Morgan Stanley evaluates MiniMax as a "globally top-tier foundational model scarce asset," with the core logic for its high valuation being that "technology determines the ceiling, and globalization determines the valuation."

Morgan Stanley initiates coverage of MiniMax with an “Overweight” rating and a target price of HKD 930, positioning it as a “Global Leader in AI Foundation Models.” In this report, Morgan Stanley’s true bet is not on short-term profits but on two main themes: whether the model capabilities are among the top in the world, and whether the revenue structure has resilience for global expansion.

According to Wind Trading Desk, Morgan Stanley analyst Gary Yu, who covers the company, believes MiniMax has entered the global SOTA model camp, with well-developed multimodal capabilities and highly scalable commercialization pathways. Based on this, the company’s revenue is expected to grow from $75 million in 2025 to $700 million in 2027, achieving a 9-10x increase in two years. Once technological capabilities establish a generational advantage, the revenue curve will show a “stepwise” leap.

The report’s explanation for the high valuation is straightforward: This is an asset where “technology determines revenue ceiling, and globalization determines valuation framework.” If model performance continues to reach the top tier globally, its ceiling is limited by the global TAM; if the revenue structure is primarily overseas, the valuation anchor naturally aligns with international comparable companies.

Technological capability is the starting point for valuation

Morgan Stanley attributes MiniMax’s core competitiveness to three factors: continuous iteration ability, multimodal layout, and cost efficiency.

In independent benchmarks, MiniMax-M2 ranked fifth globally on the LLM leaderboard at release; the latest flagship model MiniMax-M2.5 ranked sixth and was fourth among open-source models. As of mid-February 2026, M2.5 ranked first in token usage on OpenRouter, reaching 1.97 trillion tokens, with a market share of 58.8% in programming scenarios.

These data indicate that the model has entered high-frequency real-world application scenarios, rather than remaining in laboratory benchmarks.

More importantly, cost structure. The company uses MoE architecture and Linear Attention mechanisms, with inference phase Model Flop Utilization exceeding 75%, higher than the industry average of 40%-50%. Inference efficiency directly impacts API pricing tiers and gross margin elasticity, which determines whether profit margins can improve in scale expansion.

Morgan Stanley expects the company’s gross margin to rise from 12% in 2024 to 32% in 2027. However, operating losses will still expand during the same period, with non-IFRS operating losses estimated at about $484 million in 2027. This is not a profit inflection point logic but a path of “first expanding technology and scale, then looking at profits.”

Technological leadership does not guarantee profitability but determines the revenue ceiling.

Revenue structure determines growth slope

MiniMax’s business model is not driven by a single product but runs along three parallel lines:

• 2C: Agent and companion products Talkie/Xingye

• 2P: Hailuo AI, MiniMax Audio

• 2B: Open Platform API

In the first nine months of 2025, the company’s MAU grew from 3.1 million in 2023 to 27.6 million, with 1.77 million paying users. The revenue structure is becoming more diversified, with the Open Platform’s share continuously increasing.

Morgan Stanley forecasts that the Open Platform’s revenue share will rise from 29% in 2024 to 40% in 2027, with a three-year CAGR exceeding 200%. After breakthroughs in model capabilities, enterprise API demand is more likely to see “leapfrog” growth.

The report emphasizes an industry characteristic: growth in foundational model companies is often triggered by key generational model upgrades rather than smooth climbs. OpenAI’s ChatGPT 3.5, Anthropic’s Claude 3.5 Sonnet have all caused revenue jumps after model upgrades.

Whether MiniMax will replicate this rhythm depends on the next-generation model launched in mid-2026.

Globalization is a prerequisite for valuation

Morgan Stanley highlights MiniMax’s “Born Global” path.

Overseas revenue share has increased from 19% in 2023 to 73% in the first nine months of 2025. Regional distribution is: Asia-Pacific 61%, Americas 24%, EMEA 15%.

In the context of the global foundational model market expected to grow from $10.7 billion in 2024 to $206.5 billion in 2029 (CAGR 80.7%), the company’s current global market share is only about 0.3%. Slight increases in share will reveal revenue elasticity.

More importantly, the valuation framework. If revenue mainly comes from overseas markets and customers are primarily API and subscription users, the valuation logic aligns more with international AI comparables rather than traditional Chinese software companies.

This is also the core basis for Morgan Stanley’s 54x 2027 P/S valuation.

Valuation divergence centers on “next-generation models”

Three scenario assumptions are clearly delineated:

• Base case: 2027 revenue of $700 million, corresponding to a 54x P/S, with a target price of HKD 930.

• Optimistic case: 2027 revenue of $1 billion, target price HKD 1240.

• Pessimistic case: 2027 revenue of $400 million, target price HKD 300.

The only variable determining valuation differences is whether the next-generation model launched in mid-2026 reaches or surpasses global SOTA levels.

Risks are similarly focused: GPU supply and geopolitical restrictions, resource gaps compared to OpenAI and other large-scale players, pricing pressures from model commercialization, and ongoing cash burn.

This is a pricing of “technological realization capability”

Morgan Stanley does not shy away from reality: currently, no pure AI foundational model company has achieved stable profitability. MiniMax is expected to have an average monthly cash burn of about $27.9 million in 2025, with limited profit visibility.

But the core judgment of the report is—competition in the foundational model industry is not about marketing but about generational breakthroughs. Technological capability determines the revenue ceiling, and the global market sets the valuation anchor.

If model upgrades lead to nonlinear revenue expansion, the current valuation is just an early discounting of future scale; if the model fails to establish a top-tier global position, valuation will shrink just as quickly.

This is a bet on the pace of technological realization. Morgan Stanley chooses to stand on the side of “scarce assets of top-tier global foundational models.”

The above highlights are from Wind Trading Desk.

For more detailed analysis, real-time insights, and frontline research, please join 【**Wind Trading Desk - Annual Membership**】

![](https://img-cdn.gateio.im/social/moments-848acfcbd8-31de2dfa37-8b7abd-d8d215)

Risk warning and disclaimer

Market has risks, investment should be cautious. This article does not constitute personal investment advice and does not consider individual users’ specific investment goals, financial situations, or needs. Users should consider whether any opinions, views, or conclusions herein are suitable for their particular circumstances. Investment is at your own risk.