At the beginning of June 2026, three news events occurring within a week formed a clear narrative chain: Australian mining firm IREN secured $3.65 billion in investment-grade GPU financing, fully pivoting to AI cloud computing delivery; AI lab Anthropic quietly filed an S-1 with the SEC, launching its IPO process; and Google’s parent company Alphabet announced its largest-ever $80 billion equity raise, all earmarked for AI infrastructure. These developments—from the computing power suppliers, AI model developers, and infrastructure capital providers—collectively point to one trend: computing power is evolving from an ancillary asset of crypto networks into a core asset class pursued by global capital markets.
Meanwhile, the overall crypto market capitalization contracted by roughly $1.16 trillion over the past six months, while since February 2026, leading AI companies have raised about $140 billion. This capital diversion is visibly reshaping the financial landscape of both industries.
Why Mining Companies Are Abandoning "Mining" for AI Computing Power
On the surface, the immediate reason for Bitcoin miners to pivot is the rising difficulty and shrinking profit margins. In February 2026, Bitcoin mining difficulty surged by 15% to 144.4T, while hashprice (expected revenue per unit of computing power) fell to a multi-year low of just about $23.9 USD/PH/s. With the ongoing squeeze from the halving effect, miners were forced into cash flow defense mode.
However, the deeper driving force is the capital market’s intense pursuit of the AI narrative. Listed mining companies have signed over $70 billion in AI and high-performance computing contracts, with projections that AI revenue could reach 70% of their total by 2026. TeraWulf signed a $12.8 billion deal with Fluidstack, and IREN inked a five-year $9.7 billion AI cloud contract with Microsoft—these orders surpass most miners’ annual revenues during crypto bull markets.
Miners are able to quickly enter the AI space because they possess the most scarce resources for AI computing expansion: power contracts, land, cooling systems, and ready-made data infrastructure. Traditional data centers take three to five years to build, but miners can upgrade computing power in months using these assets. With AI computing power in short supply, miners’ power permits and cooling facilities have become more strategic than chips themselves. This "asset repurposing" logic has shifted miners’ valuation models from "crypto exposure" to "computing infrastructure" pricing.
How IREN’s $3.65 Billion Financing Validates Computing Power as an Asset
IREN’s $3.65 billion investment-grade GPU financing is not just a company milestone—it’s an industry benchmark. The deal includes $2.1 billion in US private placement bonds and $1.55 billion in delayed-draw term loans, with a blended debt cost of 6.00%. Fitch and DBRS awarded A and A(low) ratings, making it the highest-rated investment-grade GPU financing publicly disclosed, and the first GPU financing case in the US private market.
The funds support IREN’s $9.7 billion, five-year AI cloud contract with Microsoft. With Microsoft’s 20% upfront payment, IREN has financed about 96% of its $5.81 billion GPU capital expenditure, at an average cost of just 3.31%. IREN aims to expand AI cloud capacity to 480 MW by the end of 2026. Once fully operational, the Microsoft contract alone is expected to generate $1.94 billion in annual recurring revenue, boosting annualized run-rate revenue from $3.7 billion to $4.4 billion.
This financing structure resembles project finance more than traditional corporate debt: IREN established a risk-isolated subsidiary, using GPU assets and Microsoft contract revenue streams as collateral, shielding creditors from broader crypto business risks. This model—using computing power as collateral for investment-grade financing—creates a replicable path for other crypto infrastructure operators seeking AI transformation, marking a shift as the market begins to view AI computing power as a standalone asset class with predictable income streams.
A notable industry trend: Hut 8 is using its Bitcoin holdings as collateral for transitional capital to expand AI data centers, building a leasing base up to $16.8 billion. Marathon Digital spent $1.5 billion acquiring power assets to ensure energy supply for its AI pivot. This shows miners are deploying diverse financing strategies—including debt, Bitcoin-backed loans, and equity—to fund their transition to AI computing power.
Why the AI IPO Wave Is Draining Crypto Market Liquidity
Capital markets’ preference for AI infrastructure is evident not only in miners’ pivots, but even more so in IPO capital flows. On June 2, 2026, Anthropic quietly filed an S-1 with the SEC, formally starting its IPO process. The company had just completed a $65 billion Series H round, pushing its post-money valuation to $965 billion, with annualized revenue surpassing $47 billion. It’s expected to achieve its first adjusted operating profit in Q2 2026, around $559 million.
This IPO is not just a milestone for Anthropic—it symbolizes the AI industry’s "great migration" into public markets. SpaceX, valued at $2 trillion, is also advancing its IPO, aiming to raise over $75 billion. Cerebras Systems, valued at $56.4 billion, is raising $5.55 billion on Nasdaq. Just these three IPOs could draw over $240 billion from public markets.
Meanwhile, crypto companies face a different reality. Unicorns like Kraken, Ledger, and ConsenSys—with combined IPO targets over $20 billion—have postponed or shelved their plans. Crypto trading volumes have dropped about 75% year-over-year, putting pressure on both public and private crypto operators’ valuations. This "scissor gap" in capital market access—AI firms listing at sky-high valuations while crypto companies wait on the sidelines—is systematically shifting institutional and public investor attention away from crypto.
Investor behavior shows this diversion is self-reinforcing: capital flows to AI as IPO windows open, pushing up AI asset valuations and further widening the appeal gap between sectors. For the crypto market, this means that even if macro conditions improve, capital return will be constrained by the ongoing AI narrative’s magnetic pull.
What Google’s $80 Billion Financing Signals
On June 1, 2026, Alphabet (Google’s parent) announced plans to raise up to $80 billion through equity financing—a record for the company. The package includes $40 billion at-market sales, $30 billion underwritten offerings, and a $10 billion directed share issuance to Berkshire Hathaway. Berkshire has been steadily increasing its Google holdings since Q3 last year, now valued at $15.6 billion; this additional $10 billion investment further solidifies AI infrastructure as a long-term investment theme.
Yet, the $80 billion raise is just the tip of the iceberg. Alphabet expects capital expenditures for fiscal 2026 to reach $180–190 billion, nearly doubling the $91.4 billion spent in 2025. The five major US cloud giants are projected to invest a combined $720 billion in AI in 2026. Nvidia’s CFO recently revealed that H100 rental prices have risen about 20% since the start of the year, with A100 prices up nearly 15%. The strengthening pricing power for computing resources is driving a valuation reset across the industry.
This capital expenditure cycle, led by hyperscale cloud providers and tech giants, marks a shift from venture-driven AI infrastructure to balance-sheet-driven, deterministic expansion. For the crypto market, this means funding competition has moved from venture capital in the primary market to public capital markets. As giants like Google, Microsoft, and Amazon lock in computing power supply with hundred-billion-dollar capital outlays, crypto projects face systemic competition from AI infrastructure assets for market financing appeal.
Computing Power as an Asset: From Crypto Narrative to Industrial Logic
All these events point to a central conclusion: computing power is undergoing a transformation from a supporting production factor in crypto networks to an independent asset class. This shift is supported by three layers of logic.
First is the certainty of demand. AI training and inference needs for computing power are no longer speculative—they’re based on enterprise contracts and long-term service agreements. IREN’s $9.7 billion Microsoft contract, TeraWulf’s $12.8 billion HPC order, and over $70 billion in AI contracts by listed miners—these figures reflect hyperscale cloud providers locking in computing power supply with long-term deals, offering stability far beyond the BTC price cycles that crypto mining revenue depends on.
Second is the scarcity of supply. AI data center construction is constrained by power infrastructure, cooling systems, and land—assets miners already possess. Traditional data centers take three to five years from planning to operation, but miners can upgrade computing power using existing facilities, making them the most efficient suppliers in the AI expansion cycle. With AI computing power in short supply, miners’ power permits and cooling facilities have become more strategic than chips.
Third is the capital logic of financeability. IREN used GPU assets and Microsoft contract cash flows as collateral to secure investment-grade ratings and financing, proving that computing power assets can be independently priced and securitized. Hut 8 leveraged Bitcoin collateral to unlock $16.8 billion in AI data center leases, demonstrating a capital bridge between crypto-native assets and AI computing power. Once the cash flow models for computing power assets gain capital market acceptance, financing costs will continue to fall, accelerating infrastructure expansion.
Structural Responses and Future Evolution in the Crypto Market
Facing this capital rotation driven by AI computing demand, the crypto market’s response can be understood across short-, mid-, and long-term horizons.
In the short term, the pressure from capital outflows is real and visible. The capital absorption effect of large-scale AI IPOs, miners liquidating BTC to invest in AI infrastructure, and the migration of venture capital from crypto to AI are all systematically squeezing crypto market liquidity. However, unlike the collapse after the 2018 ICO bubble, today’s crypto core infrastructure—stablecoins, DeFi protocols, Layer 2 networks—are far more resilient, and fundamentals have not crumbled with capital flows.
In the mid-term, the convergence of crypto and AI is creating new value spaces. Miners’ transition to AI computing providers essentially reallocates resources once dedicated to PoW networks, which is not a shrinkage of the crypto industry but a revaluation of crypto infrastructure assets. As independent valuation systems for computing power assets take shape, crypto assets with real infrastructure will gain new valuation logic. Additionally, cross-sector narratives and funding windows are emerging in areas like AI agent economies, decentralized computing markets, and zkML (zero-knowledge machine learning).
In the long term, the biggest beneficiaries of the AI supercycle may not be single sectors, but the infrastructure layers that efficiently allocate and coordinate computing resources. Crypto networks’ native capabilities in computing coordination, incentive mechanism design, and decentralized resource allocation could find new use cases as computing power becomes a foundational public resource, akin to electricity. The scheduling, pricing, and trading of computing power will increasingly require the trust and efficiency provided by crypto technologies.
For crypto market participants, the core takeaway from this capital rotation is not about "chasing gains" or "turning bearish," but about reassessing computing power’s asset attributes as a fundamental resource. Capital flowing from crypto to AI infrastructure does not mean the crypto narrative loses relevance—it signals that computing power’s value creation is breaking out of single-network boundaries, entering the broader digital economy infrastructure.
Summary
AI computing demand is driving a structural capital rotation from the crypto market to infrastructure. Mining firm IREN’s $3.65 billion investment-grade financing marks a full pivot to AI cloud computing, validating the financeability and valuation logic of computing power as a standalone asset class. Anthropic’s confidential S-1 filing puts it in the trillion-dollar IPO race, with AI companies rapidly listing at sky-high valuations and drawing capital market attention. Google’s Alphabet launches its record $80 billion raise, signaling AI infrastructure’s entry into a phase of deterministic expansion led by industry giants.
The core driver of this capital rotation is computing power’s leap from ancillary production factor in crypto networks to an independent asset class—backed by stable cash flows from enterprise contracts, scarce supply capabilities anchored in power infrastructure, and capital access via asset-backed financing. Together, these form the triple logic of computing power assetization.
For the crypto market, short-term capital outflow pain is real, but the revaluation of computing power assets may open new value spaces for crypto assets with real infrastructure. In this AI-driven capital restructuring, understanding computing power’s asset attributes as a foundational resource is more meaningful for the long term than simply tracking price trends.
FAQ
Q: How does miners’ pivot to AI computing affect crypto network security?
Listed miners have liquidated over 15,000 BTC to finance their transition, with Bitcoin network computing power dropping from a peak of about 1,160 EH/s to around 920 EH/s. However, leading miners retain a portion of BTC reserves and mining capacity, so the network’s security foundation remains intact—the growth momentum in computing power has simply shifted from miners to other participants.
Q: How long will the AI IPO capital absorption effect last?
Major IPO windows for Anthropic, SpaceX, OpenAI, and others are expected from late 2026 to early 2027. If post-listing performance meets expectations, capital markets’ preference for AI may persist for an extended period, making crypto’s capital competition systemic rather than a short-term fluctuation.
Q: What independent growth narratives remain in the crypto market?
RWA (real-world asset) tokenization, stablecoin payment infrastructure, and certain AI-crypto crossover sectors (like decentralized computing markets) continue to attract institutional interest. While these narratives can’t match AI infrastructure in funding scale, they offer structural logic independent of the computing power cycle.
Q: What does computing power assetization mean for ordinary investors?
On Gate’s platform, investors can focus on two asset types: first, crypto projects or assets with real computing infrastructure; second, protocols in the AI-crypto crossover space with computing power scheduling and decentralized computation capabilities. For specific market data, refer to Gate’s real-time price quotes. (As of June 3, 2026, please rely on Gate platform’s live market data.)
