The future of compute is a scaling challenge

Freeke Heijman-te Paske did not present the future of compute as a distant, abstract horizon. She brought it back to a very concrete experience from the Delft quantum ecosystem. She arrives there in her electric car, wants to charge it, and sometimes cannot. Why? Because the power is needed for quantum. It was a small anecdote, but it captured the larger point of her contribution to the Holland High Tech Networking Event: the demand for compute power is growing dramatically, and that growth will not be solved by more of the same.

The future of compute, in her framing, is about new architectures and paradigms: quantum computing, advanced semiconductors, integrated photonics, neuromorphic approaches, and the broader stack around them. These are not fashionable labels. They are responses to a structural problem. AI, digitalisation, simulation, defence, industry and science all require more computing capacity. At the same time, energy use is becoming a limiting factor. “The demand for compute power increases dramatically,” she said. “So we need to innovate to make sure that we can deliver this.”

That makes compute both an economic and a strategic issue. The Netherlands already has a world-class control point in the global semiconductor value chain: ASML. Heijman-te Paske explicitly called that “fantastic” because it gives the country leverage. But one control point is not enough. For the next phase, she argued, the Netherlands needs more technical and strategic positions in the value chain. Compute is not just about chips; it is about the full set of enabling technologies, components, systems, sensors, software and integration capabilities that determine who can build, scale and control the next generation of digital infrastructure.

Growing fast

Her message was therefore not one of concern only. It was also an optimistic assessment of what is already present. The Dutch ecosystem in this field is stronger than it often dares to say out loud. Heijman-te Paske referred to research into a cohort of 65 companies active in areas such as quantum, photonics, neuromorphic computing and advanced semiconductors. These startups and scale-ups are growing fast. She mentioned QuantWare’s recent large Series B round as one visible example of that momentum. The capital attracted by this group is significant, she said, and its share within European tech capital is also growing. “We have something here.”

That sentence could serve as the hinge of her story. The Netherlands has something here: knowledge, companies, talent, infrastructure, and a high-tech ecosystem that already contains many parts of the future compute stack. But having something is not the same as turning it into industrial power. The gap Heijman-te Paske identified is familiar in Dutch deep tech: the early phase works relatively well. Starting companies out of strong research environments is possible. The policy instruments, networks and entrepreneurial culture have improved. But the scaling phase is harder. “From lab to fab, we need to step up our game,” she said.

That phrase gives the story its urgency. Future compute is not waiting for another strategy document. It needs execution. Asked whether she was optimistic that the Netherlands could close the gap, Heijman-te Paske pointed to the growing sense of urgency, the National Technology Strategy and the agendas now being put in place. The direction is clearer than before. The next step, she argued, is to act. Her goal, she said, would be “to stop writing papers” and start executing.

In that sense, her contribution fitted neatly into the broader tone of the Holland High Tech event. Many speakers spoke about acceleration, strategic autonomy and public-private collaboration. Heijman-te Paske translated those themes into the compute domain. If computing is becoming the engine under AI, industry, defence and scientific discovery, then the Netherlands cannot afford to treat quantum, photonics, advanced semiconductors and neuromorphic computing as separate islands. The value lies in the connections.

That was also her final point to the high-tech community in the room. The ecosystem already contains many of the necessary elements. The task is to connect them better. “We need the cables, we need the sensors, AI, the stack,” she said. The call to action is not necessarily to build a new institution, foundation or instrument. It is to “take out the walls” between existing initiatives and create the interconnect.

That is a pragmatic message, but also a demanding one. It asks researchers to think beyond the lab, startups to prepare for industrialisation, established companies to open their supply chains, and policymakers to design instruments that do not stop at the moment a company is founded. Future compute requires more than scientific excellence. It requires manufacturing capacity, energy awareness, supply-chain positioning, system integration and a willingness to scale.

The importance of Heijman-te Paske’s intervention is that she made compute tangible. It is not an invisible layer somewhere behind AI applications. It is the electricity that may prevent a car from charging. It is the ASML-like control point that may or may not exist in the next generation of technology. It is the group of 65 companies that could become a new industrial cluster or remain a promising collection of startups. It is the difference between being good at invention and being strong in production.

For the Netherlands, that difference matters. In a world where computing power becomes a strategic resource, future compute is not just a technology agenda. It is an economic sovereignty agenda. The country has the ingredients, Heijman-te Paske made clear. Now it has to connect them, scale them, and show them. The future of compute will not be won by those who only understand the science. It will be won by those who can turn that science into a working, energy-conscious, strategically relevant industrial ecosystem.