On March 10, 2026, Liquid Capital founder Yi Lihua once again put forward a contrarian viewpoint: in his opinion, ETH falling below $2000 means entering a buying zone, and the entire cryptocurrency market will be in a "major bottom phase" by 2026. Almost at the same time, data presented a different picture—Ethereum spot ETFs and ETHA products recorded continuous net outflows, and institutional funds seemingly began to withdraw in an orderly fashion. Funds are voting with their feet, and yet he chose to stack a new chip at this moment: betting that the "startup cycle of AI reconstructing Crypto" has already begun. This article will explore this set of misalignment: on one side is the Ethereum major bottom narrative in the cyclical game; on the other side is the slow construction of a new AI + Crypto pathway from infrastructure to application, attempting to restore the logic and risks of those who still choose to continue betting in a pessimistic market.
Funds withdrawing from Ethereum: ETF single-day net outflows and cooling sentiment
● The differentiated picture of funding data: According to Trader T monitoring and verified by dual sources, on a certain trading day, the Ethereum spot ETF experienced a net outflow of approximately $51.26 million, while another Ethereum-related product, ETHA, saw a net outflow of about $55.08 million (from a single source). Both sets of numbers point in the same direction: institutions are systematically reducing their ETH exposure in the secondary market, and the pace of funds withdrawing has been amplified into an easily recognizable time series. For ordinary investors, these cold outflow numbers often present a more intuitive sense of pressure than any technical indicators.
● Self-reinforcing sentiment and selling pressure: As funds from "compliant money entry" like ETFs continue to record net outflows, the market's pessimistic expectations for the short-term trend of ETH quickly accumulate. The retreating funds are interpreted as a lack of confidence in future returns, further reinforcing the market's imagination of further declines and additional selling pressure. For long-term funds that are trapped, the ETF exits are viewed as a signal that "smart money is leaving first," while waiting speculative funds tend to lower their expectations for buying prices, forming a collective game of seeking new equilibria downward.
● Contrast with Bitcoin's 'banana' sentiment: During the same period, the market generally believes that Bitcoin is emerging from a bullish pattern similar to "banana", with both price and sentiment swinging at high levels, making it the preferred allocation target for institutions and incremental funds. In comparison, Ethereum exhibits a clear funding divergence and reduced attention: on one side is the ongoing amplification of the Bitcoin narrative, while on the other side are the routine outflow data of ETH-related ETFs, the stark contrast between the two further exacerbates the market stereotype of "ETH being relatively weaker."
● Doubts about the major bottom narrative in light of voting with feet: As the digital outflow data accumulates day by day, the market inevitably raises sharp questions: at a time when institutional investors choose to leave, is the so-called "ETH major bottom opportunity" merely self-comfort for contrarian calls? The outflows from the ETF channel are seen by many as real evidence against the major bottom narrative, which places any judgments about the "buying zone" under scrutiny and debate.
Yi Lihua's contrarian bet: Understanding the major bottom from a cyclical perspective
● Public judgment on the buying zone and major bottom year: Yi Lihua made it clear in public that when ETH price falls below $2000, it enters the buying zone, and further combines his understanding of cycles to argue that 2026 will be in a major bottom phase for the cryptocurrency market. This is not merely a short-term call on a single day's price; rather, it aims to provide a rough but actionable "emotional boundary" for long-term funds: breaking below $2000 means the market is already reflecting extreme pessimism in pricing, discounting some long-term growth potential.
● Continued cyclical narrative rather than precise timing: During the argument, he did not attempt to give a precise bottom point, rather, he borrowed more from historical bull-bear rhythms and halving cycles as habitual market narratives. Multiple rounds of Bitcoin bull and bear cycles roughly anchor around halving, first driven by expectations, then by funds pouring in, followed by overextension and pullbacks. This rhythm has also transmitted to varying degrees to Ethereum and a broader range of cryptocurrency assets. What Yi Lihua relies on is this "sense of macro rhythm"—rather than specific projections down to which quarter or price point.
● Viewing ETF outflows as an end-of-cycle release: In his view, the current net outflows from Ethereum ETFs and ETHA do not necessarily indicate a long-term trend reversal but rather resemble the end liquidation of the previous round of excess expectations. When institutional funds reduce holdings in unison, it is more about passive adjustments to the previously high-positioned and structurally imbalanced situations, rather than a fundamental denial of Ethereum's long-term value. In other words, he is more inclined to see this wave of funds withdrawing as an "emotional release at the end of the cycle," rather than the beginning of a completely new long-term downtrend.
● Tension between contrarian betting and mainstream caution: Precisely because most market participants choose to shrink risk exposure in the face of continuous outflow data, Yi Lihua's statement of "starting to buy below $2000" stands out sharply. On one side is a mainstream context emphasizing defense and the safety of cash positions, while on the other side are the voices of a few claiming the major bottom is approaching, allowing for gradual betting on the future. The tension between the two forms the core of the current discourse arena: who is truly confronting greed with rational caution, and who is overly panicking, missing potential reflexive opportunities.
AI reconstructing the Crypto narrative: Building a new gambling table in a lukewarm market
● Core judgment on betting direction: Beyond discussions of price and funding, Yi Lihua focuses more attention on another dimension—entrepreneurship and technology. He clearly states, "The entrepreneurial opportunity for AI to transform the Crypto industry has arrived", which reveals the deeper logic behind his contrarian bullish stance on ETH and the major bottom: rather than engaging in timing based solely on price, it is better to place bets on the next main narrative from a longer-term perspective at the bottom of the cycle; in his view, this main thread is the deep integration of AI and cryptocurrency.
● Imagination space for AI payment and cryptocurrency integration: Taking the frequently mentioned scenario of Bank of AI payment as an example, people envision a payment and asset management system driven by AI that automatically selects optimal on-chain routing, asset combinations, and settlement paths. Users only need to provide objectives and constraints, and the underlying agent can accomplish path planning and execution among multiple chains and assets. Such scenarios are still in the early validation stage, and the research brief does not provide specific adoption data, remaining more at the experimental and conceptual level, but they provide perceivable samples for "AI reconstructing financial interaction methods."
● The real opportunity lies in deep coupling rather than conceptual piling: According to Yi Lihua's understanding, the true opportunity of AI + Crypto does not lie in simply adding the "AI" label after the project name, but in structurally coupling AI with on-chain assets, identities, and settlement layers. For example: AI models do not just read on-chain data but should control multi-signature vaults, manage permissions, and trigger smart contracts safely and verifiably; on-chain identities are not just an address but carry auditable preferences and reputation profiles, which become important inputs for AI-agent decision-making. Such deep integration requires time and infrastructural support, but once established, it could potentially reshape production relations in core aspects like trading, risk control, and settlement.
● Discrepancy between a lukewarm market and developer enthusiasm: The cooling at the price level has not completely extinguished developer enthusiasm. Even against the backdrop of continuous outflows of Ethereum ETFs, more and more developers are exploring AI + Crypto new models—whether it's automated asset management agents, on-chain reasoning markets, or privacy-protecting model calling interfaces. The cold sentiment in the secondary market stands in stark contrast to the reality of new project initiations in the developer circle, adding a layer of realism to the saying “building a new gambling table at the bottom”: funds may be retreating, but talent and code are quietly entering the scene.
The invisible foundation: From Phala to AI Agent's infrastructure experiments
● Phala's collaboration with Intel on trust infrastructure direction: At a deeper infrastructure level, Phala is collaborating with Intel to advance the layout of "trust infrastructure" aimed at AI (according to single-source information). The idea is to use hardware-level security capabilities like trusted execution environments to provide verifiable execution environments for AI reasoning and model calling, binding them to on-chain states. Although specific technical parameters and business terms have not been disclosed, and should not be overextended, this type of collaboration reflects a direction: in the future, AI will not just run on black box servers, but will need to interface with the on-chain world in auditable and demonstrable environments.
● New possibilities brought by infrastructure like CVM: Around this direction, infrastructure like CVM is highly anticipated, attempting to provide a path for AI execution on-chain that balances performance and security. On one hand, it provides efficient computing interfaces for model reasoning and data processing through specialized virtual machines and execution environments; on the other hand, concerning privacy computing, it ensures that user data remains confidential and controllable when processed by AI through encryption technologies and trusted hardware. Such combinations lay the foundational possibilities for "AI executing securely on-chain" and "privacy-friendly smart agents," features that traditional Web2 environments struggle to provide.
● The experiments and limitations of AI Agents like OpenClaw: On the application side, projects like OpenClaw have begun to experiment with allowing agents to achieve autonomous management of on-chain assets—executing transactions, rebalancing, and managing risk automatically according to the model's strategies. Currently, these explorations remain highly experimental, constrained by factors such as model robustness, attack surface complexity, and compliance boundaries, with significant distance from large-scale practicality. However, they at least prove one thing: letting AI directly hold and operate on-chain assets is not mere science fiction but a path that can be progressively validated, although caution is needed to assess its risks and boundaries at this stage.
● Temporal misalignment between technological foundation and price downturn: When we shift our focus from code repositories and technical roadmaps back to price charts, it is easy to feel a strong temporal misalignment: while infrastructure is advancing steadily but slowly, cooperation, iteration, and open-source codes keep accumulating; on the other hand, ETH and related infrastructure tokens in the secondary market are deeply mired in stagnation and disquiet. The curve of technological progress and the curve of price sentiment seem to be living in two different worlds, and this disconnection between foundational construction and market performance tests long-term believers while presenting the most common yet hardest-to-accept reality at the cycle bottom.
The team is the biggest black box: Extra risk premium in the AI track
● People are more crucial than models and parameters: When discussing the AI + Crypto track, Yi Lihua gives a frequent reminder: "The team itself is the biggest risk, not covered by data." In traditional cryptocurrency investments, people are accustomed to using on-chain data, contract calls, funding flows, and other indicators to measure the health of a project; however, in the AI domain, these quantifiable signals often fail to cover the truly critical variables—model capabilities, technical accumulation, ethical boundaries, and execution resilience, which are all deeply embedded in the team itself and hard to present completely on-chain.
● Invisible risks of execution and compliance awareness: Once AI + Crypto projects enter the practical implementation stage, they inevitably intersect with execution, compliance awareness, and safety culture. For example, a protocol claiming to use AI for asset management has real risks not only in contract vulnerabilities but also in whether the team can accurately assess the model's error probabilities, whether enough risk control switches are established, and whether adequate risk disclosures are made to end users. These soft factors are challenging to capture through on-chain addresses or performance curves, yet they decisively determine whether a project can smoothly navigate regulatory and safety storms at critical moments.
● The fate of capital chasing stories and peak explosions: In recent years, capital has repeatedly made the same mistakes while chasing "new narratives": amidst bull market noise and high valuations, overlooking long-term construction and governance design, concentrating resources on the teams with the strongest storytelling capabilities. As a result, at the peak of the cycle, it is often the projects with the most extravagant stories and weakest governance that explode first, triggering a chain reaction. The AI + Crypto narrative could easily replicate this trajectory—when terms like "AI efficiency" and "automatically making passive income" become key fundraising pitches, the genuinely complex technical and governance issues are inadvertently swept under the rug.
● Shifting due diligence focus during major bottom layouts: If we accept the premise of "approaching a major bottom by 2026" and are prepared for long-term investments in the AI + Crypto track, then the focus of due diligence must shift from short-term indicators to teams and incentive structures. Instead of fixating on short-term TVL and token price fluctuations, it is more beneficial to spend time identifying team backgrounds, technical contexts, equity and token distribution logic, and whether there exist governance frameworks capable of constraining power and countering human weaknesses. Only by doing ample homework on these "invisible parameters" can the so-called "major bottom bet" avoid becoming blind gambling in front of a black box.
Betting on Ethereum and AI between panic and imagination
As we reassemble all the clues: on one side there is the continuous net outflows of the Ethereum spot ETF and ETHA, with funding data starkly reminding us that institutions are reducing their ETH exposures; on the other side, Yi Lihua publicly asserts that "ETH below $2000 enters the buying zone" while quietly positioning himself in the new track of AI reconstructing Crypto at the same time. From a price perspective, this is a contrarian bet directly opposing mainstream cautious sentiment; from an entrepreneurial and technical view, it shifts the bets from short-term fluctuations to long-term structural changes.
It can be acknowledged that we are currently likely near the cycle bottom, but no one can provide precise answers regarding the ETH price and timing—research briefs also emphasize that specific subdivisions of price ranges for major bottoms and timing points lack verifiable data foundations. What investors can rely on is only a rough historical rhythm, the current valuation position, and judgments regarding the medium- to long-term evolution of technology and applications. This also means that any assertion that "now is absolutely the bottom" simplifies handling uncertainty, warranting sufficient skepticism and humility.
For readers, a more realistic path of action might be: on one hand, examine the current panic sentiment from a cyclical perspective, distinguishing between structural risks beneath the grand narratives of ETF outflows and price corrections, and those merely exaggerated by overblown emotions; on the other hand, assess AI + Crypto opportunities from an entrepreneurial and industrial perspective, not merely being led by a single highlight but returning to foundational elements like assets, identities, settlement, and privacy, observing whether AI can truly enhance efficiency and create new connections rather than merely generating new speculative noise.
The real game in the upcoming years might no longer be "who sells higher at the top," but rather who can endure longer during the prolonged bottom period—who has enough capital and belief to continue building AI trust infrastructure, components like CVM, and iterating real use cases of AI Agents such as OpenClaw during price downturns until these foundations and upper-layer applications genuinely mesh together. At that point, today’s debates surrounding “ETH major bottom,” “ETF outflows,” and “AI new gambling table” may be rewritten as the prologue to a longer cycle.
Join our community, let's discuss and grow stronger together!
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。
