Bernstein Picks Z.ai Over MiniMax as It Sees a $100–200 Billion China AI Market

Global brokerage firm Bernstein has officially initiated coverage of China’s publicly listed artificial intelligence laboratories, projecting an optimistic long-term outlook for the sector while drawing sharp distinctions between the leading players. According to a report by Yahoo Finance, Bernstein estimates that China’s domestic AI market could generate between $100 billion and $200 billion in annual revenue, explicitly excluding consumer applications.

This massive enterprise and industrial market potential is driven by the expectation that frontier AI development will remain a paramount technological and strategic priority for Beijing. The firm argues that China’s leading AI developers are uniquely positioned to benefit from easing domestic computing constraints, rapidly improving reasoning capabilities, and the expanding adoption of highly capable, lower-cost open-source models across the enterprise sector.

In its initial coverage, Bernstein established a clear preference among the major Chinese AI labs, launching an “Outperform” rating on Z.ai (2513:HK) while assigning a more cautious “Market-Perform” rating to its rival MiniMax (0100:HK). The brokerage set an ambitious price target of HK$1,350 for Z.ai, citing the company’s deep research pedigree and the strong competitive positioning of its GLM-5.2 model. Bernstein’s analysts, led by Robin Zhu, expect Z.ai’s revenue to exceed current consensus estimates in the coming years, projecting that the company will achieve non-GAAP operating breakeven by approximately 2028, even as it continues to pour massive capital into research and development.

Z.ai’s Competitive Advantage in Coding and Agents

Bernstein’s bullish stance on Z.ai is heavily predicated on the company’s strength in specific, high-value enterprise applications. The firm noted that despite recent share price weakness, which followed the highly publicized launch of Moonshot AI’s competing Kimi K3 model and served to reset market expectations, Z.ai remains exceptionally well-positioned. The brokerage highlighted that Z.ai’s coding-focused AI models continue to rank among the most competitive in China, a critical advantage as software development and IT operations become primary drivers of enterprise AI monetization. Furthermore, Bernstein identified the upcoming release of the GLM-5.3 model and the company’s next-generation pre-trained architectures as significant near-term catalysts that could drive the stock higher.

The preference for Z.ai also reflects a broader strategic assessment of where the most lucrative AI opportunities lie. Bernstein argues that coding and agentic AI applications—areas where Z.ai has demonstrated significant capability—offer a substantially larger and more sustainable long-term revenue opportunity than other modalities. This focus on functional, task-oriented intelligence aligns with the broader trend of Chinese enterprises seeking AI solutions that can directly automate complex workflows, optimize supply chains, and reduce operational costs, rather than simply generating creative content.

The “Make-or-Break” Moment for MiniMax

In contrast to its optimism regarding Z.ai, Bernstein adopted a more measured view of MiniMax, assigning a HK$275 price target alongside its Market-Perform rating. The brokerage characterized the upcoming release of MiniMax’s next-generation M3 Pro model as a critical “make-or-break” moment for the company. This high-stakes environment follows what Bernstein described as the “underwhelming performance” of the previous M3 iteration, which failed to capture the necessary market momentum to justify a higher valuation. While the analysts expect MiniMax’s Annual Recurring Revenue (ARR) growth to improve with the introduction of newer models, they remain skeptical about the company’s core strategic focus.

MiniMax has invested heavily in AI video generation and consumer-facing creative tools, a sector that has garnered significant public attention but presents challenging monetization dynamics. Bernstein contends that while video generation is technologically impressive, it ultimately offers a smaller total addressable market for enterprise revenue compared to the coding and agentic applications prioritized by Z.ai. This assessment highlights a growing divergence in the strategic roadmaps of China’s top AI labs, as companies are forced to choose between pursuing the massive but highly competitive enterprise automation market or attempting to build entirely new consumer ecosystems around generative media.

The Shifting Dynamics of AI Competition

Looking beyond individual company ratings, Bernstein’s coverage initiation provides valuable insight into the evolving nature of the global AI race. The brokerage expects that as more AI tasks become commercially viable, the basis of competition will increasingly shift away from raw reasoning capability and toward cost efficiency and compute availability. This transition strongly favors the Chinese AI ecosystem, which has already demonstrated a remarkable ability to optimize model architectures and drive down inference costs in response to U.S. hardware export controls. As the “intelligence premium” commoditizes, the ability to deliver “good enough” AI at a fraction of the cost will become a decisive competitive advantage.

However, Bernstein maintains that at the very top of the market, research capability and frontier model performance will remain the primary drivers of AI lab valuations. For companies like Z.ai and MiniMax, the challenge is twofold: they must continue to push the boundaries of foundational intelligence to remain relevant, while simultaneously engineering highly efficient, cost-effective deployment mechanisms to capture the massive $100–200 billion enterprise market. As China’s AI sector matures and public market investors begin to demand clearer paths to profitability, the strategic choices made by these labs in the coming months will likely determine the long-term winners of the domestic AI race.