China’s MiniMax Posts $116.6M in H1 Revenue, Up 283% on Strong B2B Growth

Chinese artificial intelligence enterprise MiniMax announced a dramatic surge in its financial top line for the first half of 2026, driven primarily by a massive expansion within its business-to-business and enterprise service divisions. According to official corporate filings for the six-month period ending June 30, the Shanghai-headquartered firm recorded revenue of $116.6 million, representing a 283 percent increase compared to the corresponding timeframe in the previous year. This rapid scaling built upon the $79 million in revenue generated across the entirety of 2025. Despite this accelerated momentum, the half-year performance represents approximately 32 percent of the $363.77 million full-year revenue benchmark projected by market analysts compiled by Bloomberg, indicating that the developer faces an ambitious upward trajectory to satisfy full-year Wall Street expectations.

The principal catalyst behind this substantial top-line momentum was the firm's enterprise-facing offerings, specifically its Open Platform alongside related corporate AI infrastructure solutions. Revenue derived from this specific segment rose by more than 703 percent year-over-year, climbing from $9.2 million in the first half of 2025 to $73.9 million in the first six months of 2026. Consequently, enterprise solutions evolved to become the core driver of corporate income, contributing 63.4 percent of total revenue compared to 30.3 percent during the prior year period. MiniMax attributed this structural transition to a growing base of enterprise clientele and a higher volume of paying users, while revenue generated from consumer-facing, AI-native applications also doubled over the same timeframe.

Profitability metrics presented a more nuanced financial footprint across the reporting period. Overall net losses narrowed by 11 percent down to $358 million. However, the adjusted net loss expanded by 111.2 percent year-over-year, increasing from roughly $139 million to $293 million as investment in technology development and operational capacity accelerated. On a gross level, performance showed substantial improvement: gross profit expanded more than five-fold from $3.7 million to $20.8 million, while gross margins widened from 12.1 percent to 17.9 percent, demonstrating enhanced cost efficiency relative to direct revenue generation. Prior to the formal earnings release, the company’s Hong Kong-listed shares recorded a modest 1.13 percent gain to close at HK$303.

Established in 2021, MiniMax has achieved prominence within the domestic landscape through its proprietary M-series foundational large language models, the H3 video-generation architecture, and consumer products such as the Hailuo AI video creation suite. Earlier in May, management reported that annual recurring revenue—a operational metric extrapolating monthly subscription figures across a full calendar year—had doubled relative to February figures to reach approximately $300 million.

Following high-profile initial public offerings in January 2026, MiniMax alongside peer Zhipu AI secured listings as the earliest publicly traded foundational model creators from mainland China. Nevertheless, valuation trajectories for both market pioneers experienced steep declines from early post-listing peaks. Shares of MiniMax have receded approximately 77 percent from a peak of HK$1,330 recorded in March. Concurrently, Zhipu AI closed at HK$1,030, representing roughly one-third of its historic high of HK$2,980 achieved in June. In response to a sharp equity contraction in July following the expiration of post-IPO share lock-up restrictions, MiniMax founder and Chief Executive Officer Yan Junjie committed to waiving his executive compensation until the venture realizes artificial general intelligence, seeking to stabilize market sentiment and internal morale.

Both early market entrants operate in an increasingly congested ecosystem marked by aggressive domestic competition from entities like Moonshot AI and DeepSeek, alongside the technical benchmarks set by global laboratories such as OpenAI and Anthropic. Analysts from JP Morgan noted that MiniMax’s recent foundational model, the M3 released in June, had not yet secured a commanding position regarding performance metrics or cost efficiency, leaving the system exposed to high-capability global models on one end and lower-cost alternatives on the other. Consequently, equity analysts emphasize that the upcoming release of the M3.1 model represents a critical milestone to determine whether MiniMax can re-establish technological parity and economic competitiveness within the fast-evolving artificial intelligence market.