Summary of MiniMax Earnings Call Company Overview - MiniMax has achieved an Annual Recurring Revenue (ARR) exceeding $150 million [1] - The average daily token consumption has increased to over six times the level of December 2025, with consumption from programming packages exceeding ten times [1] - The number of new registered users has reached more than four times the number in December 2025 [1] - In January, MiniMax released MinimaxAgent 2.0, and by February, professional users had created over 50,000 expert agents [1] - The inference cost of the M series models has decreased by over 50% per million tokens compared to December 2025 [1] Industry Insights - The programming field is expected to see advancements from L4 to L5 level intelligence, transitioning from tools to colleague-level collaboration [2] - In the office sector, the next year is anticipated to replicate the rapid progress seen in the programming field last year [2] - Multi-modal creation is expected to evolve this year into deliverable mid-to-long content, potentially achieving real-time output similar to gaming [2] - The combined demand in these three areas could lead to a potential increase in token volume by one to two orders of magnitude [2] - The market remains one where annual growth significantly exceeds existing stock, with the M3 series expected to unlock a higher ceiling [2] - The current sixfold increase in token consumption is viewed not as a short-term benefit but as the starting point for a long-term, stepwise growth trajectory [2]

Group 1 - The core issue revolves around the recent changes in Trump's tariff policy, where the Supreme Court ruled against the previous tariffs, leading to a new 10% tariff on global goods for 150 days, with potential increases to 15% [3] - The tariffs affect approximately 20% of U.S. imports, impacting countries like China, Mexico, the EU, and Vietnam, which raises concerns about inflation risks rather than alleviating them [3] - The recent economic data shows a significant slowdown, with the U.S. GDP growth for Q4 last year at only 1.4%, much lower than the expected 2.8%, contributing to fears of stagflation [3] Group 2 - OpenAI's adjustment of its total computing power expenditure target to $600 billion by 2030 has been misinterpreted as a cut of $800 billion, highlighting a misunderstanding of the time frame and scope of the figures [4] - AI stocks in Hong Kong have surged, with companies like Zhizhu and MINIMAX seeing significant increases in market value, indicating a shift in investment focus towards companies that can deliver tangible results [4] - The storage chip market is experiencing a seller's market, with SK Hynix reporting a four-week inventory and ongoing price increases due to high demand driven by AI [4] Group 3 - Public funds are preparing to enter the market with over 90 billion yuan, focusing on "technology growth" and "Chinese advantages," indicating a long-term bullish outlook on the tech industry led by AI [5] - The investment strategy for 2026 is expected to be driven by both "risk aversion and growth," with gold serving as a hedge against uncertainty and AI as a growth driver [6] - The storage sector is highlighted as a key indicator of the computing power market, with rising prices and supply shortages signaling potential opportunities for investors [7]