Chinas New Tech Champions: Pioneering AI and Autonomous Logistics

China's Tech Vanguard: From Generative AI Breakthroughs to Autonomous Logistics, New Champions Emerge

A seismic shift is underway within China's technology sector, moving beyond traditional hardware manufacturing and internet platforms into the foundational layers of artificial intelligence and its tangible, large-scale commercial applications. Two companies, operating in seemingly distinct spheres, exemplify this maturation: MiniMax, a generative AI pioneer whose valuation skyrocketed during the recent Lunar New Year, and Jiushi Intelligence, an autonomous logistics vehicle firm that has just secured unicorn status with a valuation surpassing 10 billion yuan. Their parallel ascents highlights China's evolving tech landscape, where strategic foresight, architectural gambles, and a relentless focus on global commercialization are defining a new generation of champions.

MiniMax: The Calculated Gambler Rewriting AI Economics

The story of MiniMax is one of impeccable timing and contrarian conviction. Its founder, Yan Junjie, a former vice president at Chinese AI giant SenseTime, made a pivotal decision in late 2021. On the cusp of SenseTime's IPO, which would have likely netted him significant wealth, Yan departed to establish MiniMax. His vision was predicated on a belief that the prevailing AI paradigm—training specialized models for narrow tasks—was inefficient and limiting. He envisioned a future of general-purpose models, a conviction spectacularly validated months later by the global explosion of ChatGPT.

However, MiniMax's path diverged sharply from its peers in three critical areas, each a calculated risk.

First, against the industry trend of specialization, MiniMax committed to full-stack, multimodal self-reliance from its inception. While competitors focused on text, speech, or video in isolation, Yan argued that true intelligent interaction required mastering all modalities—text, image, speech, and video—simultaneously. This "unfocused" strategy initially worried investors but has since positioned MiniMax as one of the few globally with top-tier capabilities across all modalities.

Second, while the industry largely chased enterprise (B2B) clients, MiniMax adopted a dual-pronged approach, aggressively developing consumer-facing (C2C) applications. Products like the AI companion app Talkie and the video-generation tool Hailuo AI were first launched and gained traction overseas before being introduced domestically. This global-first, consumer-centric strategy has paid extraordinary dividends. MiniMax now boasts over 200 million individual users across more than 200 countries, with overseas markets contributing over 70% of its revenue. Talkie alone has amassed 147 million users, while Hailuo AI's user base exceeds 40 million.

The third and most consequential gamble was architectural. In mid-2023, Yan insisted on shifting the company's entire technical foundation from the industry-standard Dense architecture to a Mixture of Experts (MoE) architecture. The move was perilous; MoE is notoriously difficult to train, and two initial attempts failed, bringing the company to the brink. The third attempt succeeded, positioning MiniMax ahead of the curve just as the industry began recognizing MoE's superiority in computational efficiency. In January 2024, it launched abab, China's first MoE large language model, achieving performance that places it in the global first tier with a fraction of the expenditure of rivals like OpenAI.

These strategic bets fueled meteoric growth. The company's revenue for 2024 reached 214 million yuan, a year-on-year increase of 782.2%. Its valuation, reflected in the share price of its Hong Kong entity, surged by over 100 billion Hong Kong dollars during the recent Spring Festival, cementing its status as the highest-valued among China's leading AI startups, or "Six Little Tigers."

The journey, however, was not without near-catastrophic challenges. The early 2025 open-source release of the revolutionary DeepSeek-R1 model sent shockwaves through the industry, threatening to render many startups obsolete. While peers faltered, MiniMax doubled down. In a remarkable display of resilience, it launched its own open-source reasoning model, MiniMax-M1, just six months later, followed by four other major product releases in as many days. The episode underscored a tenacity that has become a hallmark of the company.

Jiushi Intelligence: Scaling Autonomy in the Global Supply Chain

If MiniMax represents a triumph in the digital realm, Jiushi Intelligence demonstrates the massive commercialization of AI in the physical world. Founded in 2021, the company specializes in L4-level autonomous urban delivery vehicles, known as RoboVans. Its recent completion of a $300 million funding round at a valuation exceeding 10 billion yuan marks the birth of the first unicorn in its specific segment.

Jiushi's growth narrative is defined by aggressive scaling and strategic consolidation. Its fleet deployment has followed a near-vertical trajectory: from 200 vehicles in 2023 to 2,000 in 2024. A pivotal moment came in early 2025 with its strategic merger with Alibaba's Cainiao Network's autonomous vehicle unit. The integration created a dual-brand operation (Jiushi and Cainiao) with a combined fleet of over 20,000 vehicles, operating in more than 300 cities across 10 countries, including Singapore, Japan, and the UAE.

The company's value proposition is ruthlessly economic. Its flagship Z5 model, an L4 vehicle with over 5 cubic meters of cargo space, helps clients reduce last-mile delivery costs by approximately 50%. For parcel delivery franchisees, the integrated cost per vehicle per month is between 2,000 and 3,000 yuan, lowering the per-parcel delivery cost from 0.2 yuan to 0.1 yuan. This compelling unit economics has secured large-scale contracts, most notably a 7,000-unit order from China Post in late 2025.

Co-founder Zhou Qing has stated that excluding R&D costs, the company has reached breakeven. The path to full profitability, inclusive of R&D, is pegged to a fleet size of around 50,000 units—a target that appears increasingly within reach given its current growth rate. The sector is attracting intensified competition, with players like QCraft and Desay SV entering the fray. Zhou characterizes the industry as moving from a "Warring States period" toward eventual consolidation, with Jiushi positioning itself as a potential unifier.

Converging Paths: Strategic Foresight and Global Ambition

Despite their different domains, the trajectories of MiniMax and Jiushi Intelligence reveal shared characteristics defining China's new tech vanguard.

Both companies were founded by founders with deep technical pedigrees who identified and committed to a specific, forward-looking thesis ahead of the mainstream curve—Yan on general AI and MoE, Jiushi's founders on the immediate scalability of autonomous logistics.

Each has embraced a fundamentally global strategy from an early stage. MiniMax built its user base overseas first; Jiushi expanded internationally alongside its domestic deployment. This outward focus is a marked evolution from earlier Chinese internet models.

Furthermore, both have demonstrated an ability to not just innovate technologically but to translate that innovation into clear, scalable economic value—for MiniMax, through viral consumer apps and efficient model architecture; for Jiushi, through directly slashing logistics costs.

Their stories also underscore the intense, rapid-iteration competitive environment in China. Both faced existential threats—MiniMax from DeepSeek, Jiushi from an influx of well-funded competitors—and responded with accelerated product cycles and strategic maneuvering.

The rise of these companies signals a broader transition. China's tech ecosystem is nurturing firms that compete on architectural innovation and global commercial execution in deep technology fields. Their successes, built on strategic bets that defied conventional wisdom, suggest that the next phase of China's tech competition will be fought not just in applications, but in the foundational models and autonomous systems that will power the global economy.