Key Takeaways
- Nota AI has become a member of the AMD Robotics Partner Network, announced at AMD Advancing AI 2026 in San Francisco.
- The network provides an open, AMD‑validated ecosystem that brings together ODMs, ISVs, sensor providers, system integrators, and other stakeholders to accelerate end‑to‑end robotics innovation.
- Nota AI will apply its expertise in AI model compression and hardware‑aware optimization to enable efficient on‑device AI for robots operating under tight power, memory, and compute limits.
- By joining the network, Nota AI aims to expand its physical‑AI solutions—originally refined for smart cities, industrial safety, and mobility—into the robotics sector, supporting autonomous perception, reasoning, and action.
- Leadership from both Nota AI and AMD emphasize the collaborative nature of the partnership and its potential to accelerate scalable robotics deployments across industries.
Nota AI Joins the AMD Robotics Partner Network
On July 23, 2026, Nota AI, a specialist in AI model compression and optimization, announced its inclusion in the AMD Robotics Partner Network during AMD’s Advancing AI 2026 event in San Francisco, California. The partnership marks a strategic move to extend Nota AI’s on‑device AI optimization technology from its traditional edge‑AI domains into the rapidly growing robotics market. By aligning with AMD’s validated robotics ecosystem, Nota AI gains access to a broad suite of hardware platforms, development tools, and a community of partners focused on delivering end‑to‑end robotic solutions. This announcement underscores the increasing importance of efficient AI execution directly on robotic hardware, where latency, power, and thermal constraints demand highly optimized models.
Understanding the AMD Robotics Partner Network
The AMD Robotics Partner Network is designed as an open, collaborative ecosystem that enables members to build and deploy robotics solutions across embedded, edge, and cloud environments using AMD compute architectures. It brings together original design manufacturers (ODMs), independent software vendors (ISVs), sensor suppliers, system integrators, and other technology partners to co‑create scalable robotic platforms. Through shared reference designs, joint go‑to‑market programs, and access to AMD’s software stacks, the network reduces development friction and accelerates time‑to‑market for innovative robotic applications. By participating, Nota AI can leverage these resources to ensure its optimization tools are tightly integrated with AMD‑based robotic controllers and processors.
The Challenge of Running AI on Constrained Robotic Systems
Robots, especially those used in manufacturing, logistics, and field service, operate under strict power budgets, limited memory, and modest compute capabilities compared to data‑center servers. Deploying sophisticated AI models—such as those for object detection, semantic segmentation, or reinforcement learning—directly on these devices often leads to excessive latency, overheating, or battery drain if the models are not properly trimmed. Consequently, achieving real‑time perception and decision‑making requires aggressive model size reduction, computation‑aware scheduling, and memory‑footprint minimization without sacrificing essential accuracy. Nota AI’s core technology directly addresses these pain points by delivering hardware‑aware compression techniques that preserve model fidelity while drastically lowering resource consumption.
Nota AI’s AI Model Compression and Hardware‑Aware Optimization
Nota AI’s portfolio centers on advanced model compression methods—including pruning, quantization, knowledge distillation, and neural architecture search—combined with hardware‑aware optimization that tailors the compressed model to the specific characteristics of target processors (e.g., AMD’s Ryzen Embedded, EPYC, or Versal adaptive SoCs). The company’s workflow begins with profiling the target hardware to understand compute pipelines, memory bandwidth, and power envelopes, then applies transformations that align model operations with those strengths. The result is a leaner AI model that executes faster, consumes less energy, and fits within tight memory budgets, all while maintaining performance levels comparable to the original, un‑compressed version. This capability has already been proven in smart‑city video analytics, industrial safety monitoring, and autonomous mobility platforms.
Applying Optimization Expertise to Robotics via the AMD Network
Through its membership in the AMD Robotics Partner Network, Nota AI intends to port its proven optimization pipeline to the robotics domain. The company will work closely with ODMs to integrate its tools into robotic control boards, with ISVs to adapt AI perception stacks for compressed models, and with system integrators to validate end‑to‑end performance in real‑world scenarios. By aligning model compression with AMD’s compute‑centric software stacks—such as ROCm for GPU acceleration and Vitis AI for adaptive SoCs—Nota AI aims to deliver a seamless development experience where roboticists can train high‑capacity models in the cloud and then automatically generate optimized versions for on‑device execution. This approach reduces the barrier to entry for deploying advanced AI on robots that previously relied on bulky, power‑hungry solutions.
Enabling Scalable Physical AI Solutions for Autonomous Robots
The overarching goal of the collaboration is to foster scalable “physical AI” solutions—AI that interacts directly with the physical world through sensors, actuators, and mechanical systems. Optimized models will empower robots to perceive their surroundings with high fidelity (e.g., detecting obstacles, recognizing objects, estimating pose), reason about complex tasks (e.g., planning grasps, navigating dynamic environments), and act reliably in real time. By ensuring that AI workloads stay within the robot’s power and thermal limits, Nota AI’s technology helps extend operational uptime, reduces the need for external compute offloading, and enhances safety in collaborative settings. The network’s open nature further encourages cross‑industry reuse, allowing a model optimized for a warehouse logistics robot to be adapted, with minimal rework, for an agricultural drone or a medical assistance robot.
Leadership Perspective: Nota AI’s CEO Myungsu Chae
Myungsu Chae, CEO of Nota AI, emphasized the necessity of hardware‑aware AI for meaningful industrial impact. He stated, “For physical AI to deliver real value in industrial environments, AI models must be optimized for the compute and memory constraints of the robots on which they operate. By joining the AMD Robotics Partner Network, we are expanding our on‑device AI optimization technology into robotics and look forward to collaborating with fellow network partners to support the deployment of physical AI solutions across industrial environments.” Chae’s remarks highlight the company’s commitment to moving beyond laboratory demos toward rugged, production‑grade AI that can withstand the variability and demands of factory floors, outdoor sites, and service settings.
AMD’s Robotics Vision: KV Thanjavur Bhaaskar
KV Thanjavur Bhaaskar, Robotics Lead at AMD, underscored the collaborative ethos of the partner network. He remarked, “The AMD Robotics Partner Network brings together ODMs, ISVs, and system integrators to enable scalable robotics solutions across industries. It creates a collaborative ecosystem designed to accelerate innovation and expand market reach for partners.” Bhaaskar’s comment reflects AMD’s strategy of providing a robust hardware foundation while relying on partners like Nota AI to supply the software intelligence that transforms raw compute into capable robotic behavior. The synergy between AMD’s silicon expertise and Nota AI’s optimization know‑how is expected to lower development costs, shorten product cycles, and broaden the addressable market for AI‑enabled robots.
Nota AI’s Background: From Smart Cities to Mobility
Prior to its robotics focus, Nota AI has delivered on‑device AI solutions in a variety of edge AI applications. In smart‑city projects, the company’s compression techniques enabled real‑time traffic‑flow analysis and public‑safety monitoring on edge gateways with limited power. In industrial safety, Nota AI optimized video‑based anomaly detection models to run on ruggedized cameras installed near heavy machinery, providing instant alerts without overwhelming the host system. In the mobility sector, its work supported advanced driver‑assistance systems (ADAS) and autonomous‑vehicle perception stacks that needed to operate reliably within automotive ECUs. This diverse experience equips Nota AI with a nuanced understanding of how different operational environments shape model‑optimization trade‑offs—a knowledge base that will be instrumental as it tailors solutions for the unique constraints of robotic platforms.
Future Outlook and Collaboration Expectations
Looking ahead, Nota AI anticipates that its participation in the AMD Robotics Partner Network will spur joint reference designs, co‑optimized benchmarks, and shared best‑practice guides for deploying compressed AI models on AMD‑based robotic controllers. The company plans to engage in hackathons, developer workshops, and pilot programs with network members to validate use cases such as autonomous mobile robots (AMRs) in warehouses, collaborative robotic arms in assembly lines, and inspection robots in hazardous environments. By fostering an open exchange of ideas and resources, Nota AI hopes to accelerate the maturation of physical AI, making advanced perception and decision‑making accessible to a broader range of robotic applications and ultimately driving greater productivity, safety, and flexibility across industries.