Recently, the market has been dominated by a round of large-scale contract liquidations, with high-leverage long positions related to Bitcoin and Ethereum suffering concentrated liquidations. The scale of long liquidations clearly outweighs that of shorts, but there is a lack of unified and verifiable total statistics on amounts and proportions; it is more a passive de-risking in a high-leverage environment, rather than a turning point signal sufficient to declare a cycle reversal. Parallel to this "noise," there are two quieter but clearer capital flow directions: one end is represented by RWA platforms like Ondo Finance, focusing on on-chain tokenized products related to US stock ETFs and traditional assets like Nvidia. The total value locked (TVL) has been confirmed by multiple data sources to have surpassed $1 billion, and a single source estimates it to be about $1.5 billion around May 18, reflecting structural incremental funds rather than short-term leverage rotation; on the other end, the fund managed by former OpenAI researcher Leopold Aschenbrenner, with a reported scale rising from about $5.5 billion to about $13.67 billion, has significantly increased holdings in Bitcoin mining companies like Core Scientific (CORZ) and Riot Platforms (RIOT), which hold computing power and energy assets. This article will elaborate on the three main lines of "contract liquidation—RWA TVL surge—AI funds betting on mining companies," with the core judgment being: the current violent fluctuations in the high-leverage contract market are more akin to short-term noise, while the migration of funds towards increasing allocations in RWA and tangible computing power assets is quietly redrawing the capital landscape of this round of cryptocurrency and computing power asset bull market.
Panic Liquidation: Long Positions Take Center Stage
The structural characteristics of this volatility clearly point to the concentrated liquidation of leveraged longs. Various market reviews keep repeating the same remark: recent fluctuations have been severe, and high-leverage long positions have been liquidated in large numbers. Public data can only roughly confirm that this round of liquidation is "large in scale with more longs than shorts," but cannot provide precise amounts or ratios; this actually highlights a fact—it's not a single large order or specific institution that is the "culprit," but rather the overall accumulation of long risks in a high-leverage environment has been abruptly exposed.
Regarding contract subjects, Bitcoin and Ethereum-related contracts occupy the main share in this round of forced liquidations, being the core battleground for risk release. Liquidation records are more concentrated near several key price ranges, indicating that the market had piled too much "upward-only" long leverage in these ranges. Once prices deviate from expectations, automatic liquidation mechanisms are triggered continuously, forming a self-amplifying chain reaction. More importantly, it has yet to be confirmed whether there is a specific price trigger point or macro event that can explain the entire round of forced liquidations. This means that this round of panic liquidation is essentially the collapse of the high-leverage structure itself, rather than new external negative information altering the long-term value of the assets.
RWA Funds Flowing In: Ondo TVL Soars
Unlike the passive de-leveraging in the contract market, at almost the same time, the RWA sector provided another clue about capital flow. Cross-referencing multiple data sources shows that the Ondo Finance platform, which focuses on tokenizing traditional financial assets, has recently surpassed a total locked volume (TVL) of $1 billion, ranking among the fastest-growing projects in its category. In a more aggressive estimate from a single source, Ondo's TVL is said to have reached approximately $1.5 billion as of May 18. This figure, due to limited sources, can only be regarded as a reference range rather than a definitive value. However, the leap from "1 billion → approximately 1.5 billion" still indicates a continuous inflow of new funds into this platform in the short term.
Ondo is positioned as a platform for tokenized stocks and RWA, with its core assets not being the highly volatile on-chain native tokens but on-chain mappings of S&P 500 ETFs, US blue-chip stocks like Nvidia and Micron Technology, and other highly liquid assets; on Ondo Global Markets, the tokenized assets issued by Circle, along with these US stock-related subjects, constitute the main source of TVL. Some research institutions have estimated that the overall market for tokenized real-world assets has "exceeded $37.5 billion," but this figure is still under verification. A more prudent interpretation is that against the backdrop of overall expansion in RWA, platforms like Ondo, which represent "on-chain traditional assets," are amplifying their money-raising effect— the rapid rise in TVL reflects the active demand from institutions and individuals to increase exposure to on-chain US stock ETFs and blue-chip stocks rather than being driven purely by price fluctuations.
AI Researchers Shifting to Bet on Bitcoin Mining Companies
In tandem with the RWA sector attracting traditional funds, the fund managed by former OpenAI researcher Leopold Aschenbrenner has also emerged as another fund flow worth closely monitoring. Public information indicates that he currently leads a fund focused on frontier technology and infrastructure assets, with investments spanning AI and underlying computing power-related subjects. A single source disclosed that the fund's scale has recently surged from about $5.5 billion to about $13.67 billion, and the changes in the direction of incremental funds hold more indicative significance than the absolute size itself.
In the latest disclosure of holdings, this fund significantly increased its stakes in Bitcoin mining companies such as Core Scientific (CORZ) and Riot Platforms (RIOT), while also raising its exposure to AI infrastructure companies. There are market rumors suggesting a possible reduction or even shorting of traditional AI chip stocks like NVIDIA and AMD, but currently, relevant details come only from scattered reports, and no authoritative documentation can verify specific positions and operational paths. Therefore, a more reasonable description is: this is an adjustment in positions from "pure chip manufacturers" to "infrastructure holders controlling computing and energy resources." AI large model training and Bitcoin mining both heavily depend on data centers, computing power equipment, and power contracts, among other tangible assets. Mining companies have relatively concentrated bargaining power in these areas, and Aschenbrenner's increased investment is interpreted by many observers as signaling that beyond the narratives of AI and cryptocurrency, the fundamental assets of computing power combined with energy are being seen as a new core allocation focus by frontier technology capital.
Crossroads of Computing Power: The Intersection of AI and Bitcoin
From a resource structure perspective, AI large model training and Bitcoin mining are essentially competing for the same underlying elements: high-density computing power clusters and affordable electricity. Whether training models or processing blocks and recording transactions, significant investments are first made in large-scale data centers, specialized chips, and cooling systems, followed by ongoing long-term electricity costs. This makes the computing power infrastructure and energy contracts the common "foundation" for both tracks. When the macro environment is turbulent, and the positions in contracts based on pure price gambling are forcefully liquidated, funds begin to more finely reallocate risks within this computing power and electricity pool, shifting their focus from a single narrative to "who truly controls the underlying resources."
In this logic, Bitcoin mining companies are viewed by some funds as more "tangible-backed" than pure AI chip stocks: mining companies typically directly hold substantial numbers of mining machines, data centers, and electricity contracts that lock in electricity prices. These can be clearly reflected in their balance sheets while generating observable cash flow and computing power supply capabilities through daily operations, effectively turning their bets on computing power and energy into valuated infrastructure positions. Correspondingly, RWA platforms are tokenizing traditional assets such as S&P 500 ETFs and technology blue chips (including Micron Technology and Nvidia) on-chain, providing institutions with another pathway—to hold visible assets on-chain while indirectly betting on AI and technology cycles. These two types of configurations connect mining company stocks on one end and RWA assets on the other, united by a common preference: at the intersection of cryptocurrency and AI narratives, funds are more willing to pay for targets backed by real physical assets, cash flows, or computing power, viewing computing power as a core production factor that can be priced across markets.
Noisy Liquidations and the Next Scene of Capital Migration
Looking at this round of market movements in a broader context, the concentrated liquidation of high-leverage longs seems more like a periodic and technical release of risk in the contract market. Bitcoin and Ethereum remain at the core of the market pricing system, but the fragility of high-leverage longs has been once again magnified; in contrast, the TVL of platforms like Ondo Finance continues to rise, with multiple data sources confirming it has surpassed $1 billion, and a single source estimating that it approached $1.5 billion by mid-May. Additionally, some research institutions have provided a "to-be-verified" estimate that the overall scale of RWA may have exceeded $37.5 billion, reflecting a medium- to long-term willingness of capital to increase allocations to "on-chain quasi-bond/quasi-stock" assets. The fund managed by former OpenAI researcher Leopold Aschenbrenner significantly raised its scale from about $5.5 billion to about $13.67 billion and increased its holdings in mining companies and AI infrastructure firms like Core Scientific and Riot Platforms, further validating the direction of computing and energy assets being re-evaluated by frontier funds. Short-term emotion-driven liquidations and migrations of capital explored over the year are appearing on the same balance sheet, and investors need to learn to view the former as noise and the latter as trend signals, building their observation framework around several key variables: whether the actual scale of RWA and the market share of platforms like Ondo can continue to rise, whether Bitcoin mining companies' profitability post-halving can match their re-evaluated expectations, and how the evolution of AI computing demand and the global interest rate environment will rewrite the funding allocation order between "leverage contracts—RWA—computing power assets." These dimensions will jointly determine the structural outcomes in the next chapter of the cryptocurrency market.
Join our community to discuss together and become stronger!
Official Telegram community: https://t.me/aicoincn
AiCoin Chinese Twitter: https://x.com/AiCoinzh
OKX welfare group: https://aicoin.com/link/chat?cid=l61eM4owQ
Binance welfare group: https://aicoin.com/link/chat?cid=ynr7d1P6Z
免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。
