Big Tech Versus the Manufacturing Old Guard: The AI Robotics Divide at Automatica 2025

Automatica saw a panoply of new products from the established market leaders, along with innovations from new entrants in the robotics space. New products were launched by the big four robotics Original Equipment Manufacturers (OEMs) (FANUC, KUKA, ABB, and YASKAWA). These products extended their existing portfolios, along with the introduction of new controllers, Artificial Intelligence (AI)-augmented solutions, and updates to offline programming interfaces. Controllers were a focus for all four companies. YASKAWA announced the extension of MOTOMAN NEXT (which utilizes NVIDIA’s Systems-on-Chip (SoCs)); KUKA announced the launch of a virtual robot controller and the iiQKA,ai Boost Engine, a module designed for AI acceleration; ABB highlighted the AI capabilities of its OmniCore controller in conjunction with the newly unveiled ABB Ability GenAI Suite; and finally, FANUC announced its next-generation R30iC Cognitive Core Controller, again, for AI acceleration. Three out of the four chose to develop their own AI-accelerated controllers, rather than partner with big tech.

Predictably, Automatica’s show floor was packed with AI-driven automation. Two key Operational Technologies (OTs) were vision-AI models for collaborative robot pick-and-place (solutions were offered by Solomon, Mech-Mind, Basler, Sick, Cognex, InBolt, OnRobot, Cambrian Vision, and many other vendors) and multi-robot control and proprietary algorithms for path planning, which also had a large showing. Siemens’ Standard Robot Command Interface (SRCI), along with Real Time Robotics, Intrinsic, and Evasive Robotics all demonstrate robot control interfaces that could provide a manufacturing edge to Western factories.

German startup Neura Robotics had a large show stand and unveiled its partner network, the Neuraverse. However, the failed demo of Neura’s much anticipated humanoid, 4NE-1, is a bad portent for the humanoid and larger AI robotics markets.

Ambivalence surrounding big tech’s infiltration of the industrial robotics space is evident throughout traditional robotics ecosystems. Beyond plugins to Omniverse, YASKAWA is the only traditional robotics OEM to adopt NVIDIA’s hardware as an OT. Other giants are penetrating the Information Technology (IT) space via cloud-native solutions: Comau has worked with Intrinsic since 2023 to deploy streamlined programming and planning interfaces to accelerate industrial robot adoption; Microsoft has active partnerships with KUKA and Kawasaki to provide Azure Internet of Things (IoT) and digital twin tools. Cobot companies—seeking differentiation and attempting to unlock greater value for their portfolios—have no reservation about partnering with NVIDIA: Universal Robotics, Neura Robotics, and Flexiv are leading examples of hardware partnerships.

The debut of Neura Robotics’ 4NE-1 robot is indicative of the precarious state of the AI-robotics market. Neura’s humanoid robot, which Chief Executive Officer (CEO) David Reger claimed would “beat […] all the other humanoids on this planet” failed to deliver, spectacularly. Systems Integrators (SIs) and industry decision makers at all levels consider this to be indicative of the state of AI-augmented robotics and innovation: overpromising and under delivering. Events such as this and Figure AI’s unjustifiable US$40 billion valuation should sound alarm bells for the entire robotics industry. If humanoid robotics—which popular consensus holds to be the ultimate form of AI robotics—continues to fail to deliver, the market will begin to lose trust; venture capital and investors will pull out, and the reputation of emerging robotics technology will be tarnished for years to come. It is for this reason that so few industrial robot OEMs have thrown their hat in with NVIDIA’s OT offerings: risk and an underlying suspicion that Physical AI is a baseless marketing ploy.

However, AI does in fact drive robotics. Algorithms enable robots to unlock new use cases; perform tasks faster; create intuitive programming interfaces; and, ultimately, enable robotics automation within entirely new verticals such as agriculture and healthcare.

Tier One robotics OEMs are in trouble. The now slowing Chinese manufacturing market has long presented the largest addressable market for industrial robot adoption, but the rise of Chinese OEMs, and their encroachment into Western markets, is causing a sharp decline in the shipments of all established manufacturers. A further issue is presented by the ambiguity of Western industrial policy: is reshoring still on the agenda or is nearshoring closer to the de facto reality?

OEMs have begun to realize the potential of the SME market and the need to address new verticals. ABB spinning out ABB Robotics and Omron reorganizing its robotics division can be interpreted as a break from traditional go-to-market strategies and existing marketing and commercial models. A divorce from the traditional SIs is necessary if robotics is to achieve wider uptake across economies. Although “turnkey” robotics solutions remain near impossible to achieve, simplified programing interfaces, lower-cost hardware, and end-user self-management and maintenance of solutions are necessary to expand into under-automated domains. Further, existing SI revenue models do not allow for small, geographically dispersed robotics deployments—adding the complexity of AI makes these solutions untenable from a traditional support and integration point of view. Addressing this dramatic shift in focus will require a new class of SIs and extremely simplified, and robust, robotics products. Recently announced AI accelerators will aim to achieve just this.

ABI Research’s recommendations for companies attempting to navigate the evolving robotics market are as follows:

• For the Established OEM: Invest in simplified programming interfaces. This is imperative not just for breaking into new markets, but also to bridge skilled worker shortages and growing demand from adopters for autonomy. Products can be developed in-house, such as drag-and-drop interfaces or by acquiring or partnering with companies offering low-cost and streamlined offline programming tools, including Visual Components, Cenit, RobotDK, and Artiminds. Another necessity is to understand the role of vision as a major driver of value within the emerging robotics space. Vision is important for robot control and major value adds. Investing in Three-Dimensional (3D) camera vendors and integrating products into robot control interfaces is paramount if solutions are to be accepted by integrators (both old and new) to achieve scale.
• For the Innovator: Seek differentiation and defeat hype. The automation market is currently awash with AI-vision-robotics solutions; what real-world problem does your product really solve? Importantly, innovators must design products for the SI as much as for the end user. Demonstrating value to the engineers that deploy solutions is the only way to achieve uptake; making solutions as simple to maintain as possible is the only way to achieve scale.
• For the Hyperscaler: Be aware of the fastidious nature of the automation market. There’s a reason why IT is seeing success within manufacturing and OT is floundering. Adopters are wary of tech companies encroaching into protected processes. The best way into automation markets, rather than backing startups, is through SIs themselves. The manufacturing world is built on reputation, trust, and demonstrable value. You must have all three to be accepted into active deployments. With traditional OEMs developing AI solutions and providing these to their SIs, the market will develop a taste for, and understanding of, the viability of AI-accelerated robotics—however, this will take time.