One year in Silicon Valley

When I moved from France to California in the summer of 2025, I expected to find the epicenter of artificial intelligence. Like many outsiders looking at Silicon Valley from afar, I imagined a region consumed by large language models, venture capital, billion-dollar valuations and the race to build the next technological breakthrough. After nearly a year living and working here, I have come to a different conclusion. Artificial intelligence is certainly everywhere, but the most interesting conversations taking place across the Valley are often not about AI itself. They are about energy, infrastructure, manufacturing, semiconductors, cooling systems, data centers and, increasingly, the physical limits of computation.

My work brought me to California as the U.S. representative of a French technology company operating in the edge computing and AI infrastructure space. Since arriving, I have spent much of my time moving between San Francisco, Palo Alto, Stanford, Mountain View and San Jose, attending conferences, meeting founders, speaking with engineers, exchanging ideas with investors and trying to understand how the region thinks about the future. What surprised me most was not the speed at which things move, but the depth of the questions people are asking. The popular image of Silicon Valley is one of constant disruption, endless networking and rapid execution. While some of that is true, the reality is more nuanced. Many of the most successful people I have met spend an extraordinary amount of time thinking. They read extensively. They debate ideas. They explore topics far outside their immediate area of expertise. They are often obsessed with a specific problem long before they build a company around it.

One of the first lessons Silicon Valley taught me is that technology rarely evolves in isolation. Back in Europe, discussions about artificial intelligence often focus on applications. The conversation revolves around chatbots, productivity tools, automation and software. In California, particularly among people building the infrastructure layer, AI is increasingly viewed as part of a much larger system. Every conversation about machine learning eventually becomes a conversation about GPUs. Every discussion about GPUs eventually becomes a discussion about power consumption. Power consumption leads to cooling. Cooling leads to water. Water leads to infrastructure. Infrastructure leads to regulation, industrial policy and energy production. The further one follows the chain, the more artificial intelligence begins to resemble previous infrastructure revolutions rather than a traditional software cycle.

This realization became impossible to ignore throughout the year. At industry conferences, founders worried about access to compute resources. Infrastructure providers discussed cooling technologies. Investors debated energy demand projections. Researchers questioned whether future advances in AI might ultimately be constrained not by algorithms but by physical resources. For decades, software benefited from the illusion of being weightless. The cloud made computing feel distant and abstract. Artificial intelligence is forcing the industry to reconnect with physical reality. Intelligence requires hardware. Hardware requires energy. Energy generates heat. Heat requires cooling. Every breakthrough model is supported by an increasingly complex industrial ecosystem.

Perhaps the clearest example of this shift can be seen in the renewed interest surrounding edge computing. For years, the dominant assumption was that computing would become increasingly centralized. The cloud appeared to be the final destination of digital infrastructure. Yet many of the applications emerging today challenge that assumption. Manufacturing facilities, hospitals, transportation systems, warehouses and critical infrastructure environments often require real-time decision making, low latency and greater control over sensitive data. These constraints are driving a more distributed vision of computing. The future increasingly looks less like a centralized cloud and more like a network of intelligent systems operating across multiple layers of infrastructure. Rather than replacing the cloud, edge computing is emerging as a complementary architecture capable of bringing intelligence closer to the physical world.

Stanford has reinforced many of these observations. Over the past year, I have attended numerous events on campus, ranging from entrepreneurship talks and sustainability discussions to conferences on artificial intelligence and future technologies. What makes Stanford unique is not simply the quality of the speakers but the diversity of perspectives present in the room. Engineers exchange ideas with policymakers. Entrepreneurs debate researchers. Climate experts challenge technologists. The result is an environment where technological innovation is constantly examined through economic, societal and environmental lenses. In many ways, this interdisciplinary mindset may be one of Silicon Valley’s greatest strengths. It encourages people to think beyond products and ask larger questions about systems, institutions and long-term consequences.

Another lesson that has stayed with me concerns entrepreneurship itself. Before moving here, I viewed startups primarily as businesses. Silicon Valley gradually changed that perspective. The most impressive founders I encountered were not necessarily the most charismatic, the best connected or even the most technically gifted. They were often individuals who had become deeply fascinated by a problem that most people ignored. Some spent years studying energy infrastructure. Others became experts in robotics, cybersecurity, semiconductors or biotechnology. Their companies emerged from sustained curiosity rather than from a desire to build a startup. Looking back, this pattern appears repeatedly across the Valley. Successful ventures often begin with intellectual obsession. The business comes later.

This may explain why so many conversations here feel unusually future-oriented. People are not merely discussing next quarter’s results. They are debating what the world might look like ten or twenty years from now. During the past year, I have heard discussions about autonomous factories, humanoid robotics, decentralized energy systems, climate adaptation, nuclear power, space infrastructure and the future of human-computer interaction. Some of these ideas will fail. Many will take longer than expected. Yet the willingness to engage seriously with long-term challenges remains one of the defining characteristics of the region.

As I reflect on my first year in Silicon Valley, I increasingly believe that artificial intelligence is not the story. It is part of a larger story. The real story concerns the systems required to support intelligence at scale. Energy systems. Computing infrastructure. Supply chains. Education. Public policy. Human talent. The future will not be determined by any single technological breakthrough. It will emerge from the interaction between these interconnected systems. Understanding that complexity has been the most valuable lesson of my time here.

The public narrative surrounding Silicon Valley often focuses on innovation. After living and working here, I would describe it differently. At its best, Silicon Valley is a place where ambitious people attempt to solve difficult problems by combining ideas from multiple disciplines. Artificial intelligence may dominate today’s headlines, but the conversations happening beneath the surface suggest something broader is taking shape. We are not simply building smarter software. We are redesigning the infrastructure that will support an increasingly intelligent world. The software may capture the public imagination, but it is the underlying systems that will ultimately determine how far this transformation can go.