Beyond Models
Why Artificial Intelligence Is Becoming an Industrial System
For years, artificial intelligence was largely presented as a software story. The focus was on algorithms, machine learning breakthroughs and increasingly powerful models capable of generating text, images and software code. Companies such as OpenAI, Anthropic, Google DeepMind and Meta became the public faces of the AI revolution. Yet beneath the headlines, another reality has been quietly emerging.
Artificial intelligence does not exist in isolation from the physical world. Every AI-generated response depends on electricity, semiconductors, networking equipment, cooling systems and increasingly complex industrial infrastructure. Intelligence may be digital, but its foundations remain profoundly physical.
As AI continues to scale, the industry appears to be evolving from a software sector into something much larger: an industrial ecosystem built upon energy, manufacturing capacity, infrastructure and supply chains.
“Artificial intelligence is shifting from a software sprint to an infrastructure marathon.”
Much of the public conversation surrounding AI remains focused on applications and models. These are the visible layers. The invisible layers are increasingly becoming the most important.
The Physical Foundations of Intelligence
Every interaction with an AI system requires physical processes taking place somewhere in the world. Electrons move through processors. Heat must be removed. Data travels through fibre-optic networks. Massive facilities consume electricity on a scale once associated primarily with heavy industry. The popular concept of “the cloud” often obscures this reality.
In practice, AI depends on a vast physical substrate consisting of silicon, copper, photonics, advanced packaging technologies, electrical grids and industrial-scale infrastructure.
The challenge is no longer simply creating intelligence. The challenge is sustaining it.
The Return of Industrial Power
Historically, technology industries were often viewed as separate from traditional manufacturing. Artificial intelligence is beginning to erase that distinction. The construction of AI infrastructure requires enormous physical investments:
- Semiconductor fabrication facilities
- Advanced packaging and assembly plants
- Data centres
- Electrical grids
- Cooling systems
- Optical networking infrastructure
As a result, companies once considered industrial suppliers are becoming increasingly strategic participants in the AI economy.
Electrical engineering firms, infrastructure providers, equipment manufacturers and industrial automation companies now occupy critical positions within the emerging AI value chain.
This may also explain why discussions surrounding AI increasingly include topics such as energy security, workforce shortages, grid modernisation and industrial policy.
AI is not replacing industry. AI is creating demand for it.
Why Semiconductors Are About More Than Chips
Semiconductors dominate many AI discussions. Yet the most important question may not be who designs the most powerful processor. It may be who can manufacture advanced systems at scale. Without the broader semiconductor ecosystem, artificial intelligence cannot exist.
The sector depends upon highly specialised equipment manufacturers, advanced materials suppliers, precision engineering firms and increasingly sophisticated packaging technologies.
This reality helps explain why companies such as ASML, Applied Materials, Tokyo Electron, ASM International and BESI occupy positions of strategic importance far beyond their public visibility.
The value of AI is not created by software alone. It emerges from an industrial chain spanning multiple continents and involving thousands of highly specialised suppliers.
“The global race for intelligence is no longer restricted to algorithms; it is a race for industrial capacity.”
The Emerging Geopolitical Divide
The industrialisation of AI is reshaping global competition. Different regions are pursuing distinct strategies.
United States
The United States remains dominant in many of the most visible layers of the AI ecosystem, including advanced chip design, hyperscale cloud infrastructure and foundation models. Yet this dominance also highlights a strategic vulnerability.
Much of the world’s most advanced semiconductor manufacturing capacity remains concentrated outside the United States. Recent initiatives such as the CHIPS Act reflect growing recognition that technological leadership increasingly depends upon industrial capacity as well as innovation.
China
China has pursued a different approach. Rather than focusing exclusively on frontier AI models, Beijing continues investing heavily across manufacturing, supply chains, industrial policy and technological self-sufficiency.
The country’s efforts to strengthen domestic semiconductor capabilities and reduce external dependencies suggest a long-term strategy centred on ecosystem resilience rather than individual technologies.
India
India is pursuing yet another path. For decades, India’s technology story was closely associated with software and services. Increasingly, however, the country appears focused on building industrial capabilities alongside its existing strengths in engineering talent.
Recent investments in semiconductor manufacturing, packaging, assembly and electronics production indicate ambitions that extend beyond software alone.
Europe’s Underestimated Position
Europe is often portrayed as lagging behind in artificial intelligence. The narrative typically focuses on the absence of a European equivalent to OpenAI or other major foundation model developers. Yet this perspective may overlook an important reality.
Europe occupies several of the most strategically important layers of the global technology ecosystem.
From semiconductor manufacturing equipment and advanced engineering to industrial automation, telecommunications and enterprise software, European companies continue to provide capabilities that remain difficult to replace.
Companies such as ASML, BESI, ASM International, Siemens, SAP, Ericsson and Nokia operate in areas that underpin the broader AI ecosystem.
Europe may not dominate every layer of artificial intelligence. But it remains deeply embedded in the infrastructure that makes large-scale AI possible. In that sense, Europe’s role may be less about visibility and more about indispensability.
“Intelligence may be digital, but its foundations remain profoundly physical.”
Beyond Artificial Intelligence
Perhaps the most important lesson is that the future of AI may ultimately be about much more than artificial intelligence itself.
Energy generation.
Industrial production.
Supply chains.
Infrastructure resilience.
Workforce development.
Technological sovereignty.
All are becoming part of the same conversation.
For years, artificial intelligence was primarily discussed as a software revolution. Increasingly, it is beginning to resemble an industrial revolution.
The critical question may no longer be who can build the most capable model, but who can build the ecosystems capable of sustaining intelligence at scale.
In that sense, the future of AI may depend as much on power grids, factories and supply chains as on algorithms themselves.
Credit
Illustration: Altair Media (AI-generated concept artwork)
Caption
Artificial intelligence may be digital, but its foundations remain profoundly physical. From energy grids and semiconductor manufacturing to data centres and industrial infrastructure, the future of AI increasingly depends on the systems that make intelligence possible at scale.
