CreateAI Holdings Inc., an artificial intelligence technology, focuses on video game and anime production and publishing activities to digital entertainment content.
Segments
The company’s two segments are the U.S. and APAC. These segments align with its bifurcated development strategy of its L4 autonomous driving technology, independent operations and commercial business models in each region. Each segment has stand-alone engineering teams, software code base, infrastructure, technological c...
CreateAI Holdings Inc., an artificial intelligence technology, focuses on video game and anime production and publishing activities to digital entertainment content.
Segments
The company’s two segments are the U.S. and APAC. These segments align with its bifurcated development strategy of its L4 autonomous driving technology, independent operations and commercial business models in each region. Each segment has stand-alone engineering teams, software code base, infrastructure, technological capabilities, and onboard software development to fit regional demands and differences in end-market use cases.
Autonomous Freight Network
The company is in the early stages of developing its AFN in its key markets. For both the U.S. and APAC, the TuSimple Capacity model with retrofitted trucks will enable an accelerated path to commercialization. The company has developed proprietary technologies and systems necessary for the development and deployment of its autonomous trucking operations within the Autonomous Freight Network (AFN). Once fully scaled, the AFN will aim to serve as a comprehensive autonomous freight solution that provides users with access to L4 autonomous semi-trucks operating on HD digital mapped routes connecting a network of terminals.
Full-Stack L4 Autonomous Driving Technology, Hardware and Offboard Capabilities
The company is developing L4 autonomous driving solutions for the U.S. and APAC operations with ADS L4 capable onboard software, offboard capabilities and hardware. The capabilities for the U.S. and APAC segments, including their software base code, are being developed independently by the technology teams in each region.
The company’s autonomous driving technology is specifically designed for semi-trucks in the geographies where it operates. The company’s development priorities and testing activities are focused on further refining its L4 autonomous driving technology to prepare them for commercial deployment. The company’s proprietary L4 autonomous driving solutions include its L4-capable ADS, such as 1,000 meter perception range, multi-modal perception, and offboard capabilities, such as high definition (HD) maps, machine-learning, autonomy visualization and simulation capabilities, and an integrated L4 autonomous semi-truck design consisting of a redundant sensor suite, on-board computing solutions, and other components. Long-range perception, advanced planning and decision-making, and highly accurate mapping are important capabilities for the autonomous freight operation of semi-trucks, which are heavy, articulated vehicles that need to be able to operate at highway speeds.
The company operates approximately seventy L4 autonomous semi-trucks, being 35 in the U.S. and 35 in APAC, with a safety driver and safety engineer in the cabin (Driver In). Operating with a safety driver and safety engineer allows the company to continually improve its L4 technology. The company’s AFN operations consist of the following:
U.S.
Arizona: Terminal-to-terminal operations; and 35 L4 autonomous semi-trucks (Class 8 trucks).
APAC
Shanghai, China: Port-to-terminal testing operations on the Donghai Bridge; Tokyo, Japan: Testing route on the Tomei Expressway between Tokyo and Osaka; and 35 L4 autonomous semi-trucks (Chassis-over-Engine trucks).
Other Locations: The company has also conducted select testing operations in other locations in the U.S., China, Japan, and Sweden as it makes Operating Design Domain (ODD) expansions and demonstrate its capabilities to customers and partners.
Commercialization
Readiness - Tech and AFN Development
The company anticipates commercialization for it to be achieved through scaled deployment of L4 autonomous trucks using its service models in its AFN. The company is in the early stages of developing its AFN in its key markets. The company is also developing an L4 autonomy solution needed to enable commercialization in the U.S. and APAC.
Service Models
The company is developing its AFN to provide autonomous freight capacity as a service through two service models based on users’ needs:
TuSimple Capacity: The company’s fleet of owned or leased retrofitted (near-term) and purpose-built L4 autonomous semi-trucks (longer-term) aims to serve customers that desire access to safe, reliable, low cost, and more environmentally friendly freight transportation without owning semi-truck assets.
Carrier-Owned Capacity: Customers that prefer to own their fleet will be able to purchase its purpose-built L4 autonomous semi-truck from an OEM with on-board autonomous driving software solutions.
TuSimple Capacity: Path to Commercialization
For both the U.S. and APAC, the company’s the TuSimple Capacity model with retrofitted trucks will enable an accelerated path to scale.
TuSimple Capacity’s early focus on retrofitted trucks is intended to allow for continued development, integration and validation under the company’s control while OEM purpose-built capabilities are designed, validated and production infrastructure is ramped up.
Roadmap
The company’s U.S. and APAC commercialization roadmaps consist of three major phases and are strategically aligned with overall industry readiness.
TuSimple Capacity - Retrofitted Trucks (Near-Term): Primarily focused on maturing the company’s ADS technology for commercializing select freight routes and developing its AV-ready redundant base truck platforms & components to support OEM production programs.
TuSimple Capacity - Retrofitted and Purpose Built L4 Trucks (Mid-Term): Aligned with longer OEM and Tier-1 supplier timelines. The company expects that this phase will be primarily focused on expanding its self-operated L4 truck fleet and AFN lanes, continuing hardware maturity and optimization, and freezing complete ADS onboard software and hardware for OEM production programs.
TuSimple and Carrier-Owned Capacity Purpose Built L4 Trucks (Long-Term): Phase slated with the launch of OEM production vehicles. The company anticipates that it will be primarily focused on scaling AFN density.
Core Technologies
Autonomy Technology and Capabilities
The company is developing an L4 autonomy solution for both its the U.S. and APAC operations with capabilities from a hybrid data-driven and principle-based framework with redundant architecture, utilizing a proprietary camera-centric architecture. The development and testing of TuSimple’s L4 autonomy solutions in the U.S. and APAC are carried out independently as they have unique ODDs (e.g., road ecosystems, traffic rules, infrastructure and driver behavior) and hardware ecosystems (e.g., Class 8 trucks in the U.S. and Chassis-over-Engine trucks in APAC), which require separate engineering teams, software code base and systems integration to properly differentiate the product roadmaps per the end-market use cases.
ADS Onboard Software
Perception, Tracking and Fusion
Visualizing the road environment with long-range capabilities that address highway driving speeds and long stopping distance requirements is critical to the company’s L4 autonomous semi-trucks’ safe and reliable operation. The company’s perception system is powered by proprietary artificial intelligence technology and, when coupled with its proprietary camera module, enables the semi-truck’s 1,000 meter perception range. Semi-trucks can take twice the distance as a passenger car to stop and this range across lighting environments is designed to provide the company’s trucks with sufficient reaction time to safely operate at highway speeds. The company’s camera-centric system is powered by both primary and backup cameras. Augmenting the camera perception is an array of LiDAR, radar systems, GPS, and ultrasonic sensors. With the exception of its specially designed long range high definition camera, the company has sourced the balance of its sensor suite from existing third party vendors. The company’s sensor fusion technology actively integrates and fuses data from the various sensors (e.g, camera + LiDAR + radar) and HD maps to perceive, understand and create a detailed 3D representation of the surrounding environment, which is critical to operate L4 autonomous semi-trucks.
Prediction and Motion Planning
The company’s motion planning software is designed to complement its proprietary perception system by predicting the future paths of surrounding vehicles, as well as analyzing diverse traffic scenarios. By utilizing prediction models that assess vehicle speeds and driver intent and evaluating other conditions or objects, the company aims to plan better and safer driving trajectories for its semi-trucks through all types of interactions in the ODD.
Control
The company’s control algorithms receive input from the motion planning module to put into action a safe and efficient driving trajectory. These algorithms are designed specifically for the semi-truck use case, as semi-trucks are significantly less nimble than passenger cars, with characteristics including longer gear shift timeframes and the need for higher actuation accuracy. Control algorithms are tightly coupled with prediction and planning by providing probabilistic outputs for control to take defensive actions.
ADS Hardware Solutions
The company has developed and have continued to refine several critical ADS-enabling hardware in the U.S. and APAC through multiple generations which enables safety, reliability, and cost reduction with in-house designed software. The company leverages off-the-shelf sensors or components in the U.S. and APAC where possible to remain capital efficient and focused on core competencies.
Autonomous Truck Upfit Capabilities
The company has been developing autonomous semi-truck technology for over eight years, and is on its eighth generation vehicle in the U.S. and its fifth in APAC, upfitting and integrating with multiple OEMs.
TuSimple Domain Controller
In 2022, the company entered into a partnership with NVIDIA, a leading graphical computing technology company, to develop a proprietary ADC computing solution named the TuSimple Domain Controller (TDC). The TDC is a central on-board computer that is being developed by the company’s APAC team in an effort to execute hundreds of trillions of operations per second for its autonomous driving system, including its perception, planning and control systems. The TDC is designed to replace multiple existing compute units as an integrated redundant controller. This first generation TDC has the ability to address multiple use-cases, including centralized compute for L2+/L3 autonomous solutions and redundant compute for L4 autonomous solutions. The company has developed the TDC for use in its upfitted semi-trucks. The TDC has broader applications in the L2+/L3 autonomy market as a standalone compute unit or an integrated hardware and autonomous software offering.
Offboard Infrastructure and Toolchain
The company has developed in-house the corresponding toolchains that it intends to support a scalable autonomous freight operation.
Machine Learning Infrastructure
The accuracy of the company’s perception and motion planning capabilities is augmented by its proprietary machine learning algorithms. Over the course of more than 10 million road miles in both the U.S. and APAC, the company has accumulated a vast set of semi-truck specific driving data to train its perception and motion planning capabilities. The company focuses its data collection and storage efforts on both quality and quantity. This enables its database to be informative and relatively easily analyzed by the company’s machine learning software. This data analysis trains the company’s software to more accurately predict how the surrounding environment of a semi-truck will change in real time.
Simulation
TuSimple simulation capabilities are critical enablers to accelerate the development and refinement of the company’s autonomous trucking technology. The company’s simulation capabilities allow it to train and refine its onboard software (e.g., prediction, motion planning, control) in a simulated virtual environment using real-time, synthetic road mile data.
Proprietary HD Mapping Capability
Integrated mapping software with rapid updating is crucial for the operation of an L4 autonomous semi-truck. The company’s AFN consists of HD digital maps of the lanes over which its semi-trucks operate and its in-house mapping technology can quickly map new routes and provide users with more location options for shipping on its AFN. Precise localization accuracy is crucial to safely operate semi-trucks autonomously.
Fleet Operation
The company’s customer user experience is complemented by its proprietary fleet management systems, which include control center and fleet scheduling systems. This cloud-based autonomous system is being designed to ensure safe operations, reliability, and efficient capacity for the company’s users. The system will be designed to directly connect to the company’s users’ fleet management software.
Visualization
Versatile and comprehensive visualization is critical to optimize the development and refinement of the company’s autonomous driving systems. The company’s visualization platforms accommodate various scenarios from different users, including offline triage and research engineers, and test operators. In addition, the company’s visualization platforms allow for a more efficient and centralized AV development with key shared components, including 3D display, data plotting, process monitoring, and dynamic parameter reconfiguration.
System Integration and Safety
The company aims to design safety-critical best practices in all major operational processes (the TuSimple Safety Case), which includes the formal and documented safety-critical design methodologies, rigorous verification and validation against safety critical requirements and close coordination with regulators. In 2022, the company passed an independent 3rd party (TÜV SÜD) safety audit to SAE standard J3018 and AV testing best practices with zero nonconformities.
Intellectual Property
The company’s patent portfolio of 500+ global issued patents has IP spanning across the entire autonomous freight ecosystem and serves as inherent value to TuSimple shareholders. As of December 31, 2022, the company had 519 issued patents and approximately 1,025 pending patent applications globally. As of June 30, 2023, the company had approximately 590 issued patents globally and approximately 1,170 pending patent applications.
Government Regulation
As the trucks that carry its technology go into production, the company is subject to existing stringent requirements overseen by the National Highway Traffic Safety Administration (NHTSA) under the National Traffic and Motor Vehicle Safety Act of 1966 (the ‘Vehicle Safety Act’), including a duty to report, subject to strict timing requirements, safety defects with its products. It is also subject to the existing U.S. Transportation Recall Enhancement, Accountability and Documentation Act (the TREAD Act), which requires motor vehicle equipment manufacturers, such as the company, to comply with ‘Early Warning’ requirements by reporting certain information to NHTSA, such as information related to defects or reports of injury related to its products.
The company’s truck drivers and trucking operations are subject to the Federal Motor Carrier Safety Regulations (the FMCSRs) established by the Federal Motor Carrier Safety Administration (FMCSA), and it is subject to the requirements of the federal Occupational Safety and Health Act, as amended (OSHA), and comparable international, state, and local laws that protect and regulate employee health and safety.
History
The company was founded in 2015. The company was incorporated in 2021. The company was formerly known as TuSimple Holdings Inc. and changed its name to CreateAI Holdings Inc. in 2024.
The Analyst Price Target shows the analysts’ low, high, and average target at a glance.
