When MiniMax and Zhipu AI went public on the Hong Kong Stock Exchange back-to-back in January, the initial verdict from the trading floor seemed definitive. MiniMax, the consumer-focused, overseas-heavy darling from Shanghai, absolutely rocketed. It priced at HK$165 per share and shot up 109% on its very first day, commanding an early valuation that left onlookers breathless. Zhipu AI, by contrast, had a solid but much quieter debut the day before.
Fast forward to mid-2026, and the narrative has been completely flipped on its head.
By late May, Zhipu's market capitalization on the Hong Kong market touched a staggering HK$711.1 billion. MiniMax, despite its early explosive energy and a phenomenal 400% run from its listing price, sat at HK$263.5 billion. It's a massive valuation for a company founded in 2021, but it's a fraction of Zhipu's weight. The initial premium given to MiniMax’s asset-light, international consumer play has eroded in favor of Zhipu’s heavy, labor-intensive enterprise stickiness.
Now, both companies are racing toward a massive dual-listing structure, simultaneously preparing to enter Shanghai's STAR Market. But as they line up for their mainland debuts, a much larger threat looms. The market is about to find out if these multi-billion-dollar valuations can survive a brutal wave of lock-up expirations, or if the tables will turn once again.
The Mirage of Early Consumer Premium
In the initial days of the Hong Kong listings, retail investors fell in love with MiniMax's numbers. It felt easier to understand. The company draws more than 70% of its revenue from overseas markets and boasts over 300 million global users. Its 2025 revenue grew by 159% year-on-year to $79 million. Because it rents almost all of its computing power from third-party cloud services, its fixed-asset spending looked incredibly clean.
It was a textbook consumer-tech story. But the secondary market eventually looked under the hood of both businesses, and the math changed.
Zhipu AI took the hard road. It focused on the domestic B2B enterprise sector, targeting state-owned enterprises, government contracts, and deep local infrastructure. That kind of business requires massive, customized implementation teams and heavy local R&D. It's expensive to scale, and it doesn't look as graceful as an app getting millions of global downloads overnight.
The structural difference showed up clearly in the margins. Zhipu posted a 41% gross margin for 2025, while MiniMax's asset-light model only yielded a 25.4% gross margin. Why? Because when you depend entirely on third-party cloud infrastructure to run consumer-facing models, your compute costs scale almost as fast as your user base. Zhipu’s enterprise clients pay massive contract values with high switching costs, giving the company a predictable, highly sticky revenue foundation that the market ultimately priced at a premium.
The Sudden Rush to Shanghai
The race isn't staying in Hong Kong. MiniMax signed a listing guidance agreement with CITIC Securities to launch its A-share IPO process on the Shanghai Stock Exchange's STAR Market. Not even twenty-four hours later, Zhipu countered by announcing its own formal proposal to raise 15 billion yuan (around $2.2 billion) on the STAR Board, intending to pour 80% of those funds directly into general-purpose foundation model R&D.
This coordinated stampede back to the mainland isn't just about corporate vanity. It's a calculated defensive play.
Right now, secondary-market pricing for Chinese AI is undergoing a radical shift away from hardware speculation toward core software execution. For a long time, mainland investors could only bet on AI by buying hardware and infrastructure stocks like Cambricon. By securing an "A+H" dual listing, MiniMax and Zhipu can soak up massive mainland liquidity while keeping their international investor channels open via Hong Kong.
For MiniMax specifically, a Shanghai listing is a strategic necessity to rebalance its business. Relying on overseas cash flow is great until geopolitical realities or compliance barriers block you from scaling inside your home country. Gaining a mainland listing provides the domestic branding power MiniMax needs to secure the large-scale industrial projects it historically conceded to Zhipu.
The Approaching July Siege
Investors cheering the domestic expansion plans are missing the immediate storm on the horizon. The timing of these dual-listing announcements is deeply tied to a massive liquidity pressure point.
The scale of lifted share restrictions on the Hong Kong stock market is projected to hit an all-time record of HK$1.55 trillion. Both Zhipu and MiniMax are facing massive lock-up expirations. Zhipu will see 5.76% of its restricted shares become tradable, followed by a massive 40% unlock in early 2027. MiniMax faces similar structural pressure.
We’ve already seen how twitchy the market is getting. On June 5, a broader regional market pullback triggered aggressive sell-offs in large-model stocks, with Zhipu falling 9.05% and MiniMax dropping 16.65% in a single session.
When early institutional backers and cornerstone investors sit on massive paper profits, the temptation to exit is high. Unorganized selling from scattered institutional funds can decimate a stock price regardless of fundamental performance. By launching their mainland IPO processes right before the Hong Kong lock-ups expire, both executive teams are attempting to anchor market confidence and signal long-term institutional backing to prevent a post-unlock rout.
Surviving the Four-Horsemen Era
If you’re tracking this space, you can’t look at MiniMax and Zhipu in a vacuum. The competitive framework has completely changed because of what's happening in the primary market.
China's front-rank AI hierarchy has crystallized into four distinct players seeking total valuations that cross $180 billion. Behind Zhipu and DeepSeek, Moonshot AI is currently hunting for up to $2 billion in a fresh funding round that would value the company at $30 billion—a sevenfold increase in just half a year. Meanwhile, StepFun is actively preparing its own Hong Kong IPO paperwork with a targeted valuation of up to $12 billion.
The cash burn required to stay competitive against American giants like OpenAI or Anthropic is hitting unprecedented levels. Technical performance is no longer a differentiator; Moonshot's Kimi K2.6, MiniMax’s new M3 flagship model, and DeepSeek’s open-weights iterations all trade punches on coding and multimodal benchmarks.
The differentiation now lies entirely in commercial endurance and capital structure.
If you are looking to allocate capital or build partnerships within this ecosystem, do not get distracted by daily stock price fluctuations or raw user acquisition metrics. Look directly at gross margins and infrastructure dependencies. MiniMax’s push to integrate faster with domestic computing platforms is an urgent attempt to fix its 25.4% gross margin limitation. Zhipu’s massive 15 billion yuan mainland fundraising goal is an acknowledgment that maintaining an enterprise sales moat requires a permanent war chest.
Watch the trading volume during the upcoming lock-up expirations. If institutional volume holds steady without panic selling, it proves the secondary market views these firms as structural pillars of the domestic tech ecosystem rather than speculative vehicles. The companies that successfully clear the lock-up hurdles and secure their Shanghai dual-listings will hold the financial longevity required to survive the inevitable industry consolidation. All eyes are on the capital flows. Keep your focus there.
