The Event

On September 8, humanity woke up to what appeared to be one fewer Millennium Prize Problem left to solve. OpenAI—which counts Microsoft among its shareholders and key technology partners—announced that one of its internal models had demonstrated the formation of a singularity in the three-dimensional Navier–Stokes equations.

The result was reached after approximately 88 hours of coordinated work involving around 10,000 artificial intelligence agents. OpenAI claims to have shown that an initially smooth solution can develop a singularity in finite time. The proof still needs to be examined and accepted by the mathematical community. Separately, the Clay Mathematics Institute requires at least two years after publication before considering a proposed solution for its prize.

Therefore, we cannot yet consider the problem definitively closed. We can, however, consider it an event that raises an important question: what might this change for engineering?

Figures reported by OpenAI in its September 8 announcement.

Will It Change Fluid Dynamics as We Know It?

This is probably one of the first questions that comes to an engineer’s mind. The immediate answer, however, is simple: no.

The Millennium Prize Problem asks whether an initially smooth solution will always remain smooth or whether it can develop a singularity. An industrial simulation addresses a different question: how a fluid behaves around a particular geometry or inside a system under a given set of conditions.

Answering that question requires engineers to define the geometry, mesh, fluid properties, boundary conditions and turbulence models, among many other parameters. The new proof does not provide a universal formula—the holy grail of fluid dynamics. Instead, it constructs a mathematical case in which the equations develop a singularity.

Solving a fundamental property of the equations is not the same as solving every problem described by them.

So What Does This Mean for Industry?

The tools engineers use every day to predict fluid behaviour will not change immediately. We will continue to use RANS, LES and other models. Mesh quality will remain decisive, and validation through wind tunnels, test rigs and physical prototypes will still be necessary.

Something else, however, may already have changed: the way research is organized.

A problem whose mathematical foundations can be traced back to Jean Leray’s work in 1934 has been approached through a combination of massive exploration, human direction and formal verification. While thousands of AI agents investigated different strategies, human researchers connected the most promising results, guided the different lines of work and checked the consistency of the proposed solution.

This model could transform engineering research. AI agents may eventually explore thousands of concepts, compare alternatives, optimize designs, identify errors and propose experiments. Engineers would then be responsible for formulating the right problems, establishing their physical limits, identifying which proposals have real value and taking responsibility for the final decision.

The transformation will not be immediate. The process required extraordinary computing capacity and models that are not yet publicly available. Nevertheless, the precedent has now been established.

An Engineer’s Perspective

Personally, I believe this event confirms something we have been witnessing for several years: artificial intelligence is here to stay, and its future is as frightening as it is exciting.

Beyond the possible resolution of a problem that had remained open for almost a century—and that none of the great minds of our recent history had managed to close—this development represents a turning point in the way we work. Today, nobody in my circle—or at least nobody I know—approaches certain tasks without consulting AI, asking it to review their work or using it to generate new ideas.

We are entering a period in which AI will stop being an occasional tool and become another participant in the chain of work. It will not replace an engineer’s judgment or responsibility, but it will change how we explore alternatives, analyze problems and develop solutions.

I firmly believe that this transformation can be highly beneficial for industry if it is used wisely. Collaboration between artificial intelligence, formal verification and human judgment represents the future of industrial development.

For better or worse, our future is written in ones and zeros.

A European Factory, a Global Car

On September 16, Alfredo Altavilla, an adviser to Chinese manufacturer BYD in Europe, said that the brand would need three assembly plants and one battery plant on the continent in the long term. This expansion is not a new phenomenon. For years, we have seen Chinese manufacturers become part of the European automotive industry, whether through new facilities, such as BYD’s in Hungary; bringing former factories back into use, as Chery is doing with EBRO at Nissan’s former plant in Barcelona; or acquiring European brands, as Geely did with Volvo Cars.

This trend raises the question of what a car needs to be considered truly European. If it employs European engineers, is made with Austrian steel and is assembled in Hungary, is it automatically European? Or is it automatically Chinese because the company belongs to a group from that country? To answer that, we need to distinguish where value is created, who controls the technology and where decisions are made. Above all, we need to think about the consequences of this new direction.

What Lies Behind the Label

According to the IEA, China accounted for more than 80% of global battery cell production in 2025. This helps explain why the location of an assembly plant tells only part of the story. Behind every vehicle are design, intellectual property, material selection, component manufacturing, software and production processes. Companies and engineers from different countries can contribute at every stage.

It is therefore difficult to draw a line that determines how European a car is. Should every stage carry the same weight? Is manufacturing more important than design? These are questions many of us might ask when thinking about this issue.

In my view, different contributions do not carry the same value in every product. If good design reduces costs while maintaining quality, that creates significant value. If material selection improves durability, that decision matters too. If manufacturing ensures tight tolerances and consistent quality, production expertise is essential. There is no single answer, and manufacturing well is not simply a matter of following instructions.

That is why, beyond a brand’s passport, what matters to me is the philosophy a company follows and the standards it is willing to uphold. That does not determine a product’s nationality, but it does shape much of what I expect from it. And I think it is worth asking what we want to preserve about our approach to the automobile.

What Europe Stands to Gain

Beyond whether a product is European, it is important to consider what value it brings us. The arrival of these manufacturers can create jobs, attract investment, open opportunities for suppliers and encourage training and the exchange of knowledge. Nor would it be fair to assume that every foreign company comes here only to assemble cars: in 2025, BYD announced an R&D centre in Budapest, with electrification projects and university partnerships. We will have to see how those commitments take shape.

But it is not all good news. The jobs created at one plant do not tell the whole story if activity is being lost elsewhere at the same time. And manufacturing here does not guarantee that we control the technologies, component supplies or decisions that will shape that factory’s future.

In my view, the risk lies in becoming increasingly dependent on companies whose decisions we have little influence over. I worry that we may celebrate the arrival of production while losing the ability to develop our own products. Having factories matters, but so does being able to decide what they make and taking part in developing the next generation.

Competition from more affordable alternatives can benefit consumers and push European manufacturers to improve. What concerns me, however, is how they respond to that pressure. If the answer is to cut back on things that are hard to judge in a showroom—durability, repairability or the quality of certain components—to keep prices attractive and equipment eye-catching, I think we will be taking the wrong path.

I do not assume that a Chinese car is worse or that a European one is better. What I reject is a way of competing in which we end up accepting lower quality because ‘it is fine for the price’, or in which screens and lights matter more than handling, efficiency and reliability. And I would make the same criticism of a European brand if it took that path.

Personally, that is not the car industry I want for the future. I want technology to improve cars in the ways that really matter, and competitive pricing to come from doing things better, without making durability an afterthought.

The Next Generation Test

With all this in mind, we need to think about where we want to go. That question is also ours to answer as consumers, even if the decisions of executives and policymakers carry a weight that our choices in a showroom cannot replace.

I understand that price is decisive for many people. Not everyone can afford to pay more for a car, and a familiar brand does not deserve a blank cheque either. But I also think we should ask what we are comparing when we look at two prices: how long we expect the car to last, how we will be able to repair it, what service we will receive and which qualities we truly value. The same standards should apply to every brand.

Perhaps both models can coexist if competition is fair and investment helps build capabilities here. For me, a good test will be the role our engineers and suppliers play when it is time to create the next vehicle. Will we have learned to do more ourselves, or will we still depend on decisions and solutions developed elsewhere?

Perhaps the future will bring affordable, modular cars that look increasingly alike and have less personality. Perhaps we will recover some of the qualities I miss: simplicity, intuitive controls and cars designed to last and be repaired. I do not think everything was better fifteen years ago, but neither do I think every change is progress.

I do not have a crystal ball, but I know which qualities I want tomorrow’s car to keep. What about you?

How Energy-Independent Is Spain?

When we think about energy, we tend to look at the light switch at home. Yet a country’s dependence also lies in the fuel that moves its transport, powers its factories and heats its buildings.

In 2024, renewables generated 56.8% of Spain’s electricity. That same year, our dependence on energy from abroad stood at 68.4%, and petroleum products accounted for 46.5% of primary energy consumption. These are different indicators: celebrating the first does not make the other two disappear.

Using progress in renewables as sufficient proof of independence is a political oversimplification. Positive figures deserve recognition, but selecting only those that offer something to boast about turns a strategic debate into propaganda.

What Opportunities Does the Transition Offer Spain?

Spain added around 10 GW of wind and solar photovoltaic capacity in 2025, excluding self-consumption installations, and ended its fourth consecutive year as a net electricity exporter. The opportunity is there. The problem is mistaking that progress for a finished job.

An annual export surplus does not guarantee self-sufficiency at every moment. Nor do a few hours of wholesale prices close to zero guarantee cheap electricity when we need it. The system must respond when solar output falls, wind is scarce or demand rises.

That requires complementary generation, grids, storage and the ability to manage consumption. The April 2025 blackout, with voltage-control failures and cascading disconnections, left a lesson: having generation available is not enough to guarantee stability.

My criticism is simple: we celebrate installed megawatts, but we should demand the same rigour of the system that integrates them. Bringing new capacity online is not the same as planning well.

Are We Removing Dependencies or Replacing Them?

Wind and sunshine are resources of our own. The equipment needed to harness them may not be.

According to the IEA, using 2024 reference data, China accounted for approximately 85% of manufacturing capacity across the solar supply chain and 80% across the lithium-ion battery supply chain. For photovoltaic wafers, the figure was around 95%. Such industrial concentration is hard to ignore when we talk about autonomy.

This is not the same dependence as that on fuels: an installed panel keeps generating even if new equipment stops arriving; a gas-fired power plant needs a continuous fuel supply. The improvement is real, but we remain exposed when it comes to expanding, repairing and renewing installations.

That is why I find it misleading to present renewables as an automatic guarantee of independence. We are reducing some vulnerabilities while retaining others. Having the natural resource does not mean mastering the technology or being able to replace whoever sells it to us.

What Should We Demand?

Technical judgement should set limits on political promises. No technology deserves a blank cheque, and none should be dismissed simply because it sits uncomfortably with a political party’s narrative.

We should demand reliable supply, affordable costs and a demonstrable reduction in our exposure to external dependence. That requires investment, European cooperation and capabilities of our own. It also requires accountability: calling a project ‘strategic’ or ‘sustainable’ does not prove that it is well conceived.

I want a Spain that is less exposed to decisions made elsewhere. But autonomy is not announced: it is built. As long as we confuse generation records with energy independence, we will be selling as an achievement what is still unfinished work.