Changxin hasn’t rung the bell yet—Hyperliquid first trades at a 6x premium

On July 13, Goldman Sachs released a research report that was optimistic about China’s AI industry chain. On the same day, Standard Chartered maintained an Overweight rating for China stocks. Morgan Stanley’s China chief strategist also said that international investors’ interest in Chinese stocks is continuing to heat up.

Also on this day, TradeXYZ bought the ticker of the 科创 50 ETF (KSTR) on Hyperliquid. The next day, it was Changxin Memory (CXMT)’s turn.

In the same week, traditional capital and crypto capital almost simultaneously turned their attention to the same set of targets.

Foreign capital’s sight is readjusting toward China

This focus is not without reason. In the first half of this year, the 科创 50 index’s rise was multiple times that of the CSI 300. By the end of the second quarter, the scale of Northbound funds’ holdings first surpassed 3 trillion yuan. In the semiconductor sector, the market value of holdings surged by nearly 130% quarter-on-quarter—making it the most heavily added-on by foreign capital. The reports from several investment banks, including Goldman Sachs, Standard Chartered, and Morgan Stanley, essentially serve as an endorsement of a trend that has already taken off.

In China’s tech asset landscape, the truly imaginative part has yet to appear on any exchange list. Changxin Memory and Yangtze Memory—two core companies carrying the expectation of domestically replacing memory—have not had an IPO yet. Changxin Memory is the only company in China that can mass-produce DRAM memory chips, and it is advancing high-end storage technology roadmaps such as HBM. Yangtze Memory, meanwhile, has already entered the global first-tier cohort in 3D NAND processes thanks to its self-developed Xtacking architecture. Foreign investors are bullish on China’s memory-chip self-reliance narrative, but they can’t find a tool to directly bet on it. The 科创板 threshold of 500k yuan and QFII quota limits block most overseas capital from outside listed assets—let alone these targets that haven’t even written code yet.

This vacuum has been filled first by the on-chain market.

A pricing venue that doesn’t need approval

Hyperliquid has a mechanism called HIP-3: by staking 500k HYPE (about $25 million–$30 million), you can deploy your own perpetual contract exchange. After the free allocation slots are used up, for each additional underlying target you add, you pay 500 HYPE (about $30k) to bid for that ticker on-chain. TradeXYZ is currently the largest deployer, accounting for more than 90% of the related positions across the network. Previously, it has bought tickers for Nvidia, Tesla, and the S&P 500. This time is the first time it’s turned its sights to A-shares.

It’s not something approved for listing by Changxin Memory or the Shanghai Stock Exchange. TradeXYZ paid to buy the symbol itself, deploying it unilaterally—setting up the venue, taking in the quotes it provides, and managing its own leverage. In essence, it’s like registering a domain: no company consent is needed, and no regulatory approval is required. Anyone with a wallet address and some USDC can place a bet on the valuation of Changxin Memory at 3:00 a.m.

Premium pullback and mapping calibration

When CXMT went live, it shot to a peak of $8.64. Compared with Changxin Memory’s IPO issue price, that was about a 6x premium. It then kept falling, and by July 17 it had retreated to around $6.97—nearly a 20% drawdown. Trading volume also shrank step by step from the highs in the first few hours after the open.

What needs to be noted is that KSTR isn’t pegged to the 科创 50 index itself. Instead, it’s pegged to the KraneShares SSE STAR Market 50 Index ETF—that ETF trades in the US and is regulated by the SEC. That ETF itself tracks the Shanghai Stock Exchange STAR Market 50 Index. In other words, this on-chain contract is effectively two layers removed: first, the ETF maps to the 科创 50 index; then the on-chain contract maps to that ETF—not a direct mapping to the 科创 50 index.

Pushing into the vacuum zone

Changxin Memory is not listed. There is no ETF coverage, and no tool in the US, Hong Kong, or A-share market can express directional views about valuation expectations. KSTR can’t cover it, and the reach of traditional finance can’t extend here. But on Hyperliquid, it already has a perpetual contract quietly trading, with the price searching for that valuation anchor that everyone can see.

This is a pricing vacuum zone that traditional finance can’t enter at all. The on-chain market moved into it quietly.

This isn’t a question of who replaces whom. On-chain pricing won’t, in the short term, replace the pricing functions of the NYSE or the SSE. But it offers a possibility: when traditional finance can’t provide pricing services for certain assets due to various constraints, the on-chain market can.