MiniMax delivered a landmark earnings report in the first half of 2026: revenue of $117 million, up 283% year-over-year — equivalent to 1.5 times its full-year revenue from last year, achieved in just six months. B2B open platform and enterprise services revenue share surged from 30% to 63%, dramatically overtaking the consumer segment for the first time. The earnings call also revealed that Q2 total revenue grew 81.8% quarter-over-quarter.
The explosion in token consumption is the engine accelerating revenue growth: in July, MiniMax's token usage hit 20 times January's levels, with enterprise users surpassing 2 million — a tenfold increase since the end of last year. This rides a powerful industry tailwind: maturing AI coding tools and the surge of agent products have sent token consumption at large model companies into overdrive.
The revenue structure reversal is industry-wide: as B2B use cases gain real-world validation, open platform and API call revenue is becoming the primary growth engine for LLM companies. Similar signals are accumulating across the sector — token consumption is transforming from a cost item into a revenue driver, validating the "selling picks and shovels" logic of AI services.
Profitability remains a key challenge: MiniMax's H1 R&D expenses reached $297 million, up 138% year-over-year, with adjusted net losses widening. However, a promising scissors gap has emerged — revenue growth is far outpacing R&D growth, suggesting that if this trend holds, the expansion of losses could peak within the next few quarters. The burn rate is accelerating, but revenue is growing faster still.
MiniMax is placing its bets on post-training and inference efficiency: building on increased pre-training investment, intelligent growth depends equally on reinforcement learning and inference optimization. This aligns with the technical trajectory of leading players like DeepSeek and Zhipu AI — approaching performance ceilings with less compute. Refined management of compute resources and API infrastructure has become the new competitive frontier for LLM companies. StarWar Cloud's strategic positioning in the large model API marketplace and compute platform sector is a direct response to this wave of enterprise AI procurement.