To The Moon

MiniMax has released its semi-annual financial report for the first half of 2026, marking its inaugural earnings disclosure since listing on the Hong Kong Stock Exchange. The company generated $117 million in revenue during this period, representing a year-over-year surge of 283.1% and already surpassing the $79 million recorded for the entirety of 2025.

The company recorded a net loss of $358 million for the period, narrowing by 11% compared to the previous year. However, adjusted net losses widened to $293 million, an increase of 111.2% year-over-year. Gross profit reached $20.8 million, reflecting a substantial 464.8% year-over-year improvement. Overall, the company's top-line performance exceeded market expectations for the first half of the year.

The standout development during this period was a strategic shift in growth drivers, with the open platform and other AI-based enterprise services segment emerging as the primary revenue engine. This division generated $73.9 million in revenue, surging 703.1% year-over-year and climbing from 30.3% of total revenue in the prior-year period to 63.4%, overtaking AI-native products as the largest income source. The expansion is driven by a surge in reasoning demand from Agent applications and programming workloads, evidenced by increasing paid users, growing enterprise client adoption, higher API call volumes, and rapid Token plan penetration.

Breaking down the business structure, MiniMax derives its revenue from two principal segments. First, the open platform and other AI-based enterprise services—encompassing API calls and Token subscriptions for businesses and developers—generated $73.9 million, accounting for 63.4% of total revenue. Second, AI-native products, which include offerings such as Hailuo AI and MiniMax Code, contributed $42.6 million, representing 36.6% of revenue, up 100.9% year-over-year, supported by enhanced user engagement, greater willingness to pay, and continued commercialization of consumer-facing products.

These two segments display markedly divergent growth trajectories, underscoring that the company's revenue profile is now predominantly enterprise-driven. This structural transition has been building over the past two years. According to the 2025 annual report, AI-native products accounted for 71.4% of revenue in 2024, declining to 67.2% by the end of 2025 when the enterprise segment held just 32.8%. The pace of this shift accelerated significantly in the first half of this year.

During the reporting period, MiniMax released the M3 model, strengthening its programming, Agentic workflow, and professional task capabilities. Shortly after the period ended, the company also launched and open-sourced the H3 video model, reducing enterprise deployment barriers through open weights. Industry-wide, the transition toward B2B services now representing over 60% of revenue aligns with the broader trend of AI model commercialization shifting from consumer subscriptions to enterprise clients, with the latter emerging as the fastest-growing source of API and Token consumption.

This strategic direction does not come without trade-offs. The company's gross margin for the first half stood at 17.9%, an improvement from 12.1% in the prior-year period but notably lower than the 25.4% recorded for the full year 2025. The rapid scaling of API and related services has diluted overall gross margins, reflecting a clear volume-for-price trade-off.

Geographically, revenue from outside mainland China reached $70.8 million, accounting for 60.8% of total revenue with a year-over-year increase of 224.1%. Mainland China contributed $45.7 million, representing 39.2% of revenue, up 433.8% year-over-year. International markets remain the company's revenue foundation, contributing over 60% of total income and demonstrating its global monetization capabilities.

Research and development investment represents a key metric for AI model companies. MiniMax invested $297 million in R&D during the first half, up 138.8% year-over-year. Notably, R&D spending growth trailed revenue growth significantly, which the company attributes to improved R&D efficiency.

While MiniMax continues its phase of strong revenue expansion, supported by the engine transition, overseas-led income, and improving efficiency metrics, the doubling of adjusted losses and gross margins below 20% remain significant challenges. Founded in 2022, the company has gained substantial attention as a prominent industry player since its listing. This semi-annual report reflects not only MiniMax's own business restructuring but also serves as a microcosm of the broader AI model industry.

The sector currently remains in an expansion phase characterized by surging demand yet unresolved profitability questions. Reducing losses and achieving positive earnings continue to be pressing concerns across the industry. For MiniMax specifically, its market capitalization peaked above HK$400 billion following its IPO, briefly surpassing Baidu, before retreating from those highs to around HK$100 billion.

The market's recalibration reflects a fundamental shift in how AI models are evaluated. The industry has moved beyond the era of competing solely on parameter counts toward demonstrating real-world utility. Models must now perform tangible work, assist users effectively, and function as genuine productivity tools. Without such practical capabilities, valuations will remain constrained. This challenge extends beyond MiniMax to all AI model companies.

It's worth noting that despite the substantial decline from peak levels, MiniMax's current market value of approximately HK$100 billion still assumes exponential future revenue growth. The determining factor in how far this AI model dark horse can run will depend less on current revenue or losses and more on the next phase of business structure adjustments—particularly whether gross margins can improve as inference costs decline. Additionally, sustained high growth in API and Token consumption will be critical, though this remains closely tied to broader industry trends. The company's future performance will ultimately be validated by the market.