MIPI Alliance Launches Physical AI BoF Group for Humanoid Robotics

The MIPI Alliance announced the formation of a Physical AI Birds of a Feather (BoF) group on April 28, 2026, with an initial focus on humanoid robotics. The initiative examines how existing MIPI specifications can support the rapidly developing humanoid application space, addressing what the organization describes as a lack of standardized PHY, protocol, and application interfaces for physical AI markets.

This is not a working group yet. It's a BoF—a preliminary gathering where industry players hash out whether standardization is even feasible before committing to formal specification development. At the conclusion of its evaluation, the group will provide a formal recommendation to the MIPI Alliance Board of Directors regarding potential specification development opportunities.

The market timing is deliberate. Yole Group projects the humanoid robot market will grow at a 56% CAGR to reach more than $6 billion by 2030, with potential to soar to $51 billion by 2035. Pierrick Boulay, principal analyst at Yole Group, noted that humanoids are reaching an inflection point where AI progress, supply-chain maturity, and component scaling are turning prototypes into deployable systems with measurable ROI in logistics and manufacturing.

Many of today's robotics systems rely on architectures originally developed for mobile, automotive, and industrial systems. The industry is shifting toward more efficient, commercially viable designs that prioritize power efficiency, cost optimization, and reduced form factors. This evolution provides an opportunity for MIPI to address interface gaps with an architecture based on accumulated experience and proven technology.

Edo Cohen, chair of the MIPI Physical AI BoF Group and vice chair of MIPI's Technical Steering Group, explained the rationale. He has more than three decades of experience as a senior system engineer, with extensive expertise in system architecture and technical specifications. Prior to joining Valens Semiconductor, Cohen worked as a senior system architect at Intel Corporation, leading efforts in the wearables and IoT division.

The BoF group will analyze and compare current hardware and software architectures, develop system diagrams for key application areas, and identify how MIPI specifications can be leveraged or enhanced to meet market needs. This is the kind of work that happens in conference rooms with whiteboards covered in system diagrams, not in marketing decks. The physical reality of humanoid robotics involves managing data flow between sensors, processors, and actuators in real-time—latency matters when a robot arm is moving near human workers.

Member companies already participating in the BoF group include Amphenol, Binho LLC, BitifEye Digital Test Solutions GmbH, Genesys Logic Inc., Intel Corporation, I-PEX, KIOXIA Corporation, Lattice Semiconductor, LG Electronics, Inc., MediaTek Inc., Mixel, Inc., NXP Semiconductors, Robert Bosch GmbH, ROHM Co. Ltd., Samsung Electronics, Co., Silvaco, Inc., Sony Group Corporation, STMicroelectronics, Synopsys, Inc., TDK (InvenSense, Inc.), Texas Instruments Incorporated, Unisoc (Shanghai) Technologies Co., Ltd., and Valens Semiconductor.

That's a significant roster. The presence of Samsung Electronics, Sony Group Corporation, and Intel Corporation signals serious industry interest. These companies don't typically commit resources to exploratory groups without seeing concrete commercial pathways. The semiconductor bill-of-materials for humanoid platforms represents a multi-billion dollar opportunity, according to Samsung Market Intelligence.

MIPI is also launching an Emerging Technologies Initiative to encourage broader industry participation for members and non-MIPI members in the exploration of emerging technology markets. Initial areas of focus will be on the Physical AI BoF and forthcoming Die-to-Die BoF, with potentially additional groups to come in 2026 and beyond. Through the initiative, MIPI Adopter member and nonmember companies will have a pathway to continue their involvement in the effort should the BoF evolve into a working group or the activities transfer to an existing MIPI working group.

The MIPI Alliance was founded in 2003 and has over 375 member companies worldwide with more than 15 active working groups delivering specifications within the extended mobile ecosystem. Members include handset manufacturers, device OEMs, software providers, semiconductor companies, application processor developers, IP tool providers, automotive OEMs and Tier 1 suppliers, and test and test equipment companies, as well as camera, display, tablet, and laptop manufacturers.

Companies interested in participating in the Physical AI BoF may sign up through the official MIPI Alliance press release. The announcement was also distributed via Business Wire, confirming the April 28, 2026 publication date and core details.

Technical implementation will require careful consideration. Humanoid robots need to process data from multiple sensor types—cameras, LiDAR, IMUs, force sensors—while maintaining low latency for real-time control. Current MIPI specifications were designed for mobile devices, where power constraints and form factors differ significantly from industrial robotics. Adapting these interfaces means understanding the physical constraints of robot hardware: cable routing, connector durability, electromagnetic interference in factory environments, and thermal management in confined spaces.

The MIPI Physical AI BoF Kickoff event on April 28, 2026, featured a market overview presented by Zach Haines, Samsung Senior Manager, Strategy & Market Intelligence at SSIC. Haines joined Samsung Electronics in 2022, focusing on corporate strategy and demand planning across the compute semiconductor ecosystem. Prior to Samsung, he worked on the global business operations team at AMD, where he forecast PC and server compute silicon.

Key development bottlenecks include lack of software maturity and limited access to robust training data. These aren't interface problems per se, but they affect how hardware specifications get written. If software stacks are fragmented, hardware standardization becomes less valuable. The BoF group needs to understand the full stack, not just the physical layer.

China's humanoid robot ecosystem also factors into the market landscape. The semiconductor opportunity spans OEMs, software vendors, and hardware suppliers, with leading OEMs pursuing commercial deployments. Startup investment trends and funding patterns will influence whether standardized interfaces gain traction or if proprietary solutions dominate early market segments.

Whether this BoF evolves into a formal working group depends on whether member companies see enough value in standardization versus proprietary approaches. The robotics industry has historically favored custom solutions for performance reasons. Standardization only wins when interoperability and supply chain benefits outweigh the optimization gains from bespoke designs. That calculus changes as the market matures and volume production becomes the priority over cutting-edge performance.

The MIPI Alliance has experience with this dynamic. Many of its specifications started as BoF discussions before becoming industry standards. The question for Physical AI is whether humanoid robotics will follow the same trajectory as mobile interfaces or diverge due to different technical requirements and market structures. Time will tell if the $51 billion projection materializes, and whether standardized interfaces become the backbone of that market or remain a niche consideration for specific use cases.

Artūras Malašauskas is an AI Systems Integrator with 20+ years of production-grade web engineering experience. He has designed, shipped, and scaled enterprise Python/PHP systems for logistics, SaaS, and public-sector clients. For the past year, he has focused exclusively on AI integrations: deploying open-source LLMs, building generative media pipelines (image, audio, video), and engineering multi-agent workflows for real production environments. His standard: reproducibility, security, cost-efficient inference—no vaporware. He documents and evaluates emerging AI tooling, separating verified capabilities from marketing noise. Technical editor at: muza-ai.eu, ai-verslas.lt, ai-naujinos.lt Connect on LinkedIn