MiniMax doubles ARR to $1.5 billion in 60 days as AI demand surges

MiniMax's annualized recurring revenue doubled to more than $1.5 billion in 60 days, the Chinese AI startup said, as enterprise adoption of its M2 series models accelerated globally.

The company now serves more than 1 million enterprise and developer customers, a fivefold increase from six months ago, and counts roughly 300 million total users across its platform. Daily token consumption for the M2 model family grew sixfold over the same two-month period, while new registrations on the open platform more than quadrupled, according to data disclosed by the company.

The revenue acceleration — from $1.5 billion ARR in February to over $3 billion ARR currently, based on the disclosed doubling — places MiniMax among the fastest-growing AI infrastructure companies by top-line metric. By comparison, OpenAI was generating roughly $3.7 billion in annualized revenue as of early 2025, while Anthropic's ARR was estimated at about $1 billion around the same period, according to public filings and analyst estimates. MiniMax's 60-day doubling cycle far outpaces the typical 6-12 month ARR doubling cadence seen across most enterprise SaaS companies at similar scale.

What's driving the growth

The M2 series, MiniMax's flagship large language model family, has become the primary engine of the company's revenue acceleration. The models support multimodal inputs including text, images, and audio, and are offered through both API access and a consumer-facing application. The sixfold surge in daily token consumption suggests deepening usage among existing customers rather than purely new account sign-ups, a signal of strong product-market fit that investors closely track.

MiniMax competes in a crowded field of Chinese AI model providers that includes Baidu's Ernie, Alibaba's Qwen, and ByteDance's Doubao, as well as global players such as OpenAI and Anthropic. The company has differentiated itself through a focus on multimodal capabilities and cost-efficient inference, though it has not disclosed specific pricing per token or benchmark scores for the M2 series relative to comparable models from competitors.

The investment angle

The ARR milestone carries implications for private market valuations in the AI infrastructure layer. MiniMax was valued at roughly $2.5 billion in its last known funding round in 2024, according to PitchBook data. At a current ARR run rate above $3 billion, that valuation implies a price-to-sales multiple of less than 1x — a steep discount to public cloud companies like Snowflake, which trades at roughly 15x forward sales, or even to OpenAI's reported $300 billion valuation on roughly $12 billion in projected 2025 revenue, a 25x multiple.

The gap suggests either that private market investors have not yet repriced MiniMax to reflect its growth trajectory, or that the market is discounting risks around geopolitical tensions, export controls on advanced chips, and the sustainability of Chinese AI companies' growth relative to US peers. MiniMax, like other Chinese AI labs, relies on Nvidia's H100 and B200 GPUs for training and inference, access to which is constrained by US export restrictions.

The company's next funding round, if it materializes, will test whether private investors assign a multiple closer to US AI peers or maintain a China-specific discount. For now, the 60-day ARR doubling provides the strongest data point yet that MiniMax has achieved the scale and retention metrics that venture investors demand.

This article is for informational purposes only and does not constitute investment advice.