San Francisco – June 11, 2025

A profound transformation is underway in the field of robotics, particularly within Silicon Valley. The slow, incremental advancements of past decades have given way to a rapid convergence of advancements in artificial intelligence alongside leaps in hardware and software capabilities, a fusion that more and more media outlets are labeling the “robotics renaissance.”

Led by the term “Physical AI,” (a combination of artificial intelligence with physical systems) which gained significant traction after NVIDIA CEO Jensen Huang used it in his recent keynotes, this new chapter in robotics is both exciting and complex.

Yet past the captivating demos and sky-high valuations lies a more layered and nuanced reality. Are we genuinely on the verge of a future dominated by intelligent machines, or is this yet another cycle of overheated expectations, or perhaps something in between? Why is this happening now, and what does it mean moving forward? The team at Swissnex in San Francisco is closely tracking these emerging dynamics in Silicon Valley and is here to offer insight by cutting through the noise.

The association between AI and robotics is not new. To understand the current enthusiasm around the topic, we must acknowledge the cyclical nature of AI and robotics and look back to understand where today fits in the broader context.

As early as the 1960s, engineers at the Stanford Research Institute (SRI) developed “Shakey,” now known as the first mobile robot capable of reasoning about its actions. It paved the way for many AI concepts developed later. Following this milestone, American computer scientist Marvin Minsky predicted in a 1970 Life magazine article that within three to eight years, AI would achieve the general intelligence of an average human. Sounds familiar?

Another notable example is WABOT-2, a humanoid robot developed at Waseda University in Japan in 1984. This project marked a significant advancement in integrating complex reasoning and robotics. WABOT-2 was capable of performing complex, human-like tasks such as reading musical scores via a vision system and playing them on a keyboard, demonstrating a form of artistic expression.

These are just a few examples, but they illustrate a recurring pattern in the history of artificial intelligence, and by extension, AI and robotics, characterized by periods of great excitement followed by reduced funding and interest due to unmet expectations and technological limitations. These downturns are referred to as “AI winters.” The first AI winter occurred around 1974, and the second in the late 1980s.

That said, the present moment feels qualitatively different, driven by an incredibly interesting mix of advancements. Here are a few factors that, when combined, are fueling what we are seeing today:

1. The computational power of GPUs: GPUs (Graphics Processing Units) play a key role in today’s robotics, particularly for AI-driven tasks like perception and processing high-resolution images, as well as simulation and training, which is now often done in virtual environments and requires significant computational resources.GPUs accelerate calculations, allowing robots to process large amounts of data and make real-time decisions efficiently. Although not new per se, GPUs have rapidly evolved. NVIDIA, based in San Jose, now holds a strategic position in the industry.
2. The Sharpening Senses: Sensors are used in robotics to calculate various parameters depending on the robot’s function, for instance, the condition and environment, using sensors just like humans use their sensory organs. Robots are gaining an unprecedented ability to perceive their surroundings, thanks to advancements in sensor technologies. For instance, borrowing from the camera industry, vision systems now offer higher resolution and greater dynamic range; LiDAR (used in self-driving cars) and force sensors, like those found in Amazon’s new Vulcan robot, enable finer manipulation and more human-like interaction by creating a sense of touch sensitivity.
3. The Development of Generative AI: One of the key transformative elements in recent development is the rise of generative AI. For example, Google’s Gemini models are capable of responding to text, images, audio, and video. Video Language Models (VLMs), a type of multimodal AI that can understand and generate text based on both visual and linguistic information, are making their mark in robotics. Google also recently released Gemini Robotics, which adds the ability to reason about physical spaces, allowing robots to take action in the real world.
4. Data Availability: While the internet, with its over 30 years of data, has provided a massive training ground for foundational AI models, robot-specific data remains scarce. As NVIDIA researcher Jim Fan noted during a recent panel discussion: “Where do we scrape all of those robot trajectories from the internet? You can’t find it anywhere. So, we have to generate data. We have to collect data at scale.” This leads to the next important point.
5. Software and Virtual Environments: Significant advances in robotics development tools and simulation environments are accelerating progress. Platforms like NVIDIA Isaac Sim and Isaac Lab enable the creation of realistic virtual environments where robots can be trained and tested extensively before deployment. These tools help bridge the long-standing “sim-to-real” gap. While improving the generation of high-quality data, accessibility to enough training data for robotics is still a challenge.
6. Economic and Labor Market Dynamics: Finally, rising labor costs and shortages in various sectors, such as healthcare, logistics, and manual services, are creating a strong economic incentive to automate tasks. Market projections are optimistic, with the global robotics market expected to grow significantly in the coming years. According to VC firm Fusion Fund, “The Physical AI industry is a rapidly growing sector with a promising future. In North America, the sector is expected to grow from $21.5B in 2025 to $42B in 2030, with a CAGR of 14.7%.”

All these factors are reinforcing each other, creating a ripple effect for innovation and technological advancement in robotics and AI.

The Bay Area stands as the center of this robotic resurgence. While not exhaustive, here are some of the key players:

Institutions like Stanford University and their newly opened Stanford Robotics Center, along with UC Berkeley, host several prominent robotics labs that contribute to fundamental research in areas ranging from control theory and perception to reinforcement learning and human-robot interaction.

These academic centers also form the next generation of experts. Organizations like Silicon Valley Robotics, the largest cluster of robotics and AI technology innovation and investment in the world, create critical links between academia and industry. Other institutions, such as the Open Source Robotics Alliance (OSRA) play a key role in improving the governance, funding, and long-term stability of open-source robotics projects.

What is Silicon Valley without its startups? In Silicon Valley, startups are always buzzing, and AI robotics startups are part of them! Figure AI, 1X Technologies, and Foundation Robotics Labs are developing increasingly capable humanoid robots, attracting significant investor and public attention. In Silicon Valley, humanoid robots and self-driving cars are currently very popular.

Meanwhile, major companies like OpenAI, Google, and NVIDIA are all located in the Bay Area and are building the foundational AI models and GPU architectures that make this progress possible. And let’s not forget the indispensable VC scene that fuels the whole ecosystem.

What makes this particularly exciting from a Swissnex perspective is that robotics (while not necessarily humanoids) is also a core strength of Switzerland.

Sometimes referred to as the “Silicon Valley of Robotics,” Switzerland brings together mechanical engineering, precision manufacturing, and world-renowned institutions like ETH Zurich, EPFL and universities.

These institutions serve as vital incubators for robotics projects and a talent pool. Swiss startups like ANYbotics, RIVR, Ascento Robotics, Gravis Robotics and aiEndoscopic further demonstrate Switzerland’s strength in this field. While the ecosystem may be smaller than Silicon Valley’s, it benefits from a deep emphasis on quality, specialization, software development and applied innovation.

Well, it depends on who you talk to, the time frame you’re looking at, and what type of robots you’re referring to. At the recent AI Ascent Conference on May 20, hosted by VC firm Sequoia, Jim Fan offered a compelling vision of what is coming after Physical AI: physical APIs, a whole “new physical economy,” and “physical prompting” for robots trained with foundational models.

Meanwhile, at Stanford University’s Human-Centered Artificial Intelligence (HAI) conference, Rodney Brooks (MIT Professor Emeritus and Cofounder of iRobot, Founder of Rethink Robotics) cautioned against the “Humanoid Theater.” He argued that while humanoids make for great demos, robotics development is a long game.

Along the same lines, Andra Keay, Managing Director of Silicon Valley Robotics, noted: “Although the first wave of humanoid robots was largely bipedal and conformed to the human form factor, the current wave of robots of interest to industry is largely ‘human-capable’ rather than ‘humanoid’ or ‘human-like.’ A mobile base with dual manipulators is going to be more reliable, robust, effective, and affordable than a more human-like humanoid robot.”

At Swissnex in San Francisco, we take an optimistic yet realistic approach, not underestimating the complexity of the physical world or the systems involved. The field is clearly developing at an incredible pace, with exciting breakthroughs almost every week.

As a bridge between Switzerland and the North American ecosystems, Swissnex in San Francisco has a crucial role to play in this unfolding “Robotics Renaissance”. Our mission extends beyond simply observing, we aim to actively facilitate the exchange of knowledge, talent, and opportunities. Here is what has already happened since the beginning of 2025 and what the rest of the year has in store:

• February: Startup Camp (Robotics), powered by Innosuisse and managed by Swissnex in San Francisco
• February: Robotic Residency with Switzerland Innovation Park Zurich
• June: The Robot Reality Check with Switzerland American Chamber of Commerce
• October: Embodied Intelligence Program at Swissnex San Francisco
• October: Startup Camp (AI & Robotics), powered by Innosuisse and managed by Swissnex in San Francisco

More to come, check out our event overview page frequently for more updates, or subscribe to our monthly newsletter.

Are you a Swiss startup, company, artist, or researcher interested in immersing yourself in the AI and robotics scene? Get in touch with the Swissnex San Francisco team or contact Maulde Cuerel, AI Program Manager at Swissnex in San Francisco.