Test report DSG-4214 · Rev C · tested November 17, 2026
AI Datacenter InfrastructureDevice under test
US groups set to spend $740bn on AI as Europe's suppliers gain
About $740 billion in planned US AI spending is pulling investors toward European suppliers. Infineon lifted after-tax profit 49% to €1.51 billion as data-center demand boosted its business.
- Read
- 3 min
- Words
- 696
- Node
- 45nm
- Operator
- Grace Kim
Spec summary
- Alphabet, Amazon, Meta and Microsoft are expected to spend about $740 billion on AI infrastructure this year.
- Infineon reported €1.51 billion in after-tax profit for its latest financial year, up 49% year on year.
- Five hyperscalers—Alphabet, Amazon, Meta, Microsoft and Oracle—now operate capital-intensive infrastructure with high fixed costs.
- Three European groups named by Scalable Capital: Hochtief, Siemens Energy and Siemens.

By Grace Kim
Alphabet, Amazon, Meta and Microsoft are set to spend about $740 billion on high-performance chips, data centers and other AI infrastructure this year, according to Pascal Kielkopf, an analyst at asset manager HQ Trust.
AI has driven equity markets for some time, with sharp gains and reversals. Investors now question the cost of the build-out more often. Analysts at Scalable Capital liken the mood to a Wild West gold rush: few prospectors got rich, while the shovel sellers profited.
Who supplies Europe's AI build-out?
Christian Röhl of Scalable Capital, speaking to ARD's finance desk, said: "Without ASML from the Netherlands, or the unlisted German companies Carl Zeiss and Trumpf, TSMC in Taiwan could not produce microchips for Nvidia. Europe is, so to speak, the backbone of the data-center boom."
Stephan Kemper of BNP Paribas called Aixtron a structural beneficiary and said that without European firms and the equipment they supply, not one AI server would run worldwide. Germany's Jenoptik supplies important components, while Aixtron, listed in the MDAX, Germany's mid-cap index, builds equipment for compound semiconductors.
What do the hyperscalers' balance sheets show?
Trillion-scale AI investment is no longer merely announced: it has rapidly reshaped the balance sheets of Alphabet, Amazon, Meta, Microsoft and Oracle, analyst Christof Schürmann of Flossbach von Storch Research Institute wrote.
Schürmann called the five groups' infrastructure commitments staggering and said they will appear in future balance sheets. He said the pace of expansion would have been hard to imagine a few years ago.
The companies, known as hyperscalers because they invest far beyond normal levels to scale future AI revenue, are becoming capital-intensive infrastructure operators with high fixed costs. Flossbach von Storch now describes them as the new heavy industry. Against investor skepticism about those outlays, many European companies already earn money during the build-out, including data-center contractors, electricity-grid operators and semiconductor-equipment makers.
Which European groups are already paid by the boom?
Röhl named three:
- Hochtief, which is active in the United States.
- Siemens Energy, with its power-grid and energy solutions.
- Siemens, whose digital factories connect software and AI with the physical world.
Further down the semiconductor chain, Infineon, a DAX-listed chipmaker whose business originated in automotive electronics, supplies power semiconductors for data-center electricity supply and energy management. Kemper said Infineon's role has grown because Nvidia chips and comparable processors require substantial power, making energy management more important for data centers.
What do Infineon's results show?
Infineon, headquartered in Neubiberg near Munich, earned €1.51 billion after tax in its latest financial year, a 49% increase from a year earlier. Data centers drove the improvement, and the Munich chipmaker expects them to remain a growth driver. Its power semiconductors feed the electricity supply of those facilities.
Will revenue outrun the infrastructure bill?
Many analysts still expect the investment to pay off over the medium term, with expectations for the end of the decade correspondingly high. Kielkopf said uncertainty remains over when and how the sums will earn returns.
He expects additional revenue mainly from cloud services, AI applications and software licenses, while infrastructure costs stay immense.
What changes for investors?
Robert Halver of Baader Bank, in an interview with ARD's finance desk, points to the latest Nasdaq sell-off after the Democrats won the US midterm elections as evidence of how sensitive markets have become to AI investment news. He sees that sensitivity as an argument for heavier allocations to German and European stocks.
"Large pools of capital still describe AI as an attractive long-term sector, but they now want to focus on cyclical businesses, machine builders and suppliers that benefit from AI," Halver said. He calls the move a rotation away from the New Economy toward the Old Economy.
For investors, the case now rests on companies already earning revenue from data-center construction, power grids and semiconductor equipment, rather than on expectations for AI services alone. The main question is timing: infrastructure spending has reached the hyperscalers' balance sheets, while much of the return still depends on cloud, application and licensing revenue arriving later.
via tagesschau.de/xml/rss2/ (Source)
More from Grace Kim
Show full bio
Market editor covering regulation, policy and their impact on hardware layer of artificial intelligence at Die Signal.
665 articles
Same lot · LOT-362D
- DSG-20967nmSWI raises €800 million to fund data center and AI compute expansion
- DSG-98965nmASML Sees Q4 Sales Topping Estimates on AI Chip Demand
- DSG-315565nmSMCI-Backed Argentum AI in Talks for $5 Billion to Buy GPUs
- DSG-57425nmScaleway to build AMD-powered AI supercomputer in Sweden
- DSG-90315nmGlobalFoundries to Produce AI Chip Component for TSMC in $2B Deal