquidity and focus on building stronger fundamentals. But in the short term, crypto assets need to find new footholds in both narrative and fundamentals to regain institutional capital attention.