On July 22, 2024, a seemingly ordinary Monday, the Hong Kong stock market opened with a chill that only the brave would call a correction. Minimax, the AI darling that had been propped up by promises of a "new paradigm" in machine intelligence, plunged over 9% in a single session. Zhipu, the academic heavyweight backed by Tsinghua's pedigree, followed with a 3% decline. The crypto-native AI token market, still nursing wounds from the 2022 meltdown, reacted almost in sympathy: FET dropped 4%, AGIX slid 2.5%, and the entire AI narrative on-chain began to tremble. This was not a coincidence. It was the beginning of a narrative collapse that both traditional and decentralized markets had been ignoring for months.
From the ashes of 2017 to the fluidity of DeFi, I have tracked every major narrative shift in this space. When a market loses its story, it loses its soul. And right now, the AI story — whether in equity or token form — is losing its soul faster than a bear market rug pull.
Context: The Grand AI Narrative Cycle
To understand why Minimax and Zhipu—two companies that had raised billions in combined valuation—fell so hard on a Tuesday, we must first examine the narrative architecture that propped them up. In 2023, the AI industry experienced a "Sputnik moment" with the release of ChatGPT. This sparked a global mania that transcended asset classes: tech stocks surged, crypto AI tokens (like FET, AGIX, and OCEAN) exploded by 10x or more, and venture capital poured into any startup with "large model" in its pitch deck. The narrative was simple: AI is the new internet, and everyone must own a piece.
But as any seasoned crypto historian knows, narratives collapse when they become too universal. By mid-2024, the AI story had become a commodity. The same venture capitalists who funded AI startups were now whispering about the "institutional friction" of high compute costs and the "societal stratification" that arises when only a few companies control the models. The narrative began to fray at the edges.
Minimax and Zhipu were poster children for this fraying. Minimax, with its ambitious linear-attention architecture, promised to outperform Transformer models at a fraction of the cost. Zhipu, with its GLM series, leveraged academic credibility to win government and enterprise contracts. But the market, both in Hong Kong and in crypto, started to ask a dangerous question: "Where is the revenue?"
This is the same question that killed countless ICOs in 2018, the same question that undid the DeFi yield farms of 2020 when incentives dried up. Narratives can survive a lack of technical perfection, but they cannot survive a lack of narrative sustenance. The AI narrative needed new stories — new models, new killer apps, new reasons to believe. By July 2024, the story well was dry.
Core: The Mechanism of Narrative Decay in AI Markets
Let me break down exactly how the narrative decay unfolded. Based on my experience analyzing on-chain liquidity flows during the 2020 DeFi Summer, I have identified three phases of narrative collapse: Disillusionment, Recalibration, and Reallocation. Minimax and Zhipu's stock dip is a perfect case study of phase one and two.
Phase 1: Disillusionment (June–July 2024)
The first signal came in June, when AI model benchmarks began showing diminishing returns. GPT-4o, released by OpenAI, still dominated the leaderboards. Anthropic's Claude 3.5 was close behind. Meanwhile, the Chinese models — Minimax, Zhipu, Baidu's ERNIE, and ByteDance's Doubao — were locked in a brutal price war. API pricing dropped by 80% in six months. The narrative of "AI as a premium service" collapsed. Instead, the market began to see AI as a commodity, a utility with razor-thin margins.
Token prices in the crypto AI sector mirror this. Fetch.ai, which had rallied on the promise of autonomous agents, saw its trading volume drop 70% month-over-month. The narrative that "AI needs blockchain for trust" was failing to gain real traction, because the actual use cases (decentralized compute, data markets) remained niche.
Phase 2: Recalibration (July 22, 2024)
On July 22, the Hong Kong market panicked. But why this date? Three catalysts aligned: 1. Macro: A stronger-than-expected US dollar and hawkish Fed remarks triggered a risk-off sentiment in global equities. Growth stocks with no earnings were the first to be cut. 2. Company-specific: Rumors circulated that Minimax's cash burn rate had exceeded internal projections, and that its latest model (MiniMax-VL2) failed to improve over GPT-4o in third-party evaluations. Zhipu faced whispers of a talent drain to ByteDance. 3. Narrative saturation: On Twitter and crypto Telegram groups, the AI hype had turned to fatigue. The same accounts that had shilled $FET ten months earlier were now laughing at "AI bros."
The result was a waterfall. Stop-losses triggered. Margin calls hit. And the narrative that once held everything together— "AI is the future" — suddenly sounded hollow.
Phase 3: Reallocation (Ongoing)
Money does not sit still. When a narrative dies, capital flows to the next story. In crypto, that story has been Real World Assets (RWA) and DePIN (Decentralized Physical Infrastructure). In equities, it's been "defensive consumer goods" and "value stocks." The AI money is being harvested and replanted elsewhere. This is the same pattern we saw when the 2017 ICO bubble burst: capital moved to DeFi in 2020, then to NFTs in 2021, then to nothing in 2022. The narrative cycle is brutal but predictable.
Contrarian: The Bull Case That No One Is Talking About
Let me offer a contrarian view, because narrative hunters must always look for the blind spot. The collapse of AI hype might actually be the best thing that could happen for long-term, genuine innovation.
Here's the logic: When capital is abundant, it fuels mediocrity. Every company with a whitepaper or a research paper can raise money. But when capital dries up, only the strongest survive. This is precisely how crypto survived the 2018–2019 bear market: the weak ICOs died, and the resilient ones (like Binance, Chainlink, and Uniswap) emerged stronger.
For Minimax and Zhipu, the stock dip forces them to focus on product–market fit rather than valuation games. For crypto AI tokens, the price drop cleanses the market of speculative froth and leaves behind only projects with genuine demand (e.g., decentralized compute networks like Io.net, or AI-driven prediction markets like Gnosis).
Moreover, the institutional adoption of AI is still in its infancy. The 2024 ETF era for Bitcoin was a massive validation of crypto; something similar could happen for AI if a major index fund adds AI stocks or tokens. But for now, the market is punishing greed, not technology.
I also note an interesting divergence: while Minimax dropped 9%, the underlying sentiment in the crypto AI community remained less drastic. Why? Because crypto AI tokens are often tied to specific protocols with token sinks (staking, governance, fee burn), which create a floor that equity does not have. Zhipu's stock has no burn mechanism; FET tokens do. This token design difference may explain why the crypto AI narrative has a layer of resilience that pure equity lacks.
Takeaway: How to Hunt the Next Narrative
As I write this, the narrative needle is twitching. The capital leaving AI is not being destroyed; it is being reallocated. Where will it go? My analysis of on-chain flows and equity fund rotations suggests three candidates: 1. RWA tokenization (real estate, bonds, commodities) – The most boring, but the most sustainable narrative for 2025. 2. DePIN (decentralized GPS, wireless, energy) – Physical infrastructure narratives are gaining traction because they are harder to fork than software. 3. Privacy and ZK (zero-knowledge) proofs – As AI becomes commoditized, the ability to verify facts without revealing data becomes a premium.
The lesson from Minimax and Zhipu is clear: narrative dominance is fleeting. In 2017, I learned that even the most brilliant cryptographic ideas can be killed by a broken story. In 2024, the same is true for AI. The question is not whether AI will change the world—it will. The question is whether the market will survive the narrative winter to build the future.
From the ashes of 2017 to the fluidity of DeFi, I have seen enough cycles to know that the time to buy is when the narrative is dead and everyone is laughing. The price may not recover for months, but the story will. It always does.