The silence between the code and the chaos is a familiar place for me. I’ve spent years mapping it in crypto, where narratives shift faster than block times. Now, I see the same pattern forming in the AI market. A recent CITIC Securities report on tech stock adjustments caught my attention—not for its stock picks, but for its narrative framework. The report argues that AI stock pricing has moved from ‘macro-rate-driven’ to ‘industry-fundamental-driven,’ with three variables: commercialization pace, computing power conversion, and model gap evolution. And it flags one wildcard: anti-distillation. This is the same narrative cycle I’ve observed in DeFi, L2s, and now AI. The market is no longer paying for imagination; it’s paying for execution. The narrative is the only immutable ledger.
Context: The Great Pricing Reset
CITIC’s report is a rare institutional document that cuts through the noise. It states that AI stocks have entered a new phase: the ‘expectation verification period.’ In 2023, the market priced AI on breakthroughs—GPT-4, multi-modal, AGI whispers. In 2024, the anchor shifted to commercialisation—revenue growth, customer retention, gross margins. The report’s core insight is that macro factors like US Treasury yields are no longer the primary driver. Instead, internal industry variables dictate pricing. This mirrors the crypto bear market of 2022-2023, where projects that survived were those with real users and revenue, not just whitepapers. I’ve seen this before. During the 2020 DeFi Summer, I wrote ‘Liquidity as Ethics,’ predicting that yield farming without ethical frameworks would collapse. The same pattern holds: narrative without execution is a dead token.
Core: The Three Variables and the Anti-Distillation Blind Spot
The report identifies three pricing variables. First, commercialization pace: whether AI companies can convert tech hype into recurring revenue. Second, computing power conversion: whether GPU advantages translate into market share and pricing power. Third, model gap evolution: whether the gap between frontier models widens or narrows. The wildcard is anti-distillation—technical measures to prevent competitors from using a model’s outputs to train new models. This is the most critical variable, and the report treats it as a mere footnote. Based on my experience auditing tokenomics for AI-crypto protocols, I can tell you that anti-distillation is not just a technical problem; it’s a narrative and economic trap.
Let me explain. The report’s logic is: if anti-distillation works, model gaps become permanent, computing power advantages solidify, and the market concentrates into an oligopoly. This sounds like a bullish scenario for incumbents (OpenAI, Anthropic, Google). But in practice, anti-distillation is a double-edged sword. In crypto, we’ve seen projects try to protect their code with licenses or obfuscation. It never works. The community forks, the code is decompiled, and the narrative shifts to ‘open source vs. closed source.’ The same will happen in AI. The moment a model tries to lock its outputs, a decentralized alternative emerges. I’ve been tracking the AI-crypto convergence since 2024—I published a report, ‘Agents Without Borders,’ forecasting a 300% increase in AI-crypto integration by 2027. The key insight: trustless execution requires verifiable compute, not proprietary model weights. Anti-distillation is a battle against entropy. The market will eventually price in the failure of such attempts.
The Commercialization Trap
The report argues that commercialisation is the first variable. It points out that top AI firms still rely on ‘new customer acquisition’ rather than ‘deep monetisation of existing customers.’ The unit economics are unproven. This is exactly where crypto’s lesson applies. In 2021, DeFi protocols had high TVL but low retention. The market eventually punished them. The same is happening now to AI companies. The market’s patience window is narrowing. If the next 2-3 quarters don’t show improving LTV/CAC ratios, the valuation framework will shift from PS (price-to-sales) to PE (price-to-earnings). That’s a systemic de-rating. I’ve seen this script before. In the 2022 bear market, I retreated to a cabin in Jiuzhaigou to process the narrative breakdown. The lesson: execution is the only story that survives the winter.
Computing Power as a Moat, Not a Castle
The report treats computing power as a structural advantage. It says that access to GPUs creates a ‘generational advantage’ that new entrants cannot match. This is true, but only partially. In crypto, we call this ‘hashrate centralisation.’ Bitcoin’s mining landscape is dominated by a few pools, yet the network remains secure because the economic incentives align. For AI, the same logic applies: computing power is a moat, but it is not a castle. The report misses the nuance that efficiency matters more than raw power. I’ve worked with projects using MoE architectures and speculative decoding to reduce inference costs by 40%. The market will reward those who squeeze more output per watt, not just those who hoard H100s. The contrarian angle: the computing power advantage is already being eroded by algorithmic innovation, just as ASIC resistance was once a narrative in crypto.
Contrarian: The Anti-Distillation Paradox
Here is the counter-intuitive angle that the report overlooks. Anti-distillation, if implemented, will accelerate the very fragmentation it aims to prevent. Why? Because it forces smaller players to either copy openly or innovate from scratch. History shows that when barriers to imitation rise, the incentive to collaborate increases. In crypto, the best security comes from transparency. The same will happen in AI. The most valuable models will be those that embrace verifiable provenance—not just output watermarks, but on-chain audit trails of training data. This is where blockchain meets AI. I call it the ‘Trustless Training Economy.’ The report’s focus on anti-distillation as a ‘largest potential variable’ is correct, but it fails to see that the real battle is not between models, but between architectures of trust. The narrative will shift from ‘my model is better’ to ‘my model is honest.’
Takeaway: The Next Narrative Cycle
The market is at a narrative inflection point. The CITIC report provides a solid framework, but it lacks the crypto-native perspective that the next cycle will be about decentralised intelligence. The three variables—commercialisation, computing power, model gap—will be redefined by on-chain verification. The next takeaway is not about which AI stock to buy, but about which narrative architecture to trust. In the wild west, stories are the only compass. The AI market’s compass is now pointing toward execution. But the deeper truth hides in the bear market’s quiet shadows: the real value is not in the model, but in the proof. The silence between the code and the chaos is where the next narrative will be born.
Truth hides in the bear market’s quiet shadows. The next cycle will be about verifiable AI, not just powerful AI. I hunt for the story that the data cannot speak. And right now, the data is whispering: anti-distillation is a trap, and the market will soon have to price it as such.
