← Back to Library

Zhipu and MiniMax ipo

Jordan Schneider delivers a rare, forensic look inside the Chinese AI boom by dissecting the actual IPO prospectuses of Zhipu and MiniMax, revealing a sector far more pragmatic and state-entangled than Western hype suggests. While the global narrative often fixates on a singular political figure's trade wars, Schneider's analysis bypasses the noise to expose the gritty reality of how Chinese AI firms are actually surviving: through a mix of "light-asset" compute strategies, deep state-sector integration, and a surprising pivot toward digital companionship.

The Model as a Service Gambit

Schneider immediately identifies a strategic divergence in how these two companies are selling themselves to investors. Zhipu is betting heavily on "Model-as-a-Service" (MaaS), a concept the prospectus mentions 96 times. Schneider writes, "MaaS customers buy access to the AI model, rather than products built on top of, or outputs generated by, the model." This framing attempts to rebrand the pure-play AI market into a B2B software model, centering API calls as the revenue engine. The author suggests this is a dual-purpose move: to justify the massive capital expenditure required for research and to gain credibility as a frontier lab in a landscape often viewed with suspicion in the West.

Zhipu and MiniMax ipo

However, the reality of who is buying this "service" tells a different story than a purely commercial tech play. The prospectuses reveal that Zhipu has aggressively courted the public sector, particularly state-owned enterprises (SOEs). Schneider notes that in the first nine months of 2025, nearly 30% of Zhipu's on-premise revenue came from public-sector clients, with the telecommunications giant China Telecom standing out as a major customer. This reliance on government contracts is not accidental; as Schneider points out, "With its Tsinghua roots, Zhipu is a state-fund darling."

"The impulse to constantly assert that the technology itself is the product is an interesting one... Is this a move to persuade investors to support costly R&D? Or to gain credibility as frontier labs in a hostile Western-dominated landscape? Or is it both?"

Critics might argue that framing this as a "hostile landscape" ignores the internal complexities of China's own regulatory environment, where state alignment is often a prerequisite for survival rather than just a defensive posture. The evidence suggests these companies are not just fighting external pressure but are deeply embedded in the domestic state apparatus, a nuance often lost in broader geopolitical summaries.

The Compute Conundrum and Circular Deals

Perhaps the most startling revelation in Schneider's analysis is the fragility of the hardware backbone. Despite the image of massive, sovereign supercomputing clusters, MiniMax admits to a "light-asset" strategy, possessing no meaningful local compute of its own. Schneider observes, "MiniMax has used a diverse range of cloud computing suppliers as its compute demand skyrocketed," listing anonymized suppliers from China and Singapore. This reliance on third-party cloud providers highlights a critical vulnerability: the sector is running on borrowed time and borrowed silicon.

The financial ecosystem supporting these firms is equally intricate, characterized by what Schneider calls "circular deals." Major tech giants like Tencent and Alibaba are not just competitors; they are investors in these startups. "Zhipu's investors include Meituan and Tencent, while two of MiniMax's major pre-IPO investors were subsidiaries of Tencent and Alibaba," Schneider writes. This creates a dynamic where the giants compete with the startups they fund, a structure that blurs the lines of market competition and raises questions about true independence.

"Beijing is careful not to bet too much of the country's economic future on unpredictable developments in AI and watches out for bubble dynamics closely."

This caution from the central government adds a layer of tension. While the companies race to scale, the state remains wary of the "unpredictable developments" and potential bubbles. This mirrors the historical caution seen in other state-directed sectors, such as the initial regulatory tightening on the Hong Kong Stock Exchange listings for tech firms, where the state sought to balance innovation with systemic stability.

The AI Companion Economy

The most human-centric, and perhaps most controversial, part of Schneider's commentary focuses on MiniMax's flagship product: Talkie/Xingye, an AI companion app. The company is banking on a future where emotional entanglement with algorithms is the primary revenue driver. Schneider highlights the company's optimistic, almost naive, framing of this trend: "As model intelligence and memory capabilities continue to advance, we foresee a future 'Her'-style moment where everyone has an AI companion that truly knows them."

The numbers are staggering but the margins are razor-thin. With over 212 million users across 200 countries, the app has massive reach, yet the average customer spent only $5 in the first nine months of 2025. Schneider notes that this figure actually decreased as the company tried to attract cost-conscious international users. The prospectus even acknowledges the risks of "companion misuse" but largely dismisses them in favor of a "sunny attitude" toward a future of digital intimacy.

"MiniMax thinks the future of digital companionship stretches as wide as humanity does, with a new generation of AI-native internet users 'naturally inclined to interact with AI companions.'"

This section draws a fascinating parallel to the gaming industry, specifically referencing the cross-pollination with miHoYo, the maker of Genshin Impact. Just as miHoYo leveraged animation and gaming communities to build a global empire, MiniMax is courting similar demographics for its AI companions. This connection underscores how the "AI boyfriend" business is not a standalone novelty but an evolution of the engagement models perfected in the mobile gaming sector.

Defining the Future of Intelligence

Finally, Schneider contrasts how these two companies define the endgame of artificial intelligence. Zhipu presents a rigid, ideological roadmap of five stages, culminating in "consciousness," while admitting they cannot guarantee reaching those heights. MiniMax, conversely, takes a more flexible, multimodal approach, focusing on "human intellectual tasks" rather than abstract sentience.

Schneider points out that Zhipu is positioning itself as the safety-conscious player, noting it was the only Chinese firm to sign the Frontier AI Safety Commitments in Seoul. Yet, the market reality remains mixed. Zhipu holds a 6.6% market share, trailing behind iFlytek, a state-affiliated player with a 9.4% share, which Schneider attributes to iFlytek's ability to roll out services to its massive existing user base.

"Zhipu doesn't think either open- or closed-source AI will 'win.' As its prospectus describes, '[in] the future, open-source models will play a key role in driving technical innovation and fostering collaboration within the community, while closed-source models will take the lead in commercial applications and enterprise services.'"

This pragmatic split suggests that the Chinese AI sector is not looking for a single "winner" but is instead building a hybrid ecosystem where open innovation fuels commercial exploitation. The argument here is compelling because it moves away from the binary "open vs. closed" debate often seen in the West and instead presents a functional division of labor that suits the specific economic and political constraints of the region.

Bottom Line

Schneider's analysis succeeds in stripping away the geopolitical theater to reveal the operational DNA of China's AI sector: a state-backed, compute-constrained, and increasingly consumer-focused industry that is finding its footing through digital companionship and public-sector contracts. The strongest part of the argument is the exposure of the "light-asset" reality, which challenges the assumption of massive, sovereign hardware independence. The biggest vulnerability, however, lies in the reliance on the "AI companion" revenue stream, where razor-thin margins and regulatory uncertainty could easily derail the growth narrative. Readers should watch closely for how these companies navigate the tension between state mandates and the volatile global market for consumer AI.

Deep Dives

Explore these related deep dives:

  • Hong Kong Stock Exchange

    Both Zhipu and MiniMax chose HKEX for their IPOs, making it the venue for the world's first pure-play AI company listings. Understanding HKEX's unique role as a bridge between Chinese companies and international capital markets provides crucial context for why these AI firms chose this particular exchange.

  • Tsinghua University

    The article mentions Zhipu's 'Tsinghua roots' as key to its success in attracting state funding. Tsinghua is China's premier technical university with deep ties to government and industry, and understanding its role in China's tech ecosystem explains the political economy of Chinese AI development.

  • MiHoYo

    The article notes miHoYo's investment in MiniMax reflects 'cross-pollination between AI companions and other entertainment industries.' Understanding miHoYo's rise with Genshin Impact and its massive global success provides context for why gaming companies are investing in AI companion technology.

Sources

Zhipu and MiniMax ipo

by Jordan Schneider · ChinaTalk · Read full article

This month both Zhipu (also known as Z.ai) and MiniMax made initial public offerings (IPOs) on the Hong Kong Stock Exchange (HKEX), making them the world’s first two pure-play AI companies to go public. Securities laws generally require companies to submit lengthy prospectuses disclosing information relevant for investors before offering shares to the public. In the cases of Zhipu and MiniMax, these are gold mines of information about not only their corporate fundamentals, but also their views on AI, internal culture, and how they fit into the Chinese AI puzzle.

I spent the past few days with these prospectuses and came out of reading with a plethora of observations and questions. Below are some findings and early thoughts, featuring:

Zhipu’s Model as a Service (MaaS) = SaaS + AI?

China’s competitive, ever-changing cloud computing landscape;

AGI is what you want it to be;

And an early look at how good of a business AI boyfriends are…

What is the product?.

Going public requires a company to be very explicit about what they are selling. Here, the two companies diverge the most. Zhipu frames its product strategy around model-as-a-service (MaaS — an acronym which appears 96 times in the prospectus), while MiniMax has an array of diverse products that consumers are already familiar with, from chatbots and video generation platforms to its signature companion app Talkie/Xingye. But MiniMax, self-reportedly, also wants to deliver “technology as products.”

MaaS customers buy access to the AI model, rather than products built on top of, or outputs generated by, the model. In other words, this emphasis on MaaS tries to turn the pure-play AI market into a kind of (mostly B2B) SaaS, with API calls at the center.

The impulse to constantly assert that the technology itself is the product is an interesting one. Both Zhipu and MiniMax are eager to describe themselves as foundation-model companies first, even if they have more specific application products that are clearly profitable (in the case of MiniMax). Is this a move to persuade investors to support costly R&D? Or to gain credibility as frontier labs in a hostile Western-dominated landscape? Or is it both?

Who’s buying from them?.

We learn from Zhipu’s prospectus that it considers the public sector to be a significant source of revenue. It has particularly courted the telecommunications sector, which is heavily dominated by state-owned enterprises (SOEs) in China. Of all the revenue it ...