Two of China's largest generative AI startups are set to list in Hong Kong this week, testing investor interest in an industry that has rapidly risen in the post-ChatGPT era. However, as some of the earliest domestic challengers to leaders like OpenAI, KNOWLEDGE ATLAS and MiniMax have chosen starkly different paths to their public market debut. KNOWLEDGE ATLAS and MiniMax will begin trading on January 8th and 9th, respectively, serving as a litmus test for whether investors believe China's emerging AI industry can compete globally. Both companies exemplify China's pragmatic and efficient approach to AI development, requiring significantly less capital, fewer chips, and leaner teams compared to firms like OpenAI Inc. and Anthropic PBC. Yet their trajectories diverge considerably. KNOWLEDGE ATLAS, established six years ago by professors from one of China's most prestigious universities, primarily serves and is supported by domestic institutional clients. MiniMax is a younger, more conventional startup that aims to surpass DeepSeek and OpenAI globally with a consumer-facing chatbot, backed by Alibaba Group Holding Ltd. and an Abu Dhabi sovereign wealth fund. MiniMax operates like a Silicon Valley startup, deriving most of its revenue from 210 million users of its native AI applications through subscriptions and advertising. In contrast, the state-backed KNOWLEDGE ATLAS has found favor with institutional clients, boasting over 8,000 corporate customers. "These listings reflect a policy-driven capital cycle as China accelerates investment in its domestic semiconductor and AI supply chains," said Aadil Ebrahim, Head of Equities at Singapore-based Klay Group. "The significant stock price surges of GPU and AI-related listed companies in the recent A-share and Hong Kong markets clearly indicate strong investor interest." In 2024, KNOWLEDGE ATLAS achieved revenue of 312.4 million yuan (approximately $44.7 million), surpassing MiniMax's $30.5 million. These figures are a fraction of OpenAI's estimated $13 billion annualized revenue and Anthropic's projected $9 billion, though KNOWLEDGE ATLAS and MiniMax's pre-IPO valuations are also much lower, hovering around $4 billion each. KNOWLEDGE ATLAS has assembled top AI scientists from Tsinghua University and is hailed as a pioneer in China's AI sector. Led by the prominent scholar Tang Jie, the company was founded in 2019. Following OpenAI's release of GPT-3 in mid-2020, KNOWLEDGE ATLAS embarked on building a similarly scaled open-source alternative. According to co-founder and CEO Zhang Peng, an 18-person research team, including Tang Jie, accomplished this goal two years later with a computing power budget of just 4 million yuan. "Looking at the balance sheet, it was clear that the worst-case scenario was not having enough funds to finish the project," Zhang Peng stated in an interview. "We essentially had only one shot at success." KNOWLEDGE ATLAS's management believes the Beijing-based startup embodies the Chinese approach to AI development—deeply integrated with the real economy and its industrial giants. Zhang Peng acknowledged that private, localized deployment solutions are less flashy than consumer apps but argues they are precisely what Chinese enterprises need. State-owned enterprises are unlikely to rely on open-source models or public cloud infrastructure when building AI systems from scratch, creating an opportunity for KNOWLEDGE ATLAS. In contrast, Shanghai-based MiniMax began in the gaming sector. Co-founder Yan Junjie, an avid player of the battle arena game Dota 2, first took notice of OpenAI in 2019 when its bots defeated top human players worldwide. Deeply impressed, he immersed himself in the lab's research papers, shifting his career focus from computer vision to natural language processing. Yan Junjie recounted in an interview that he pitched investors on a plan to create AI agents approaching the threshold of the Turing test—a proposal met with skepticism by most. He secured his first investment from miHoYo, the studio behind Genshin Impact, whose founder is obsessed with using AI to revolutionize the gaming industry. miHoYo remains a key client for MiniMax. Both KNOWLEDGE ATLAS and MiniMax are seeking to push their technologies into global markets, prioritizing scale over immediate profitability. KNOWLEDGE ATLAS now offers an AI programming solution similar to Anthropic's Claude but at a fraction of the price. MiniMax competes with Sora and numerous Chinese firms monetizing AI video generation. Both companies are delving deeper into the "model-as-a-service" business—cloud services that allow users to deploy proprietary models and integrate them into their own applications. However, similar to concerns in the US, there are worries in China about overinvestment in AI infrastructure without clear paths to profitability. OpenAI challengers face mounting losses: from 2022 to 2024, R&D spending accelerated at both KNOWLEDGE ATLAS and MiniMax, while profits continued to decline. Source: IPO filings Following a brutal price war known as the "war of a hundred models," both companies are poised to raise hundreds of millions of dollars. MiniMax has brought on several cornerstone investors for its IPO, including Alibaba, the Abu Dhabi Investment Authority, and Korea's Mirae Asset Securities. Constrained by US trade restriction blacklists, KNOWLEDGE ATLAS's investors are predominantly domestic, including institutions like China International Capital Corporation and Taikang Insurance. This wave of listings follows a series of strong debuts by prominent Chinese chipmakers, viewed as crucial to China's quest for technological self-reliance and success in the AI race. In December, AI chipmaker Moore Threads Technology Co. saw its shares surge 425% on its first day of trading on Shanghai's STAR Market, followed by MetaX Integrated Circuits Shanghai Co., whose stock skyrocketed 693% post-listing. "This round of IPOs is critically important for the Chinese government, which is clearly fully supportive of these companies accessing public equity capital," noted Klay Group analyst Ebrahim. "Simultaneously, investors need to recognize that most of these newly listed companies are still loss-making, and this situation may persist for some time as they scale up."