What happened
MiniMax announced unaudited financial results for the six months ended June 30, 2026. The company reported total revenue of $116.6 million, up 283.1% year over year, with Open Platform and other AI-based enterprise services accounting for $73.9 million, or 63.4% of total revenue.
MiniMax Group Inc. said its total revenue increased from $30.4 million in the first half of 2025 to $116.6 million in the first half of 2026, a 283.1% year-over-year increase. The company described the figures as unaudited. First-half revenue exceeded the $79.0 million that MiniMax reported for all of 2025, although the source does not provide a comparable full-year 2026 forecast or say what portion of the increase came from new versus existing customers. The largest reported change was in Open Platform and other AI-based enterprise services. Revenue in that category rose from $9.2 million to $73.9 million, an increase of 703.1%, and its share of total revenue increased from 30.3% to 63.4%. MiniMax attributed the growth to more paying individual and enterprise users, higher API call volumes and adoption of its Token Plan. The source does not provide the number of paying users, the number of enterprise customers, average contract values, retention rates or a breakdown between platform, API and other enterprise-service revenue.
Revenue from AI-native products increased 100.9%, from $21.2 million to $42.6 million. MiniMax attributed that increase to higher user engagement, greater willingness to pay and continued adoption and monetization of Hailuo AI and other products. During the reporting period, the company said it released its MiniMax M3 model, with improvements in coding, agentic workflows and professional work. Shortly after the reporting period, it said it released MiniMax H3 with open weights for video generation and broader enterprise and developer deployment. The company also reported gross profit of $20.8 million, up 464.8% from $3.7 million, while gross margin rose from 12.1% to 17.9%. Research and development spending increased 138.8% to $296.9 million, primarily because of cloud services used for training and model iteration. Selling and distribution expenses fell 17.9% to $27.0 million, while administrative expenses more than doubled to $30.2 million.
Adjusted net loss reached $293.0 million, compared with $138.7 million a year earlier. MiniMax reported $1.323 billion in cash and related financial holdings at June 30, up from $1.050 billion at the end of 2025.
Why it matters
The results indicate that enterprise and developer demand, rather than only consumer AI products, supplied most of MiniMax’s reported first-half revenue. They also show the tension between rapid AI revenue growth and the substantial cost of training, upgrading and serving models.
The revenue mix matters because it places production use by enterprises and developers at the center of MiniMax’s business, according to the company’s own accounting. Open Platform and other enterprise services supplied nearly two-thirds of first-half revenue, compared with less than one-third a year earlier. That is a meaningful shift in the company’s reported commercial profile: demand for API calls, model inference and enterprise access appears to have grown faster than revenue from AI-native products aimed at individual users.
The figures also illustrate the cost structure of competing in advanced AI. MiniMax reported a 283.1% increase in revenue alongside a 138.8% increase in research and development spending, which the company presents as evidence of improved R&D efficiency. But R&D spending of $296.9 million was still far above reported revenue, and the company recorded a $358.0 million accounting loss for the period. The adjusted loss was smaller after excluding share-based payments, fair-value losses on financial liabilities and listing expenses, but it remained substantially larger than the prior-year adjusted loss. Gross-margin improvement to 17.9% suggests that MiniMax says its infrastructure became more efficient as usage expanded. The company’s chief executive also said token consumption had reached 20 times its January level by July, though the source does not provide the underlying token count, a methodology or a comparison with revenue growth. Higher usage can support scale, but it can also increase computing and energy costs. The results therefore do not establish that MiniMax has reached sustainable profitability or that every additional unit of model usage is economically attractive.
The filing is significant for users and customers because it connects product strategy to financial pressure. MiniMax says it is investing in foundation models, multimodal capabilities, inference efficiency and products such as MiniMax Code, while making models available through an Open Platform. Yet the source gives no independent performance tests, customer case studies, service-level data or detailed pricing analysis. Its claims about global reach, affordability and model capability remain company statements rather than independently established findings in this source.
What to watch next
The key unresolved questions are whether the enterprise growth can persist, how much revenue comes from a small number of customers or products, and when MiniMax can narrow its losses. The company did not disclose detailed customer concentration, unit economics, model-level revenue, or profitability forecasts in the source.
The next important evidence will be whether Open Platform revenue continues to grow in subsequent reporting periods and whether that growth is diversified across enterprise customers, developers and individual users. MiniMax attributes the increase partly to API call volumes and Token Plan adoption, but does not disclose usage-based revenue, customer concentration, churn, contract duration or the share of revenue generated outside its home market. Those figures would help distinguish broad adoption from rapid growth concentrated in a limited number of accounts.
Investors and customers should also watch the relationship between inference efficiency, model capability and spending. MiniMax says infrastructure efficiency improved and that inference efficiency enables more post-training, experimentation and deployment. The source does not quantify cost per token, gross profit by product, compute capacity, energy use or the effect of the M3 and H3 releases on margins. It is therefore not possible from this announcement to determine whether model improvements are lowering the cost of service, increasing demand enough to offset costs, or both.
The company’s cash position provides context for its continuing investment, but not a complete picture of financial durability. MiniMax reported $1.323 billion in cash and related holdings at June 30, while also reporting a $293.0 million adjusted net loss for the half-year. The source does not provide a cash-flow statement, a runway estimate, debt maturities, future capital plans or a profitability target. It also does not establish the commercial availability, adoption or financial contribution of H3, which was released after the reporting period.
Finally, the claims about model performance, user reach and enterprise adoption warrant continued verification. MiniMax says its products and models serve hundreds of millions of users and more than one million enterprises and developers across multiple geographic markets, but the announcement does not define those counts or explain how active, paying and trial users are measured. Future disclosures that clarify these definitions, report independent evaluations or show customer-level outcomes would make the company’s growth claims easier to assess.
