Europe sets its sights on multi-billion-euro gigawatt AI factories
2025-07-29 05:02:29
Data storage tapes are stored at the NESC Scientific Research Center (NERSC) in the Lawrence Berkeley National Laboratory, which will include the United States super computer to be run by Vera Robin coming from NVIDIA, in Berkeley, California, May 29, 2025.
Manuel Orbigoso Reuters
Europe puts its attention on Gigawatt factories in an attempt to enhance the late artificial intelligence industry and meet the rapid challenges of the sector.
A sensation about the concept of factories that manufacture manufacturing by Amnesty International has gained ground in recent months, especially Nafidia CEO Jensen Huang stressed the importance of infrastructure at the June event. Huang praised a new “industrial revolution” at the GTC conference in Paris, France, and He said His company was working to help countries build artificial intelligence factories through Partnerships In France, Italy and the United Kingdom
For its part, the European Union describes factories as a “dynamic ecosystem” that combines computing power, data and talents to create artificial intelligence models and applications.
The bloc has always been behind the United States and China in the race to expand the scope of artificial intelligence. With 27 members of the Federation, the region is slower in acting when it comes to agreeing on new legislation. High energy costs, and allow delay and network in the urgent need to modernize, obstruct developments.
Henna Virkkunen, CEO of the European Commission for the European Commission, told the Tec Sovereignty, CNBC that the goal of the mass is to combine high -quality data groups, the ability of computing and researchers, all in one place.
“We have, for example, by 30 % of the individual researchers from the United States, focus on artificial intelligence. We also have about 7,000 startups [that] They develop artificial intelligence, but the main obstacle for them is that they have a very limited computing capacity. This is why we decided, along with our member states, invested in this very decisive infrastructure. “
These are very large investments because they are four times stronger when it comes to computing capabilities from the largest artificial intelligence factories.
Henna Ferkconine
Executive Vice President of the European Commission for Technical Excellency
“We have all that is required to be able to compete in this sector, but at the same time we want to build our technological sovereignty and competitiveness.”
To date, the European Union has set 10 billion euros (11.8 billion dollars) to finance to establish 13 factories of artificial intelligence and 20 billion euros as a starting point for investment in GigaFactories, which represents what he says is “the largest public investment in artificial intelligence in the world.” Virkkunen said that the bloc has already received 76 expression of interest in Gigafactories from 16 member states across 60 sites.
Virkkunen pointed out that the call to pay attention to Gigafactories was “overwhelming”, exceeding the expectations of the bloc. However, in order for factories to add a worthy addition to the ability of computing in Europe, more investment from the private sector will be needed to finance expensive infrastructure.
“Intelligence Revolution”
The European Union describes the establishments as a “one comprehensive store” for artificial intelligence companies. They aim to reflect the process that is carried out in industrial factories, which transforms raw materials into goods and services. With the AI factory, raw data enter intoput, advanced artificial intelligence products are the expected result.
It is essentially a data center with an additional infrastructure related to how to adopt technology, according to Andre Kukhnin, UBS stock research analyst.
“The idea is to create a graphics processing unit [graphics processing units] “The capacity is, so creating data centers mainly with graphics processing units that can train forms and run inference … then create an infrastructure that allows you to make this available for small and medium -sized companies and parties that you will not be able to go and build.”
Martin Wilki, CITI research analyst.
“You are creating a platform by owning these chips, which have crazy levels of account,” he said. “And if you connect it to a network capable of getting strength to actually use it in full power, the world is at your feet. You have this tremendous ability to do something, but what is its success, it will be defined by what you use for it.”
Telecom and Wireless Communications Company Telenor Possible cases of use of such facilities are already exploring the launch of the AI factory in Norway in November last year. The company currently has a small set of graphics processing units that operate, as it looks at the market test before expansion.
Senior Innovation officials at Telenor and head
Telenor
“The journey began with faith – NVIDIA had a belief that every country needed to produce its own intelligence,” the chief innovation in Telenor and head of the Ai Factory Kaaren Hilsen Factory told CNBC.
Helson emphasized that the database is the key. “If you want to use artificial intelligence to innovate and make business more efficiently, you will likely put decisive and sensitive works for business in these artificial intelligence models.”
The company works with Babelspeak, which Hilsen described as a Norwegian version of ChatGPT. Technology translates sensitive dialogues, such as its pilot with the border police, which cannot use public translation services due to security problems.
“We are facing a” intelligence revolution “where sovereign intelligence factories can really help society’s progress.”
One billion euros investments
Virkkunen said that the first Amnesty International factory in the region to work in the coming weeks, with one of the largest projects in Munich, Germany was launched in the early days of September. It is a different story from Gigafactories.
“These are very large investments because they are four times stronger when it comes to computing capabilities from the largest artificial intelligence factories, and this means billions in investments. Each of these needs three to five billion [euros] In the investment, the Commissioner said, adding that the bloc will look forward to creating a consortium from the partners and then opened an invitation to invest in later this year.
Berin Martins, an older research colleague in Brug, asked why such investments needed support by government funds.
He said: “We do not yet know the amount of private investment that was proposed as a supplement to support taxpayers, and what is the ability and how much are these factories. This is still very clear at this stage, so it is very difficult to determine how much it will add in terms of computing capacity.”
Energy consumption is also a major problem. Martins noted that the construction of GigaFactory from artificial intelligence may take from two to two years – but the construction of power generation of this size requires more time.
“If you want to build a large group of husbands with hundreds of thousands of NVIDIA chips, you must rely on energy consumption at least from Gigawat for one of these factories. If there is enough space in the electricity network in Europe in all these countries to create these factories still must be seen … this will require a great investment in the power to distribute energy,” he told CNBC.
UBS expects that The current 85 GB of 85 GB WAM will have the need for a high demand. Based on the investment of the European Union of 20 billion euros and the plan of each factory to operate 100,000 advanced processors, UBS estimates that each factory may range between 100 and 150 megawatts with a total capacity of all facilities of about 1.5-2 GB.
This can add about 15 % to the total ability of Europe – a big batch, even when compared to the United States, which currently possesses about a third of the global capacity, according to data.
After announcing the trade framework of the European Union, European Union President Ursula von der Layen He said On Sunday, American artificial intelligence chips will help operate Gigafactories from AI in an attempt to help the states “maintain the advantage of technology”.
“One can argue that it is relatively easy, provided that you have money. It is relatively easy to buy chips from NVIDIA and create these hardware factories, but to make them work and make them viable economically, a completely different question,” Martins told CNBC.
He said that the European Union is likely to start on a smaller scale, as the region is unable to build its border models immediately in artificial intelligence because of its expenses.
“I think in time, Europe can build its infrastructure and business models on artificial intelligence to reach this stage, but this will not happen immediately,” said Martins.
https://image.cnbcfm.com/api/v1/image/108152481-17486125992025-05-29t230728z_945976731_rc2yreayezu6_rtrmadp_0_usa-nvidia.jpeg?v=1748612626&w=1920&h=1080
Post Comment