Why MiniMax and Zhipu AI Are So Difficult to Short — and Why July Could Change the Trade

Chinese AI startups MiniMax Group and Knowledge Atlas Technology (Zhipu AI) have become two of the most explosive recent IPOs in Hong Kong. Their shares rallied sharply after listing, even though both companies are still deeply unprofitable and spending heavily on research and computing infrastructure.

For investors trying to bet against them, the puzzle is obvious: why are these stocks so difficult to short? The answer lies less in fundamentals and more in market structure—specifically limited share supply, short‑selling rules, and strong demand tied to AI enthusiasm and index inclusion expectations.

1. Weak Profitability Doesn’t Automatically Create Short Opportunities

Both companies are still burning cash as they scale their AI models.

MiniMax’s financial disclosures show extremely rapid growth but very large losses. Revenue rose from $3.46 million in 2023 to $30.52 million in 2024, while net losses widened to $465.24 million. In the first nine months of 2025, revenue reached about $53.44 million, but the company still recorded a net loss of roughly $512 million.

Zhipu AI (the company behind Knowledge Atlas Technology) shows a similar pattern: strong revenue growth but massive spending on R&D and computing capacity. For example, the company reported a net loss of about 2.36 billion yuan in the first half of 2025, driven largely by rising development costs for AI models.

Despite these losses, the market has rewarded both companies with strong valuations because investors see them as early leaders in China’s generative‑AI ecosystem.

2. The Real Reason They’re Hard to Short: Limited Float

Short sellers need shares to borrow before placing a short trade. In Hong Kong, naked short selling is generally prohibited, meaning investors must have a borrowed or otherwise deliverable share position before selling short.

For newly listed companies, this can be a major obstacle.

At the time of their IPOs, only a relatively small number of shares were available to trade. For example:

• Knowledge Atlas offered about 37.4 million H shares in its Hong Kong IPO.
• Many early investors and cornerstone investors agreed to lock‑up periods, restricting when they can sell shares after the listing.

When the freely tradable float is small and existing holders are unwilling to lend shares, the borrowing market becomes tight. That can lead to:

• limited availability of stock to borrow
• very high borrow costs
• increased risk of short squeezes

Even if a bearish thesis exists, traders may find it difficult or expensive to express it.

3. Strong Post‑IPO Demand Amplifies the Supply Problem

Both companies benefited from intense investor demand around their listings. Their shares surged shortly after debuting, and their market capitalizations quickly reached tens of billions of Hong Kong dollars.

Several forces helped support prices:

• strong investor appetite for AI‑related stocks
• limited free float after the IPO
• speculation about index inclusion

When demand is high but tradable supply is limited, prices can move sharply upward—even if the companies are still far from profitability.

4. Index Inclusion Could Bring Billions in Buying

One of the biggest near‑term catalysts is the possibility that both companies join the Hang Seng Tech Index, which tracks major technology firms listed in Hong Kong.

Analysts at Morgan Stanley estimated that inclusion could trigger $1.25 billion to $1.75 billion in passive inflows, because funds that track the index must buy the new constituents in proportion to their weight.

Index inclusion often pushes prices higher temporarily because:

• passive ETFs must buy the stock
• benchmarked active funds may add positions
• trading volumes increase significantly

However, markets sometimes experience a “buy the rumor, sell the news” pattern if the expected inflows are already priced in.

5. July Lockup Expirations Could Be the First Major Supply Shock

The next major test may come when early investors’ lockups expire.

Lockup expirations allow insiders, venture investors, and cornerstone backers to begin selling shares that were previously restricted. Reports suggest that July will mark a key lockup expiration period for these AI IPOs, potentially releasing additional shares into the market.

That matters for two reasons:

1. More sellable supply – Early investors may take profits after large post‑IPO gains.
2. More lendable shares – A larger float increases the pool of stock available for securities lending, which can make short selling easier.

If both happen simultaneously, the stocks could face downward pressure for the first time since their listings.

6. What Happens Next: Supply vs. Demand

The future path of these stocks will likely depend on the balance between new demand and new supply.

Bullish forces include:

• passive inflows from index inclusion
• continued enthusiasm around AI startups
• expansion of investor access through Hong Kong market programs

Bearish forces include:

• lockup expirations increasing the tradable float
• early investors taking profits
• persistent losses and high spending on AI infrastructure

In other words, the difficulty of shorting these companies today is less about their business fundamentals and more about market mechanics. If the share supply expands in the coming months, the trade dynamics could change quickly.