In Milan, Italy has signed technological sovereignty. Too bad everything else is missing.

The first World Tech Conference, organized by Micromegas Comunicazione, concluded at the Allianz MiCo in Milan. Over four days (perhaps a bit too many for a first edition) from June 24 to June 27, a call was made from Milan to the international community for technological progress to be guided by responsibility and cooperation, featuring 220 speakers, 18 panels, and around twenty hours of streaming.

However, the most substantial moment came on the second day, with the signing of the Volta Declaration, the alliance between government, research, and businesses designed to accelerate Italy's race in AI and quantum computing, presented by Undersecretary for Innovation Alessio Butti in front of Henna Virkkunen, Executive Vice President of the European Commission with a mandate for technological sovereignty, as well as security and democracy. The backdrop from the stage was one of great occasions: if two centuries ago, Volta made a new form of energy available, today it is up to us to build the infrastructure for knowledge and computation.

I tried to look beyond the formal phrases, starting from the text of the declaration and from the two interviews I conducted after the Conference, one with Butti and the other with Andrea Lenzi, president of the CNR and among the signatories.

Let's start with the name. The declaration is called Volta because in 2027, it will mark the bicentennial of his death; thus, formally we are approaching the anniversary. Volta, an Italian, invented the battery, and today an Italian alliance has signed on a future where however the most valuable part of the hardware is produced by others.

In reality, the Volta Declaration is a three-year memorandum of understanding open to new parties, signed by the government via Butti, by the CNR of Lenzi, by the National Quantum Science and Technology Institute, by the Centro Nazionale Volta, and by Confindustria with Emanuele Orsini, along with a group of companies including Enel, Almaviva, Sielte, and the Canadian D-Wave. It establishes the Volta Forum, a permanent table with a general secretary, aimed at fostering dialogue between research and industry, training young experts, and taking stock every year of the country’s progress. It is a pact that commits the signatories to meet and coordinate, which is already something, and it is appropriate to call it by its name before treating it as an already active production line.

Butti was the most concrete voice of the day. On artificial intelligence, he acknowledged directly that Italy is destined to lag behind, while regarding quantum computing, he believes we can be among the front runners. Quantum computing, explained simply, is a family of machines that leverage particle physics to tackle problems that regular computers find prohibitive, and for now, it still mostly lives in laboratories. According to Butti, Italy’s bet lies in its people: he says there are about fifteen mathematicians in Europe capable of working at high levels in quantum, and that half are likely to be with us. Quantum costs less than AI in capital and requires more inventiveness, and on this point, Butti is probably right, as it is the only area where the word sovereignty rests on something solid, considering that the raw material here is the skills of people rather than the availability of billions.

Then comes the concrete part, and here Butti was more interesting than he perhaps intended. His reasoning is that hosting technologies is one thing, governing them is another, and that to govern them, data centers need to be located on-site, warehouses full of servers where data is processed and stored. So far, so good. Only that data centers consume a lot of energy, and Butti made it clear: feeding them with solely renewable energy is considered unrealistic by the scientific community, and therefore the government intends to reintroduce next-generation nuclear power, SMRs (Small Modular Reactors), small-size fission reactors that can be built more or less in series and that should generate the 250-300 megawatts needed to operate a computing center without relying on energy from third countries. I give him credit; it’s the kind of thing usually avoided in an interview.

However, there is a passage regarding data centers that made me raise my eyebrows. In recent months, about fifty mayors from the Milan area have asked the region to review the data center law, stating in writing that they are not at all opposed to technological development but want more say on water, energy, and land consumption. Butti framed that type of resistance as a communication problem, a matter of moral suasion, meaning persuasion needed to overcome misoneism, the learned word for fear of the new, and to make people understand that technology is not an enemy but a friend. He reassured that future data centers will be smaller, camouflaged in the environment and without health risks. Those mayors had already expressed in writing that they were not afraid of technology, and what they were asking for were answers on water, electricity, and land consumption. Shifting the discussion from verifiable numbers to intangible sentiments benefits especially those who would have more uncomfortable answers regarding the numbers.

This is also the point where the real costs risks arriving late. The gain, in sovereignty and qualified jobs, is the part that is announced today, from the stage, with the signature and the photograph. The material costs, the energy that these plants burn, and the land they occupy, as well as the effect on the communities that find themselves with one at a highway exit, are the parts that spread over the years in which the infrastructures are built and activated, after the spotlights have long gone dark. It is the time lag between the gains that are immediately accounted for and the costs that materialize later, what in my book I call Temporal Ghost GDP, and data centers are one of the cases where this is most evident.

Professor Lenzi, however, in my interview, takes the matter in another direction, and for the better. He has placed research at the core of everything, the famous last mile that transforms scientific discovery into tools for businesses and citizens, and when it was pointed out that today AI is in the hands of a few American private companies, he responded in a way that made me pause. He said that AIs do not originate from the companies that sell them; they originate from the ideas of researchers, and that while businesses limit themselves to packaging today’s systems, science is already moving beyond AI. It is a reversal of perspective that has a solid foundation, as the physics and mathematics underlying these models are public, stemming from decades of work largely done outside companies, and those who market the products work downstream of all that material, rather than producing it.

Lenzi also downplayed the excitement over generative algorithms, reminding that they remain a tool and often replicate and produce incorrect texts, illustrating it with an image as old as the world: the first researcher was the one who learned to use fire, and fire is good if you use it to cook, but if you use it poorly, it burns your house down. Said by a gentleman who chairs the CNR, after months in which we have read interviews in which AI confesses to wanting to see the sea, to hear that it is merely a tool is almost a breath of fresh air. Saying that it is a tool does not, however, resolve the question from which Lenzi started, that is, who holds the knife by the handle. Even fire is a tool, yet over which and under what rules we have built several millennia of institutions regarding who can ignite it, while the rules around AI's handle are currently being written primarily by the companies that sell it, and the fact that the underlying science is public does not shift who controls the finished product that people use by a whit.

The entire framework of the day revolves around the word sovereignty, sovereign computation, and sovereign infrastructures, with Europe needing to stop merely being a user of technologies and become a designer of them. Then you look at the list of signatories, and among those contributing quantum hardware is D-Wave, which is Canadian, and in the broader circuit of the conference are IonQ and Rigetti, which are American. The same Q-Alliance states that it aims to fill the algorithmic gap that currently exists between quantum machines and the real-world problems of businesses, which is a fancy way of saying that others build the machines, and we try to write the software for them. I do not say this to diminish, as developing competence in quantum software is serious work. I say this because the word sovereignty, repeated continuously, tries to justify a situation in which we have no capacity to decide anything since everything starts from buying someone else's hardware, which raises several questions that no one in the room answered.

The Volta Declaration is a reasonable starting point, and the Volta Forum, if it truly keeps research and businesses at the same table for some years, can serve a purpose. What remains in my hands, as a citizen who might encounter a data center near home and a modular reactor in some valley in a few years, is that the big decisions about where to place infrastructures, how much energy to allocate, and which nuclear to implement under what conditions are being made by the same people who were on stage explaining that technology is a friend. They may well be right. Nonetheless, they remain decisions of industrial and territorial policy, and I would like to see them discussed with the vocabulary of costs and choices, rather than with that of friendship between us and machines. For now, in Milan, the second language has been felt almost exclusively.