On August 31, Momenta (6880.HK), a leading physical AI company, released its first-half 2026 earnings report — its first since going public. The results showcase robust financial performance across multiple metrics, underscoring how its industry-leading world model foundation platform is accelerating both technological advancement and commercial growth.
**Revenue Surges 76% in H1, Steering the Industry Toward Sustainable Profitability**
According to the financial report, Momenta achieved record highs in both revenue and gross profit during the first half of 2026. Revenue reached RMB 1.60 billion, representing a 76% year-on-year increase. After excluding share-based compensation expenses, gross profit climbed to RMB 1.21 billion, up 75% from the prior year, with a gross margin of 75% — maintaining a leading position within the industry.
Notably, Momenta’s adjusted net loss narrowed sharply to RMB 14.10 million in H1, a substantial 97% reduction year-on-year. Operating cash outflow also contracted by 47% compared to the same period last year. These marked improvements in core profitability metrics not only demonstrate the company’s stable operational trajectory but also signal a positive shift for the autonomous driving sector, which has long been characterized by heavy cash burn — moving instead toward a more sustainable profitability model. Following its IPO fundraising in July, Momenta’s cash reserves have surpassed RMB 15 billion, providing ample resources to accelerate its scaled autonomous operations and global strategic layout.
Positioned as “the preferred choice of global brands,” Momenta retains the largest market share among third-party providers of city NOA (Navigate on Autopilot) solutions. The company has established partnerships with 26 automakers worldwide — the highest number of international clients in the industry — making it the first Chinese enterprise to achieve global mass production of intelligent driving solutions. Its production-ready systems are now deployed across more than 10 countries and regions in Asia, Europe, Oceania, Latin America, and North Africa, continuously supporting major automakers in their overseas expansion.
The sustained breakthroughs in its mass-production business have fully unlocked Momenta’s “flywheel effect”: Over 1.1 million production vehicles now carry Momenta’s systems, with more than 110 production models successfully delivered and cumulative design wins exceeding 230 models. The latest 100,000-unit delivery milestone was achieved in just 36 days. Furthermore, Momenta has commenced L3 autonomous driving R&D, with mass production targeted for 2027.
Xu Dong Cao, CEO of Momenta, commented: “Guided by our mission of ‘Better AI, Better Life,’ we remain committed to data-driven development, building a world-leading world model foundation platform. With the world model as our unified foundation, we leverage three layers — pre-training, simulation, and reinforcement learning — to enable autonomous driving to deeply understand the physical world, progressively surpassing human-level performance. This technological capability has already extended from mass-production scenarios to large-scale unmanned businesses such as Robovan and Robotaxi, and is rapidly expanding into overseas markets, continuously strengthening our global leadership.”
**World Model Foundation Platform Drives Multi-Scenario Generalization, Accelerating Technology and Commercial Advancement**
The world model has emerged as the next core growth track following large language models. In H1, Momenta’s R&D investment accounted for 73% of revenue, with R&D staff comprising 79% of its workforce. This high-caliber R&D investment has propelled rapid model iteration. In March of this year, Momenta unveiled the R7 world model, marking the “prelude to physical AI.” Currently, leveraging over 15 billion kilometers of real-world driving data accumulated from production operations, more than 100 million segments of premium data extracted from that corpus, and over 10 million kilometers of daily simulation, the R7 world model enables self-play evolution through reinforcement learning training, delivering substantial improvements in critical scenario capabilities.
Momenta is committed to building a world model foundation platform that employs the same large model architecture to support the large-scale deployment of passenger vehicles, Robovan, Robotaxi, and Robotruck — with future expansion into embodied intelligence and other domains.
Built upon this unified world model foundation platform, all business lines share the same core large model, while the massive volumes of real-world driving data generated across diverse scenarios simultaneously drive up model parameter counts, enhancing model capabilities and product experiences while also diluting R&D costs. Momenta’s L4 and overseas operations have become new growth engines, accelerating the opening of the trillion-yuan Robo-market.
Momenta’s Robovan business has been deployed in Xiangcheng District, Suzhou, partnering with mainstream express delivery companies including China Post, JD.com, SF Express, and the “Three Tong One Da” courier services. Operations cover scenarios ranging from parcel delivery and nighttime distribution to fresh food, fast-moving consumer goods, community group buying, and hardware components — achieving round-the-clock, high-efficiency unmanned logistics. Built on the R7 world model, Momenta’s Robovan technology meets automotive-grade standards, delivering exceptionally safe products and services to the unmanned logistics industry. Additionally, Momenta has pioneered the introduction of its “map-free solution” — validated through years of mass-production passenger vehicle practice — into the Robovan sector, dramatically improving city rollout speed and fleet deployment efficiency while safely handling scenarios such as temporary road conditions.
Within this year, Momenta’s Robotaxi will commence operations in Shanghai and Abu Dhabi. Notably, Momenta secured the autonomous driving operation license for Shanghai’s Pudong area in just eight months, setting a new industry record. In July, Momenta received approval from Germany’s Federal Motor Transport Authority to conduct L4 autonomous driving tests on urban roads nationwide — marking the first time a Chinese company has obtained this qualification. To date, Momenta has deployed testing operations in cities including Shanghai, Suzhou, Wuxi, Shenzhen, Abu Dhabi, and Munich, with plans to expand to more than 10 cities globally by year-end. The company has also established partnerships with global mobility platforms such as Uber, Grab, Lumo, and SAIC Mobility, continuously expanding its worldwide ecosystem.
**Positive Growth Flywheel Achieved, Physical AI Long-Term Value Recognized by Capital Markets**
Leveraging its world model foundation platform, Momenta has become the first in the industry to complete the “Data-Compute-Model-Business” positive growth flywheel, continuously deepening its dual moats in technology and commerce.
Specifically, as real-world driving data continues to accumulate (Data Scaling) and training compute grows (Compute Scaling), the parameter count of the world model rapidly increases (Model Scaling), driving continuous improvements in model capabilities and product experience. This, in turn, leads to more vehicle model design wins and commercial revenue (Business Scaling). More design wins further accumulate real-world scenario data, while higher revenues feed back into compute investment — ultimately forming the “Data-Compute-Model-Business” positive growth flywheel that continuously reinforces Momenta’s competitive advantages.
Recently, a number of prominent investment banks — including CICC, Deutsche Bank, Bank of America, CITIC, and BOC International — have initiated coverage on Momenta, all assigning “Buy” or “Overweight” ratings, with the highest target price set at HK$520 per share.
In July, Momenta was officially listed on the Hong Kong Stock Exchange, drawing enthusiastic market response with unanimous support from dozens of top-tier long-only funds. For example, sovereign wealth funds focused on long-cycle value investing — including CIC, GIC, ADIA, Temasek, and OIA — alongside asset management and private equity institutions such as Fidelity International, BlackRock, J.P. Morgan Asset Management, Boyu Capital, and Hillhouse, formed an “all-star lineup,” underscoring strong consensus recognition of Momenta’s financial performance and its long-term value in the physical AI domain.
