CiDi Inc. (HKEX: 3881), a global pioneer in Heavy-Duty Embodied AI, reported robust first-half 2026 results that mark a decisive inflection point: revenue nearly doubled, gross profit more than tripled, autonomous driving deployments expanded, and the company extended its full-stack autonomy platform into Heavy-Duty Embodied AI.

For the six months ended June 30, 2026, revenue reached RMB 803.7 million, up 97.0% year on year. Gross profit rose 204% to RMB 211.9 million, while gross margin increased to 26.4% from 17.1%.

With adjusted net loss narrowing to approximately RMB 19.7 million on revenue of RMB 803.7 million, CiDi has all but reached break-even—a threshold that signals heavy‑duty embodied AI is no longer a frontier experiment but financially sustainable and scalable technology.

Autonomous haulage moves deeper into fully commercial production

In autonomous driving, CiDi has moved decisively beyond pilots. The business generated RMB784.5 million in revenue in the first half of 2026, up 107.3% year on year.

By June 30, CiDi MetaMine had deployed more than 1,900 heavy autonomous mining trucks across 40 mining sites worldwide. Seven sites are running fleets of more than 100 autonomous trucks now, while the largest single-site fleet exceeded 220 vehicles.

One hundred percent of these vehicles were deployed as product sales to existing owners and operators in mining. Hence, the scale of CiDi’s deployments underscores both technical robustness and high customer approval.

These trucks execute full production workflows—overburden removal, coal loading and haulage, crusher feeding, and battery swapping—in mixed-traffic, all-weather conditions.

CiDi also advanced multiple large-scale autonomous deployments during the period. In February, the company secured an order from Guangna Group for 500 autonomous mining trucks, the largest single procurement ever recorded in the global autonomous haulage sector.

Autonomous logistics trucks are unlocking a new growth vector, with non-binding commitments for more than 400 units secured from global customers during the first half.

International expansion is advancing in parallel. Building on its presence in Australia, Brazil, and Indonesia, CiDi is targeting further growth across Africa, Central Asia, and the Middle East through partnerships with resource owners, operators, and equipment manufacturers.

From autonomous vehicles to Heavy-Duty Embodied AI

CiDi defines Heavy-Duty Embodied AI as the application of Physical AI to heavy equipment operating in demanding industrial environments. This means enabling machines to perceive their surroundings, make decisions, and act with increasing autonomy in environments that are hazardous, repetitive, or difficult for people to work in directly.

The company has built a common technology stack around three layers: at its base that ingests multi-modal sensor streams; a heavy-duty world model that predicts environmental evolution, enabling multi-agent collaboration; and an orchestration engine at the top layer that coordinates multi-agent fleets and performs global optimization. This architecture allows CiDi to transfer capabilities honed through years of autonomous driving deployments across a heterogeneous fleet of heavy-duty machines.

Task Robots extend embodied AI beyond haulage

CiDi is extending embodied AI beyond transport into task-specific heavy machinery. During the first half of 2026, CiDi expanded its robotics portfolio into excavation, rock drilling, explosives-charging, metallurgical material handling, underground operations, and industrial cleaning.

The company delivered 34 roboexcavator units, as well as heavy-duty ladle-handling robotics for high-temperature material transfers in metallurgy. During this period, CiDi has also begun engineering and fabrication of its underground mining and tunnel robots as well as completing development for open‑pit explosives‑charging robots to automate hazardous blasting workflows. These robots provide comprehensive autonomy solutions across global resource excavation industries.

A broader technology base

CiDi’s automotive-grade perception and V2X businesses form a complementary layer of its technology platform. These capabilities generate standalone commercial opportunities while serving as foundational components of the company’s autonomous driving and embodied AI systems. Perception endows machines with environmental awareness; V2X complements that awareness beyond onboard sensors by connecting vehicles, infrastructure, equipment, and people. CiDi’s Train Autonomous Perception System (TAPS) for railway has achieved SIL 4, the highest safety integrity level under the applicable functional safety framework.

Rather than treating autonomous driving as a capability confined to an individual truck, CiDi is building a shared intelligence layer that coordinates vehicles, equipment, infrastructure, and production workflows across industrial operations.

Its defining advantage may ultimately lie not in fleet scale alone, but in a unified intelligence architecture that extends across multiple classes of heavy machinery, giving CiDi a defensible edge as Physical AI expands across industrial operations.