Italy awards 2D Photonics €211 million to speed up AI data centers

Italy has awarded 2D Photonics a grant of €211 million, or about $249 million, amid growing competition for artificial intelligence infrastructure. The recipient is a company trying to speed up data processing in AI data centers, and the authorities present the decision as part of a broader strategy to support local startups. For the market, this is an important signal: European governments are moving more actively into not only applied AI but also the foundational technologies without which modern models cannot be scaled.

The logic behind the funding is clear. The main problem in AI infrastructure today is not only the number of available accelerators but also how fast data moves inside a computing system. As models grow, requirements for bandwidth, latency and power consumption rise. Against this backdrop, companies offering new ways to speed up data transfer and processing inside data centers have a chance to become a critical link in the whole chain. According to the project description, 2D Photonics is aiming at exactly this niche.

For Italy, this investment is not just support for a single company but an element of industrial policy. The authorities are trying to show that local deep-tech startups can count not only on university grants or limited venture capital but also on large amounts of state money when a strategic technology is at stake. This matters especially for Europe, where dependence on outside suppliers of chips, cloud infrastructure and key components for AI systems has long been debated. If such companies grow inside the region, countries have a better chance of keeping intellectual property, manufacturing and engineering teams at home.

The size of the sum matters too. A €211 million grant is no longer symbolic early-stage support but a resource that could potentially let the company speed up research, bring the technology to industrial level, hire strong specialists and reach deployment noticeably faster. In projects like this, money is needed not only for lab research but also for prototyping, compatibility testing, pilots with data centers and preparation for serial production.

In the AI data center market, where development often takes years and experiments are very expensive, access to this much capital can be a decisive factor. Even if part of the funds goes to a long R&D cycle, the funding itself lowers the project's risk and makes it more visible to partners, equipment makers and future customers.

The decision also shows how governments' view of the AI economy is changing. Not long ago, the main focus was often on applications, chatbots and consumer services. Now the center of gravity is shifting deeper, to the compute base, data transmission networks, power supply, cooling and other layers that determine the real cost and speed of AI systems. These layers often become the bottleneck when demand for compute grows faster than the market can add capacity.

In this sense, betting on a company that promises faster data centers looks pragmatic: the winner will not be whoever merely announces support for AI, but whoever helps make infrastructure cheaper, faster and more scalable. For 2D Photonics, this is a window of opportunity; for Italy, it is a test of its ability to turn public money into a technology asset. If the project delivers a notable result, the country will be able to show that it can grow not only AI service companies but also players in critical infrastructure.

If not, this case will serve as a reminder of how expensive and risky the race for hardware and next-generation technology remains. But one thing is already clear: the fight for AI leadership is waged not only in models but also in the systems those models run on.