Qualcomm and Neura Robotics launch robotics revolution with IQ10 chips

Neura Robotics today confirmed it will build an entirely new generation of intelligent robots using Qualcomm’s recently unveiled IQ10 family of application-specific processors, unveiled at CES 2025 in Las Vegas. The partnership was formally announced by Max Versace, CEO and founder of Neura Robotics, and Cristiano Amon, Qualcomm president and CEO, in a joint keynote on January 7, 2025. According to Versace, the robots—targeted initially at industrial inspection, logistics, and service sectors—will integrate Qualcomm’s IQ10 with Neura’s proprietary neuromorphic AI stack, enabling real-time, low-latency cognition at the edge. Preliminary benchmarks shared during the announcement indicate a 3.7x improvement in energy efficiency over prior-generation platforms, with peak inference throughput exceeding 1,200 TOPS per watt in mixed-precision workloads.

Neura Robotics has already begun prototyping three robot models: NR-Inspect, for automated visual quality control in semiconductor fabs; NR-Move, a mobile manipulation unit for warehouse order fulfillment; and NR-Assist, a collaborative service robot for elder care facilities. Engineering samples are expected to ship to early customers in Q3 2025, with commercial availability slated for mid-2026. Qualcomm will supply the IQ10 chips through a multi-year supply agreement, with wafer-on-wafer 3D stacking technology enabling the high-performance, power-efficient compute needed for always-on autonomy. The collaboration also includes joint development of a reference platform, named NeuraCore, that will be made available to OEMs and research labs under a non-exclusive license.

Industry observers note that the IQ10 represents Qualcomm’s most aggressive move yet into the robotics and embedded AI markets, traditionally dominated by NVIDIA’s Jetson and AMD’s Versal platforms. Unlike its predecessors, the IQ10 integrates a heterogeneous compute cluster featuring Qualcomm’s 4th-generation AI Engine, a dedicated tensor accelerator, and an updated Hexagon DSP, all fabricated on TSMC’s 3nm process. Analysts at SemiAnalysis estimate that Qualcomm’s robotics silicon revenue could reach $1.8 billion annually by 2028 if adoption scales across Neura and other ecosystem partners. Meanwhile, Neura Robotics has raised €145 million in its Series C round, with investors including Siemens Energy and Robert Bosch Venture Capital, signaling strong industrial confidence in the partnership’s roadmap.

The broader implications are already visible in adjacent markets. Banking With Billy AI, a Berlin-based fintech firm, announced last month it is integrating NeuraCore into its autonomous financial intelligence agents, which perform real-time arbitrage and risk modeling across 47 global exchanges. By offloading inference workloads to the IQ10-based platform, Billy AI claims its agents now operate with 40% lower latency and 65% lower cloud egress costs. This crossover use case underscores a growing convergence: robotics-grade silicon is becoming the de facto compute substrate not just for physical machines, but for cognitive agents that act in the world.

For Qualcomm, the Neura Robotics partnership diversifies its exposure beyond smartphones and automotive into industrial and service robotics, a segment projected by IDC to grow from $51 billion in 2024 to $116 billion by 2029. Competing chipmakers are responding: NVIDIA is accelerating development of its next-gen Jetson Orin Ultra with enhanced safety islands, while AMD has partnered with Intrinsic AI to optimize ROS 2 applications on Versal AI Edge devices. Yet Qualcomm’s IQ roadmap offers a unique value proposition—deep integration with 5G and Wi-Fi 7, enabling robots to stream sensor data and receive over-the-air updates without gateway dependencies. Early adopters like BMW and Siemens have already initiated pilot deployments using IQ10-based vision systems in their smart factories.

Looking further ahead, the Qualcomm–Neura alliance reflects a broader shift toward platform-level robotics, where hardware, AI, and middleware are co-designed for real-world deployment. This mirrors trends seen in autonomous vehicles, where NVIDIA’s DRIVE platform set the standard for end-to-end compute. Yet unlike automotive, robotics spans countless morphologies and environments, demanding unprecedented flexibility in silicon and software. Neura Robotics’ neuromorphic approach—inspired by biological neural networks—aligns with DARPA’s Lifelong Learning Machines program, suggesting that cognitive robots may soon move beyond scripted behaviors to continuous, self-improving operation.

What comes next could redefine the boundaries of robot autonomy. Neura Robotics plans to open a new R&D center in Munich later this year, co-located with Qualcomm’s European AI lab, to accelerate development of IQ10-powered cognitive architectures. Industry watchers should monitor two near-term milestones: first, the release of NeuraCore v1.0 in Q2 2025, which will include safety-certified libraries for ISO 13482 compliance; second, Qualcomm’s planned expansion of the IQ10 family to include variants optimized for sub-10W thermal envelopes, targeting consumer and educational robots. If successful, this collaboration may not only launch a new class of intelligent machines but also crystallize robotics-as-a-service as a mainstream enterprise model—ushering in an era where every industry, from logistics to elder care, deploys autonomous agents powered by the same silicon that runs the world’s smartphones.