Dresden-based semiconductor startup FMC has secured €100 million to scale its low-power memory chips for AI data centres — one of the largest funding rounds in Europe’s memory segment to date, and a notable milestone as the EU pushes to reduce dependence on US and Asian suppliers.
The round includes €77 million in equity from an oversubscribed Series C led by HV Capital and the DeepTech & Climate Fonds (DTCF), with participation from Vsquared Ventures and returning backers eCAPITAL, Bosch Ventures, Air Liquide Venture Capital, M Ventures (Merck), and Verve Ventures. A further €23 million comes from public funding, including the IPCEI ME/CT programme and the European Innovation Council.
Tackling AI’s growing energy appetite
FMC is developing a new class of memory chips based on hafnium oxide thin films, packaged under its DRAM+ and 3D-CACHE+ technologies. The company claims these chips offer faster performance, lower cost and — crucially — dramatically reduced energy consumption compared to conventional DRAM.
“With bandwidth long dominating the AI conversation, energy efficiency is now becoming the decisive metric,” CEO Thomas Rückes said. “We’re building memory that is not only faster and cheaper, but far more sustainable.”
As AI data centres expand globally, analysts warn they could soon account for a significant share of total energy production. FMC argues that most of today’s power drain comes from constant data movement between fast, volatile memory and slower storage layers. By replacing these volatile tiers with persistent memory, the company says it can cut data-transfer energy and boost system-level efficiency by more than 100% in some workloads.
A standout round in Europe’s 2025 chip boom
FMC’s raise lands amid a broader uptick in European semiconductor funding in 2025. Other notable rounds reported this year include:
- Q.ANT (Germany) — €62M Series A for photonic processors (July 2025)
- Arago (France) — €22.1M Seed for light-powered AI chips (July 2025)
- Scintil Photonics (France) — €50M Series B for integrated photonics (September 2025)
Together, these rounds represent around €134 million flowing into Europe’s semiconductor and AI-hardware ecosystem — and FMC’s €100 million now ranks among the largest deep-tech investments of the year, particularly in memory technologies where Europe has historically had almost no domestic presence.
“Memory chips are the main bottleneck in the AI stack,” Rückes said. “Securing an equity round of this magnitude highlights the importance of solving this challenge within Europe.”
A shot at European memory independence
Today, global memory manufacturing is dominated by South Korea, the US and Taiwan, with China rapidly scaling up. Europe, meanwhile, has lacked a meaningful foothold in this strategic layer of the semiconductor supply chain.
FMC, founded in 2016 and emerging from the Silicon Saxony cluster, positions itself as one of the first credible European players attempting to close this gap. The startup recently transitioned into a fabless model, focusing on design and development while relying on external foundries for manufacturing.
Investors say FMC’s tech is now proving itself in real-world pilots. “Design wins with leading OEMs are already in place,” said Paul-Josef Patt, Managing Partner at eCAPITAL. “FMC shows that European DeepTech can deliver — and even take the lead — in next-generation memory.”
What comes next
The funding will accelerate FMC’s push to commercialize its DRAM+ and 3D-CACHE+ product lines and expand its presence in global markets. HV Capital partner Fabian Gruner called the technology “unique” and said it has the potential to “redefine global industry standards.”
DTCF’s Torsten Löffler added that the company’s technology is strategically aligned with strengthening European semiconductor sovereignty: “By tackling the growing energy needs of AI infrastructure, FMC’s memory technology enables more efficient computing — and it’s made in Germany.”
FMC previously appeared in the 2021 ranking of “59 Top Semiconductor Startups and Companies in Germany.”




































