The “Roadmaps to new nuclear” conference held by the Nuclear Energy Agency in Paris a couple of weeks ago saw the participation of over 300 delegates from countries, companies and institutions all united in the effort to stimulate the efforts and increase the potential to realize the Nuclear Renaissance the world is hoping for.
A large part of the debates was dedicated to the difficult issue of financing, and the conclusion, reiterated by both the public and the private sector, was that the current situation does not allow the development of nuclear programs without heavy commitments (including financial) from the public hand. All the speakers insisted on the fact that there needs to be long-term commitment assurances as well as large public money support in order to ensure that a national nuclear program could be properly funded.
This all sounds quite reasonable at first glance. After all, how can private capital be involved and galvanized without political certainties? Or without efficient administrative and bureaucratical responses from the authorities deputized to bring along the licensing, the certifications and all the other steps needed in order to complete construction and commissioning of a nuclear installation?
However, if the same logic was applied to Artificial Intelligence, it is safe to say that most of the advances the industry has made in the last 15 years would probably not have seen the light:
It is not a secret that most AI companies have lost and keep losing tons of money, while their stock evaluation not only prospers but seems not to end in its setting of new records
It is not a secret either that it is these sky-rocketing evaluations that allow AI companies to invest, to develop new applications, to nurture and cultivate talents, to thrive.
And it is not a secret that the size of this money-moving activity is so large that AI companies are now even entering – with their own logic – the basic infrastructural business of data center construction and O&M, a business far nearer to nuclear reactors’ construction and O&M.
How does AI do it? How have AI players been capable of mobilizing these sums of money with no public intervention? Is there one or two lessons to be learned here for the nuclear industry? We believe so.
It’s innovation, stupid !
First of all, there is the theme of innovation. The whole AI’s narrative dwells in the idea that new ideas, unconventional approaches, not by-the-book logics are good and should all be tried, because in the end that is how progress and technological advances happen. It is not by chance that it is the SMR’s and Advanced Reactor’s approach that is allowing nuclear today to be front news again. But investors do not look only at innovation in the product, they also want to see the single companies in the value-chain to invest in innovating their own internal processes, in optimizing their structures, in a word, in becoming full-stack AI companies. And in this respect the nuclear industry has a very long journey in perspective: how many nuclear companies have a C-level AI/Innovation Manager? How many invest in new administrative tools? How many look at ways to improve their manufacturing processes by integrating AI tools? (we talk about AI here, not mere digitalization) How many think about Infrastructure, and not just about the reactor itself?
But innovation should not only be pursued, it should also be communicated and valued by leveraging on the specific value-added features and functionalities it brings to a specific industry. While some aspects of industrial innovation are generic and always apply, others should be elaborated by looking at the peculiar status of the nuclear industry. For example
it is possible to demonstrate how AI increase compliance with safety procedures by adding automatic checks on top of human checks;
AI can considerably speed up workforce certification and on-the-job training, especially in the knowledge-transfer space;
and last but not least, innovation can be a powerful element in making an industry attractive to younger generations looking for career indications.
These communications elements are not felt as vital as long as an industry relies solely on public spending; they become fundamental and even shape their own specialists once an industry recognizes the need for them. Maybe it is time for the nuclear industry to do just that.
Merging of strategic infrastructures
All of the above are relatively short-term indications and rely essentially on the willingness for single players in the nuclear industry to engage in innovative ways of thinking. The larger picture, where AI and nuclear can really unite for their common interest, is in beginning to conceive the investment costs to launch a nuclear program as part of the AI’s infrastructural budgeting. This change in perspective, the unification of nuclear and AI under the same strategic-type investment umbrella, can narratively be a powerful force for the acceptance of nuclear as a source of energy production.
At the same time, this merging of interests would open the possibility for nuclear programs, or parts of it, to be financed with funds available for datacenters deployments. In part this is already happening by private initiative, where large datacenter players through private PPAs activate their purchasing powers to finance nuclear initiatives. Rather than direct public investments, it would be wise for the nuclear industry to require public incentivization in this direction. This may take various forms in different countries or continents, of course.
“AI for nuclear and nuclear for AI” (the theme of this blog) is not just a slogan, and neither a mere question of technological cooperation. In our vision it is a strategic convergence of interests which could (and should) result in a complete resetting of rules for both industries, possibly even giving birth to a new race of companies, for a world and an economy both AI-driven and nuclear-powered.
