A significant redevelopment at Bull's historic manufacturing site in Angers, France, reflects the company's commitment to scaling European supercomputing production and reducing reliance on overseas supply chains. The €80 million investment has enabled Bull to establish new production capabilities while simultaneously preparing to extend its industrial infrastructure to emerging technology companies across the continent.
Bull, a specialist in high-performance computing, artificial intelligence and quantum technologies, operates across 32 countries with 3,000 employees and generates approximately €720 million in annual revenue. The company maintains a portfolio of around 1,600 patents and provides hardware, software and services to businesses, research institutions and government bodies worldwide.
Inside Bull's new Angers manufacturing facility
The Angers site currently employs nearly 300 people, with plans to recruit between 50 and 100 additional staff as part of a broader hiring initiative targeting 500 new positions. Production capacity has doubled from six computing racks per week to twelve, with Bull's Chief Executive Emmanuel Le Roux indicating that output could eventually reach 24 racks weekly.
The manufacturing environment differs markedly from traditional industrial assembly lines. The production floor comprises three distinct zones dedicated to high-performance computing, artificial intelligence infrastructure, and centralized testing and validation. Rather than crowded assembly lines, technicians work at individual stations performing precision assembly tasks using intelligent screwdrivers that automatically calibrate torque settings and guide the assembly process, minimizing the risk of fastening errors or component damage.

The workforce combines individuals with diverse academic and technical backgrounds, with many positions prioritizing manual dexterity and fine motor skills over formal credentials. The facility's modular design enables production lines to adapt to different equipment configurations within approximately ten minutes, while relocating an entire production line requires roughly three days.
The expanded facility now permits simultaneous assembly of multiple large-scale supercomputers, a capability previously unavailable. Current projects include France's 94-rack Alice Recoque supercomputer and Finland's €388 million LUMI-AI system, which represents Bull's largest contract to date.
The redevelopment incorporates sustainability measures including more than 3,500 photovoltaic panels generating 1,500 kWp of peak capacity and a heat-recovery system designed to eventually capture sufficient energy to serve approximately 600 households. Bull intends to plant approximately 1,500 trees adjacent to the facility. The project received €1.2 million from France's Green Fund, €213,000 from ADEME's Heat Fund and €1.5 million from the Pays de la Loire Regional Council.
A cornerstone technology at Bull is its proprietary Direct Liquid Cooling system, which circulates water through cooling plates attached to heat-generating components within supercomputers. This approach permits water temperatures up to 40°C, substantially reducing energy demands for refrigeration and cooling fans compared to conventional air-cooling methods. Bull's latest generation operates entirely without fans, enabling greater computing density while facilitating waste heat capture for reuse in adjacent facilities.

Building European sovereignty
Europe must have the industrial capabilities required to produce tomorrow's strategic digital infrastructures on its own territory.
Emmanuel Le Roux, Bull CEO
Bull is actively reducing dependence on non-European suppliers through multiple initiatives. The company has developed its proprietary BXI high-performance interconnect technology, maintains a stake in European chip designer SiPearl and now sources electronic boards manufactured within France. According to Le Roux, these boards were sourced entirely outside Europe a decade ago. He also highlighted the Angers facility's uninterrupted operation during the COVID-19 pandemic as evidence of the strategic value of maintaining local manufacturing capacity.
European-manufactured components now represent approximately 70 per cent of systems produced at Angers, compared with 20–30 per cent five years earlier. Le Roux stated that the forthcoming LUMI-AI supercomputer will incorporate 70 per cent European-made components, compared with 40 per cent for its predecessor and approximately 5 per cent for competing systems. These figures have not undergone independent verification.
For Le Roux, technological autonomy involves maintaining alternatives rather than eliminating international suppliers. He described a situation where a shortage of a US-sourced electronic component costing less than €2 threatened to halt supercomputer production. Bull's engineering team redesigned the circuit board to accommodate an Infineon component, enabling manufacturing to resume.
The investment arrives as European institutions work to expand computing infrastructure. The European Union has allocated approximately €7 billion to the European High Performance Computing Joint Undertaking over 2021–2027. According to Le Roux, Bull has secured 15 of 18 EuroHPC tenders.
An Open Factory for European startups
Bull has unveiled an Open Factory initiative designed to make its manufacturing capabilities accessible to other technology enterprises. The programme targets European technology development and aims to assist companies in transitioning from laboratory research and prototype stages to industrial-scale production.
Petr Bednařík, founder of Czech artificial intelligence company DataSentics and now Bull's AI Strategy Director, discussed the potential implications for emerging companies. For smaller organizations developing artificial intelligence hardware or high-performance computing technologies, Bednařík views the Angers facility as a potential solution to the substantial expenses associated with constructing proprietary manufacturing infrastructure.
The factory is a unique asset.
Petr Bednařík
Bull's value proposition for startups extends beyond physical infrastructure. A specialized startup might develop applications leveraging Bull's software platforms or incorporate its technology into existing projects for large enterprise customers, thereby accessing business relationships that would otherwise prove difficult to establish independently. Bednařík is also investigating approaches to provide engineering resources as a form of investment.
For example, we might assign three engineers to work with a company for two years, with that contribution forming part of our investment.
Petr Bednařík
Providing money isn't necessarily the most distinctive thing we can offer.
Petr Bednařík
Bull does not intend to operate as a conventional venture capital firm. For early-stage deeptech enterprises, technical expertise and engineering capacity frequently prove as valuable as financial resources. The investment model and factory-access arrangements remain under development, with Bednařík indicating that Bull is progressing through business planning and board discussions. The concept could be finalized by the end of 2026, with outreach to startups commencing in early 2027. Software startups interested in developing applications on Bull's existing platforms may already contact the company.
Why DataSentics turned Bull away five times
Bednařík's perspective on startup engagement derives from his experience building and exiting DataSentics. Established in Prague in 2016, DataSentics creates artificial intelligence and machine learning solutions for enterprise customers. French information technology group Atos acquired the company in 2021, subsequently operating it under the Eviden brand before Bull assumed ownership in 2026.
Prior to founding DataSentics, Bednařík spent ten years at Deloitte. He sought to establish a more nimble organization, yet scaling as a relatively small Czech technology supplier proved challenging, particularly when engaging major banking institutions. Bull and Atos approached the company multiple times regarding acquisition. "To be honest, we turned them away five times," he recalled.
Eventually, Bednařík determined that joining a larger organization could provide DataSentics access to customers it struggled to reach independently. However, he harbored concerns about preserving the company's culture and autonomy.
As a startup founder, you hear all the horror stories about acquisitions. A company gets bought, and within a year or two, the original team has disappeared, and much of the value has been lost.
Petr Bednařík
According to Bednařík, Bull largely preserved DataSentics' team composition and operational practices, concentrating integration efforts on sales, enterprise relationships and technology roadmap alignment. He acknowledged that maintaining relative independence can complicate financial reporting and administrative processes, yet believes this approach enabled the company to retain its personnel and entrepreneurial character.
Even today, my team operates much as it did when we were an independent startup.
Petr Bednařík
This experience persuaded him that startups could benefit from corporate resources without necessarily undergoing acquisition. He references startup programmes operated by hyperscalers and Databricks as examples of organizations supporting businesses that build on their technologies, and perceives an opportunity to develop a comparable model centered on European industrial infrastructure.
Why Europe's AI factories need more than GPUs
At the factory inauguration, Sebastian von Alfthan, Director of CSC – IT Centre for Science, addressed the scarcity of computing capacity available to European researchers and businesses, and the necessity of making advanced infrastructure accessible beyond major technology corporations. CSC manages Finland's research computing infrastructure and operates LUMI, the EuroHPC supercomputer inaugurated in 2022.
What we lack is sufficient computing capacity to support their ambitions.
Sebastian von Alfthan
LUMI ranked third globally upon its launch. Although initially developed primarily for high-performance computing, its graphics processing unit resources have also supported artificial intelligence workloads. Its successor, LUMI-AI, is among the systems scheduled for production at Angers.
Von Alfthan explained that an artificial intelligence factory encompasses more than a data centre filled with graphics processing units. It integrates computing resources, datasets, storage, technical assistance, regulatory guidance and business development support. This proves particularly significant for smaller enterprises that may lack personnel or expertise to effectively utilize advanced computing systems.
He also explored the potential to merge conventional simulation with artificial intelligence in scientific research and, eventually, quantum computing. Applications included employing artificial intelligence to examine agricultural data and evaluate wood quality in industrial production, illustrating how advanced computing can benefit established industries alongside emerging technologies.
From French state ownership to new European contracts
The French government finalized its acquisition of Bull from Atos in March 2026, establishing the company with a long-term public shareholder as France endeavors to reinforce its computing sector. Bull has subsequently announced multiple commercial agreements throughout Europe. In November, it executed a memorandum of understanding with Schwarz Digits, the digital division of Schwarz Group, to investigate developing European artificial intelligence and cloud infrastructure. The proposed partnership encompasses Bull-manufactured servers, direct liquid cooling and availability of the Bull Sequana AI platform through the STACKIT Marketplace.
In October, Germany's Climate Computing Centre awarded Bull a €40 million contract, including five years of support, for its next-generation HLRE-5 climate supercomputer. Scheduled to commence operation in early 2028, the BullSequana XH3500-based system is engineered to execute two global climate simulations simultaneously at 5 km resolution while reducing energy consumption per simulated year by approximately 75 per cent compared with DKRZ's existing Levante system. The supercomputer will employ Bull's direct liquid cooling technology.
DKRZ already distributes recovered heat to the University of Hamburg's adjacent chemistry department and intends to make additional waste heat available to university facilities upon HLRE-5 installation.
Bull's expanded Angers facility and proposed Open Factory programme present meaningful opportunities for European deeptech startups, particularly those confronting the substantial expenses of advancing from prototype to production. The challenge ahead involves translating these ambitions into actionable support that startups can genuinely leverage.
Source: Tech.eu



