Microsoft CEO Nadella: Sovereignty does not depend on data center location

Microsoft CEO Satya Nadella explained in Davos that sovereignty must be thought of differently in the age of AI than before. And he also sees no AI bubble.

listen Print view
Larry Fink in conversation with Satya Nadella

BlackRock CEO Larry Fink (right) in conversation with Microsoft CEO Satya Nadella.

(Image: Screenshot)

3 min. read

Microsoft CEO Satya Nadella currently sees no speculative bubble around artificial intelligence. Speaking at the World Economic Forum in Davos, he explained in conversation with BlackRock CEO Larry Fink that it would be an unmistakable sign of a bubble if everyone were only talking about tech companies and their AI offerings – but not also about the benefits they provide. Examples include medications whose clinical trials could be accelerated with AI. Accordingly, he is confident that AI will reach people on the same track as cloud and mobile devices, drive up the productivity curve, and create local added value.

Admiring the technology itself is not enough, Nadella emphasized. “We will quickly lose social acceptance for using a scarce resource like energy for token generation if these tokens do not improve healthcare, educational outcomes, public sector efficiency, and private sector competitiveness in all areas, whether small or large.” The generation of tokens per dollar per watt must become constantly more efficient, he demanded. Currently, the price is falling approximately every three months. Tokens are the smallest unit of data for a language model.

Regarding Europe and the increasingly loud question of digital sovereignty there, Nadella urged that isolation and protection are not the path to global competitiveness. Data protection and AI regulation are to be welcomed, but one must build locally and think globally. In the age of AI, a company's sovereignty must be understood as much more than sovereignty over its data. A company that can map its data and the implicit knowledge of the organization into the weights of its AI models – that is sovereign.

Where the data center for the services used runs is, in turn, the least important of all factors. Sovereignty can also be well represented through encryption and keys held locally, Nadella believes.

Empfohlener redaktioneller Inhalt

Mit Ihrer Zustimmung wird hier ein externes YouTube-Video (Google Ireland Limited) geladen.

Ich bin damit einverstanden, dass mir externe Inhalte angezeigt werden. Damit können personenbezogene Daten an Drittplattformen (Google Ireland Limited) übermittelt werden. Mehr dazu in unserer Datenschutzerklärung.

However, Microsoft's specific corporate policy on digital sovereignty currently looks different from Nadella's considerations: here, the company relies on the construct of EU data borders, invests in data centers in Germany and the EU, deposits source code in Switzerland, and expands Azure with on-premise functions. However, the company cannot change the fundamental problem that, as a US company, it is subject to the CLOUD Act, which allows US authorities to access data under certain circumstances.

Nadella also praised artificial intelligence as a cognitive enhancer for knowledge workers. The fact that many companies have not yet seen growth from this enhancement so far is explained by the fact that a transition to new ways of thinking is also necessary. Information hierarchies in companies are flattened with AI; managers must rethink workflows and, above all, gain their experience with AI. This is hard work and a lot of change management, especially in large companies.

It is not surprising that Nadella speaks so vehemently for AI – after all, it is Microsoft's biggest future bet. The software giant is a central partner of ChatGPT maker OpenAI and invests billions in AI data centers and their power supply. And in its software and cloud ecosystem, comprehensive AI services are integrated.

(axk)

Don't miss any news – follow us on Facebook, LinkedIn or Mastodon.

This article was originally published in German. It was translated with technical assistance and editorially reviewed before publication.