We do not build for today. We build for the chain that outlasts the hype. Yet when a traditional medical company—Jiuan Medical—throws $100 million at three AI startups, the market cheers. But as a core protocol developer who has audited smart contracts for reentrancy, I see something else: a financial bet dressed in innovation, with zero technical substance, and a valuation that defies verification.
Here is the data: Jiuan Medical, flush with cash from COVID-era antigen test sales, invested 750 million RMB (~$104M) for 0.21% of DeepSeek, plus $30 million in Kimi and 100 million RMB in LeapStar. Total exposure: roughly 1.2 billion RMB. For that, they get no board seat, no operational control, and no technical insight. They are passive holders of illiquid private equity in companies that are essentially black boxes.
Context Jiuan Medical is a household medical device manufacturer. Their core competency is thermometers and blood pressure cuffs. They have no AI team, no data centers, no cryptographic infrastructure. Their investment thesis is not technical—it is financial. They are buying a lottery ticket on the AI narrative, hoping to ride the wave of a bull market that has already priced in dreams. The three companies they picked represent the spectrum of Chinese large language model (LLM) players: DeepSeek (open-source leader), Kimi (long-context specialist), and LeapStar (general-purpose platform). But the common thread is centralization. Every API call, every model update, every governance decision is controlled by a closed board. There is no proof of work, no proof of stake, no on-chain transparency.
Core Analysis: The Technical Debt of Private AI From a protocol perspective, this investment is a textbook case of technical debt without the code. Let me break it down using the framework I have applied to DeFi audits and zk-rollup benchmarks.
First, valuation opacity. DeepSeek’s 0.21% stake for 750M RMB implies a valuation of ~357 billion RMB ($50B). For a company whose revenue is not publicly audited, whose costs (GPU clusters, engineering salaries) are hidden, and whose model performance is only self-reported. In blockchain, we have on-chain treasuries, verified token supplies, and open-source code. Here, we have a press release and a promise. The financial market is accepting a valuation that would make most Web3 projects blush—yet there is no smart contract to verify the capitalization.
Second, liquidity illusion. Jiuan Medical now holds shares that cannot be traded on any secondary market. There is no AMM, no order book, no instant exit. The only path to liquidity is an IPO or acquisition, both uncertain. Compare this to a decentralized AI protocol like Bittensor (TAO) or Akash Network (AKT). In those cases, compute and inference are traded on-chain, with transparent fees, active staking, and verifiable work. An investor can sell their token in seconds, with the price set by a global market. Jiuan Medical’s bet is a 5-year lockup with no price discovery.
Third, governance risk. The three startups are directed by a handful of founders and venture capitalists. They can change their business model, pivot technology, or dilute shareholders at will. Jiuan Medical, as a 0.21% holder, has zero influence. In a DAO, governance tokens give proportional voting rights on treasury allocations, protocol upgrades, and fee structures. Here, the only vote is the check they already wrote.
Fourth, security and decentralization. LLMs are increasingly vulnerable to adversarial attacks, data poisoning, and censorship. Jiuan Medical has no ability to audit the model weights, the training data provenance, or the inference pipeline. In a trust-minimized system, we demand cryptographic proofs—zero-knowledge proofs of inference, verifiable randomness, auditable logs. None of that exists in these private silos. The art is the hash; the value is the proof. Without the proof, the value is speculation.
Contrarian Angle: The Hidden Reentrancy The contrarian insight here is not that Jiuan made a bad investment— it is that the entire AI investment thesis is a form of financial reentrancy. The market is reentering the same narrative cycle: hype drives valuation, valuation attracts more capital, and the capital flows to closed systems that cannot be audited. Reentrancy does not scale. It creates a recursive feedback loop where each new round of investment increases the surface area for failure without adding any protocol-level safeguards.
We saw this in the ICO boom of 2017, where projects raised millions on whitepapers alone. Many collapsed because they lacked the infrastructure for trustless execution. Now, the same pattern is repeating with AI. Jiuan Medical’s investment is not a hedge—it is a leveraged bet on the status quo. If any of these startups faces a regulatory crackdown, a security breach, or a talent exodus, the value evaporates with no recourse. The medical company is not buying a protocol; they are buying a promise. And promises are the most expensive form of technical debt.
Takeaway The market is short-sighted. It applauds the splash of a traditional firm “going AI” without asking the hard questions: Where is the code? Where is the proof? Where is the decentralization? We do not build for today. We build for the chain that survives the next bull, the next regulatory purge, the next paradigm shift. Jiuan Medical’s bet will likely yield headlines, not wealth. The real infrastructure—the one with immutable hashes, permissionless access, and verifiable computation—remains underfunded. The question every investor should ask is not “Is AI a good trend?” but “Are we funding a protocol or a function call?” The answer will reveal whose scrutiny truly matters.
