AI Navigator #3: Germany and Europe have lost touch with AI
Now is the time to invest in AI and build an AI infrastructure so that Europe can catch up internationally, says Mirko Ross.
(Bild: Diego Barbieri/ Shutterstock.com)
- Mirko Ross
Welcome to the third edition of the DOAG AI Community's AI Navigator column!
At the Hannover Messe trade fair, corporate PR departments were full of success stories about AI. The reality away from PR wonderlands, however, is bleak. The European and German economy is at risk of being left behind in global competition when it comes to AI. Massive efforts and investments are more necessary than ever to play a serious role.
Because beyond the glamorous world of exaggerated cheerleading and pretty PowerPoint slides for board reports, sober facts show that Europe and the German economy are only bringing up the rear in the global AI race. There is a real danger that German and European companies will be reduced to the role of consumers and service users when it comes to artificial intelligence, instead of playing an active role with powerful AI models.
Dependence in the cloud
In the cloud revolution, the momentum of hyperscalers from overseas has already overtaken the European economic area. European companies have recognized too late the positive effect of massive investments in the provision of virtually unlimited computing resources for the business models of the future.
In the absence of European alternatives, the European and German economies are largely dependent on the dominant US hyperscalers, which siphon off a significant proportion of industrial value creation via usage fees. This pattern threatens to repeat itself with artificial intelligence. It doesn't help to make excuses: the European and German economy is not investing to the necessary extent in the development of AI infrastructures and models that can withstand global competition. Those who do not invest now will soon end up on the bench. Instead of being a player, you become a spectator in the global AI game.
(Bild: DOAG)
On November 20 and 21, 2024, the AI Navigator will enter its second round. The event, organized by DOAG, Heise Medien and de'ge'pol, will once again take place at the Nuremberg Convention Center East. KI Navigator is the conference on the practice of AI in the three areas of IT, business and society. It is dedicated to the concrete application of artificial intelligence.
The program will go online at the end of May. If you would like to get an impression of the conference before then, you can read the follow-up report on KI Navigator 2023.
This paralysis of the European and German industry can be demonstrated by the global orders for powerful AI processors that are required to train current and future models. In 2023, no European company was in the top ten for the Nvidia H100 processor generation. Of the approximately 500,000 processors ordered (as of Q3/2023), Omdia's market researchers list twelve buyers, led by Microsoft and Meta with 150,000 orders each, followed by Google, Amazon, Oracle and Tencent with 50,000 orders each. There are no European or German companies on the list.
Major processor purchases only from the USA
There are also no European companies in the top ranks for the next processor generation, Nvidia Blackwell. According to a leak of confidential documents to Business Insider, Microsoft is planning to double its computing power from 2024 to 2025 just to use this number of ordered processors and is investing 2.9 billion US dollars in Japan and 3.16 billion US dollars in the UK, among other places.
The only logical conclusion from this is that anyone developing powerful AI models also needs processors and computing power. Europe and Germany are effectively the red lantern on the global AI express train. This failure will not only have a creeping negative impact, but disadvantages are already being felt. Major industrial AI applications from Europe are based on models and infrastructures from overseas. As in the case of cloud platforms, infrastructure providers from the USA are siphoning off significant economic benefits generated by artificial intelligence.
Holger Hoos, Professor of Artificial Intelligence Methodology at RWTH Aachen University, expressed his criticism of the growing dominance of US companies in the AI sector to the Tagesschau: "The problem is that the economy and society are becoming critically dependent on these companies. This is particularly sensitive in the field of AI because these technologies will play a key role in all areas of our economy and our global competitiveness will depend directly on them."
Something needs to change now
To avoid falling into the dependency trap again, massive and, above all, rapid efforts are needed. It won't help us if company boards continue to build Potemkin villages for politicians with euphoric reports. What is needed is an unsparing look behind the scenes and a courageous approach by corporations and financial investors in close cooperation with politicians.
A new president will be elected in the USA on November 5 this year. The outcome of this election is open. With the Inflation Reduction Act, the current Biden administration has already proven that it can give the domestic economy a market advantage through protectionism. Economic dominance by US companies is also part of the political agenda of the moderate Biden administration. If Donald J. Trump wins the election, the increased dependence on the AI sector poses an uncontrollable social and economic risk for Europe.
European companies must invest significantly more and more courageously in the development of AI infrastructures and AI models. At least two to three companies from Europe must be among the top ten global buyers of AI processors in Q3 2024. There is still time to secure a seat on the AI express. Let's get on with it!
(mki)
