On February 23, 2026, a Monday that should have been calm, IBM's stock price experienced the most severe single-day decline since October 2000. At the close, it was down 13.2%, with about $40 billion of market value evaporating in just a few hours. The catalyst was not a financial report disaster or regulatory crackdown, but a product announcement: the AI startup Anthropic announced that its Claude Code tool could modernize COBOL programming language running on IBM systems, and COBOL happens to be IBM's highly profitable "moat" business.

Three days later, a similar storyline played out in a completely opposite manner. On February 26, Block, the fintech company founded by Jack Dorsey, announced layoffs of about 4,000 employees, nearly 50% of its workforce, citing AI-driven efficiency improvements. However, the market's reaction was starkly different—Block's stock price surged over 24% in after-hours trading. Dorsey admitted in a letter to shareholders: "I believe that within the next year, most companies will reach the same conclusion and make similar structural adjustments."

Two events, the same driving force—AI; two completely different market responses—one plummeted, the other soared. What exactly happened behind this? The answer may point to a deeper proposition: AI is redefining "what is a valuable asset." For executives of public companies, investors, and decision-makers in traditional businesses, understanding this revaluation logic is no longer just forward-looking strategic thinking but an urgent matter of survival.

1. The Same AI, Different Market Judgments

To understand the contrast between these two events, it is necessary to clarify their respective asset structures.

IBM's drop was, on the surface, a technical threat from the Claude Code tool but was essentially a market re-evaluation of its core asset model. The COBOL programming language, born in the late 1950s, still supports approximately 95% of ATM transactions globally and many core systems in critical fields such as finance, aviation, and government. Anthropic stated in its blog, "Every day, hundreds of billions of lines of COBOL code run in production environments fueling these critical systems. Yet, the number of people who understand COBOL decreases year by year."

For a long time, modernizing COBOL systems has been a complex and costly project, serving as a moat for IBM's profitable business. However, Anthropic claims: "With the power of AI, teams can modernize COBOL codebases in just a few seasons without spending years." The market heard the underlying message: the labor-intensive system maintenance income and services revenue surrounding mainframes that IBM relies on are being eroded by AI technology.

Interestingly, IBM's stock price rebounded 2.68% the following day. Wall Street analysts from firms such as Wedbush and Evercore ISI quickly stepped in to defend the stock, claiming that this drop was "an unfounded overreaction." Their reasoning pointed directly to the core of the issue: enterprise customers are unlikely to abandon their mainframe systems just because a new AI tool can translate legacy code. There exists a massive gulf between code translation and the modernization of deeply integrated hardware-software systems.

IBM itself released a response on the same day, proposing a key argument: the challenge of modernization is not a COBOL language issue but an IBM Z platform issue—code translation almost fails to capture the actual complexity, and the value of the platform lies in decades of software-hardware integration that cannot be migrated through code translation.

Now let’s look at the Block incident. While it also involved massive layoffs driven by AI, the market's verdict was a 24% increase. The key lies in Block's changing asset structure. Since 2024, Block has been restructuring its business model and personnel allocation while heavily investing in AI tools to improve operational efficiency, including developing an in-house tool called Goose.

Block's CFO Amrita Ahuja emphasized when explaining the layoffs: "We are taking bold and decisive action, but we are doing so based on a foundation of strength." This "foundation of strength" is backed by data: for the entire year of 2025, gross profit reached $10.36 billion, a year-on-year increase of 17%. This strong financial performance provides a buffer for the company to push forward with large-scale restructuring at this time.

The market’s interpretation is clear: Block is not passively contracting under AI pressure but is actively optimizing its asset structure—exchanging fewer "human assets" for higher "technology asset" output efficiency. Raising its full-year guidance while cutting 50% of its workforce implies that the value of unit human output is being amplified by AI.

2. In the AI Era, Four Types of Assets Are Being Repriced

These two cases reveal a trend that is unfolding: AI is becoming a "re-pricer" of asset value. Different types of assets exhibit distinctly different value curves under the AI evaluation framework.

The first type is human capital-intensive assets. The value of IBM’s COBOL maintenance teams, traditional analysts, programmers, and other "information processors" is being diluted by AI. When introducing Claude Code, Anthropic mentioned that this tool can identify "risks that would take human analysts months to discover." This does not mean humans are no longer important, but rather, jobs relying on information asymmetry and procedural knowledge are seeing their value compressed by technology.

However, it is important to note that AI is replacing "information processing," not "value creation." Analysts at Futurum Group point out in their research report that successful COBOL modernization projects require multiple dimensions such as business scope definition, technical assessment, data migration planning, behavioral equivalency verification, observability, and organizational change management; code translation is just one part of it. Human capabilities that can navigate complex systems, understand business essence, and make strategic judgments remain scarce.

The second type is data assets, which are becoming value high grounds in the AI era. With the rapid development of generative AI, the value attributes of data are being reshaped. Researchers like Tang have pointed out in a study published in PLOS One that generative AI is changing the ways data are obtained, processed, and utilized. The value of data assets is not only dependent on their inherent quality and relevance but also closely tied to their application scenarios, transformation capabilities, and market demand within the generative AI framework.

This means that the uniqueness, continuity, and governability of data are becoming core value dimensions. A data set may be highly valuable in one scenario but worthless in another. Companies that can provide exclusive, continuous, high-quality data for AI model training are gaining new pricing power.

The third type is algorithm and model assets. The EVMbench launched by OpenAI in partnership with Paradigm, which evaluates AI abilities in detecting, patching, and exploiting vulnerabilities in smart contracts, itself demonstrates that algorithms are becoming quantifiable assets. Model weights, algorithm frameworks, and training methodologies are becoming identifiable, controllable, and monetizable intangible assets.

The fourth type is traditional tangible assets, which are experiencing differentiation. Assets that rely on "information asymmetry" and "human mediation" are facing depreciation pressure, while tangible assets that possess "AI-resilient" characteristics—such as energy facilities, scarce resources, and core infrastructure—maintain relatively stable value. The reason is simple: AI can analyze and optimize the operations of these assets but cannot replace their physical presence and value-bearing functions.

3. From "Asset Revaluation" to "AI Immunity"

Based on the analysis above, companies need a systematic framework to judge whether their assets appreciate or depreciate in the AI era. The RWA Research Institute proposed the "AI Immunity" asset identification framework, which includes three core characteristics.

The first characteristic is non-codability. This refers to value elements that cannot be completely learned or replicated by AI. COBOL code itself can be translated by AI, but the transaction processing capacity built at the chip level in Z series mainframes running COBOL systems, quantum-safe encryption, and eight nines reliability are things that AI tools cannot replicate. Research from Futurum Group notes that "code translation does not capture actual complexity, and platform value comes from decades of software-hardware integration." Similarly, offline scenario control rights, implicit industry knowledge, and complex relational networks—these non-"codable" elements form the first layer of immune barriers for assets.

The second characteristic is data moat. Does the company possess exclusive, continuous, governable data assets? Is it merely using publicly available data or able to generate data that others cannot access? CITIC Bank has started to explore using large models to assess data asset value, attempting "data asset incorporation." The underlying logic is: in the AI era, data is not just a raw material for production but an asset itself. However, not all data come with a moat—public internet data will quickly be "digested" by AI models, while companies that have exclusive data sources can achieve a premium under the AI valuation framework.

The third characteristic is AI-enabled resilience. Can the asset itself be enhanced by AI rather than replaced? This is key to distinguishing IBM-style shocks from Block-style transformations. IBM's core business—maintaining COBOL legacy systems—is subject to "replacement" by AI, while Block's business model—payments, financial services—can be "empowered" by AI. In fact, IBM has also developed the watsonx Code Assistant for Z, a dedicated tool that allows customers to securely refactor and modernize legacy code directly on the platform while maintaining enterprise-level security. When assets can form synergies with AI rather than confront it, their value increases.

Conversely, AI-vulnerable assets also exhibit three characteristics: reliance on "information processing" as core value, potential for substitution by standardized processes, and lack of data generation and accumulation capabilities. By contrasting with these three characteristics, companies can conduct "stress tests" on their asset portfolios.

4. New Opportunities for RWA: What Assets Are Worth Tokenizing?

Extending the above framework to the field of RWA (Real World Asset Tokenization) leads to a clear conclusion: RWA is not "any asset can be on-chain," but rather filtering out hard assets that can traverse the AI cycle amidst the wave of AI revaluation.

In March 2026, the total value of on-chain RWA had surpassed $25 billion, nearly quadrupling year over year. However, the RWA industry white paper released by the Hong Kong Web 3.0 Standardization Association in August 2025 clearly stated: "The proposition that everything can be RWA is a fallacy." Assets that successfully achieve large-scale implementation need to meet three thresholds: value stability, legal clarity of rights, and off-chain data verifiability.

Combining the "AI Immunity" framework, we can further specify that assets worth tokenizing are those whose value is stable in the AI revaluation context.

The first type is tangible assets with "AI Immunity" characteristics. These include energy assets, infrastructure, and scarce resources. The value of such assets does not rely on information processing but is derived from physical existence and practical utility. New energy RWAs mentioned in the white paper (such as charging stations, photovoltaic assets), and GPU computing power assets fall into this category. Among them, GPU computing power assets, bolstered by the "rigid demand" of the AI industry and reliable "digital lineage," are becoming ideal anchor assets for RWA.

The second type is programmable data assets. Assets with exclusive data sources that can monetize automatically via smart contracts possess both the "data moat" and "AI-enabled resilience." The white paper categories data along with intellectual property, carbon credits, etc., as intangible assets. However, it is crucial to be aware that not all data can become assets—only those that can continuously generate, are verifiable, and can establish rights have the foundation for tokenization.

The third type is hybrid assets, which combine non-"codable" physical control with "programmable" digital rights. For example, the ownership of commercial real estate can be tokenized, but the actual operation, maintenance, and leasing of the property—these offline scenario control rights—remain in the hands of professional institutions. This "physical + digital" dual structure takes advantage of blockchain's liquidity benefits while retaining the offline value anchor of "AI Immunity."

Conversely, two types of assets require caution regarding tokenization in the AI era. One type is financial assets that heavily rely on human mediation, which are likely to have their value compressed by AI; the other type is standardized assets without data moats, which lack bargaining power under the AI valuation framework.

5. Action Guidelines: From Cognition to Decision

IBM's $40 billion evaporation signals an era—those assets reliant on information asymmetry and human accumulation are being repriced by AI. Block’s counter-trend surge is another clarion call for a new era—companies that can embrace AI and optimize asset structures are receiving a market revaluation.

For decision-makers in public companies and traditional enterprises, this is not only about technological anxiety, but a fundamental reconstruction of asset value systems. CEOs need to answer an unavoidable question: How much is my asset portfolio worth in the eyes of AI?

Based on the analysis in this article, three actionable suggestions can be made.

First, immediately initiate an "AI stress test" of assets. Evaluate each core business unit against the three characteristics of the "AI Immunity" framework—non-codability, data moat, AI-enabled resilience. Identify which businesses are most susceptible to value shrinkage under AI pressure and which might experience the amplification effect from AI.

Second, establish a dynamic asset portfolio management mechanism. Under the backdrop of AI revaluation, asset allocation is no longer a static strategy of "buy and hold." Companies need to consciously increase the proportion of "AI Immunity" assets while formulating transformation or divestiture plans for those AI-vulnerable assets. This responsibility lies not just with the finance department but also requires collaboration among strategic, technical, and business departments.

Third, re-examine RWA strategy. Before considering asset tokenization, first screen underlying assets with the "AI Immunity" framework. The core value of RWA is not "going on-chain" itself, but obtaining better liquidity and pricing efficiency for quality assets through tokenization. If the underlying asset itself is depreciating in the AI era, then tokenization merely accelerates the loss of value.

Finally, it should be explicitly stated that according to the document No. 42 jointly issued by eight departments in China, any form of token issuance and token trading is strictly prohibited within mainland China. The RWA tokenization discussed in this article only refers to asset digitization practices under compliant frameworks outside of China. Companies exploring related businesses must strictly adhere to the regulatory red line of "strictly prohibited domestically, registered abroad."

When AI begins pricing assets, the only sense of security comes from those things that AI cannot price—not code, not data, but the human ability to judge value itself.

(This article is based on publicly available information and data, with sources including Nasdaq, Tencent News, Futurum Group, PLOS One, 21st Century Business Herald, and other authoritative media and research institutions. The views expressed in this article do not constitute any investment advice.)

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。