In recent years, the European space industry has frequently discussed supply chain security, which sounds like it wants to pursue "deglobalization," but the anxiety behind it is actually very specific.
Kate Underhill of the European Space Agency told a small story in Budapest that best illustrates the problem: they wanted to order 20 laser diodes from a German company, and the reply was that the minimum order quantity was 10,000.
Between 20 and 10,000, there is a difference of 500 times.
This is not a price issue, but a business logic issue.
Space projects do not launch many times a year, and the number of parts used each time is extremely small. Suppliers are unwilling to open a dedicated line for such fragmented demand.
Consumer electronics, automobiles, and AI data centers are the big customers of semiconductor companies, while space orders are at the bottom of the list.
The European Space Agency itself admits that space systems are usually "at least 10 years behind consumer electronics."
The reason for this backwardness is that the market is too small.
Satellite launch volumes are low, and buyers are scattered among national space agencies, research institutes, and commercial companies. Each buyer purchases only a little. Faced with such customers, suppliers either quote sky-high prices or simply refuse. The European Space Agency tries to solve this through an "inward" strategy, applying high-volume commercial off-the-shelf electronic equipment to space, while also seeking "outward" opportunities to bring space-tested designs into the commercial ground market.
But this path is slow because space certification cycles are too long, and the commercial market cannot wait.
What is more troublesome is export control. Underhill said: "If your satellite contains any American components, you must comply with American regulations."
This is the extraterritorial jurisdiction of the International Traffic in Arms Regulations.
European satellite manufacturers want to build a satellite not subject to American jurisdiction, but certain specialized components can only be purchased from the United States or China.
As a result, part of Europe's satellite sovereignty is in Washington's hands.
The Real Weakness of European Space
The European Space Agency's response is to identify components still purchased from the United States or China and see what can be done in Europe.
This approach is correct, but it is extremely difficult to implement.
Because the space supply chain is not a line, but a network. If you replace one American part, you may uncover three new dependencies. If you replace one Chinese part, you may find that Europe has no substitute at all.
The real problem is that Europe's space demand is too fragmented.
National space agencies each go their own way, procurement standards are not unified, and certification systems are not mutually recognized.
When a German company gets an order from the French space agency, it may have to redo certification. This fragmentation puts Europe in a passive position in the entire supply chain. Suppliers know you have no alternative, so they dare to quote a minimum order quantity of 10,000. If you had an order for 800 satellites, the situation would be completely different.
This is the significance of the IRIS² project.
Europe plans to build a secure satellite constellation, changing the scale from "I want to build one satellite" to "I want to build 800." Underhill said the order quantity will reach a semi-industrial level, and it will be possible to begin leveraging economies of scale.
800 satellites cannot solve all problems, but at least they can make suppliers sit down and talk.
When your order is large enough, the minimum order quantity is no longer an obstacle, and prices will return to a reasonable range.
The European Space Agency is also promoting the application of artificial intelligence in design, testing, and manufacturing. Artificial intelligence allows engineers to run complex simulations and reduce the need for physical testing.
Digital twin models of rocket engines can monitor health status and predict service life.
During 3D printing, artificial intelligence checks each layer to ensure parts are strong and defect-free. These technologies cannot directly solve supply chain problems, but they can shorten certification cycles, reduce manufacturing costs, and make the European space industry more competitive.
Launch cost is another key variable. Europe's Ariane 5 rocket retired in July 2023. At its peak, it performed six to seven missions per year, with a cost of about $10,000 per kilogram of payload.
SpaceX plans to reduce launch costs to $30 per kilogram through the fully reusable Starship. This is simply not competition on the same level.
The European Space Agency has set a goal of launching 9 to 10 Ariane 6 rockets per year starting in 2027. The goal may not sound too high, but it is already 1.5 times the peak period of Ariane 5.
Underhill said Europe needs to build the space transportation industry like the aviation industry, achieving rapid turnaround, predictive maintenance, and frequent launches. Traditional launch vehicle construction methods cannot do this, and it must be rethought. The essence of this idea is to transform space from a "project-based" model into an "industrialized" one.
Project-based means each launch is unique. Industrialized means standardization, repeatability, and scale. Only industrialization can reduce costs, only cost reduction can expand the market, and only market expansion can gain supply chain bargaining power.
Europe's regulatory environment is stricter than that of the United States, which may slow project progress in the short term. Underhill mentioned the EU's proposed PFAS restrictions. These synthetic chemicals persist in the environment but are still necessary under extreme aerospace conditions.
She explained that it is impossible to build cryogenic systems without using PFAS in seals.
The European Space Agency is working with regulators to seek temporary exemptions while funding research and development of environmentally friendly alternatives.
The development of PFAS-free seals will take five to seven years.
In the long term, this regulatory approach is very likely to become an advantage. Europe's strict standards in environment, labor, and safety help build the industry's long-term strength. If Europe takes the lead in setting standards for green space, space debris mitigation, and safety certification, it can turn regulation into a non-tariff barrier and a tool for exporting standards.
The United States has fewer rules and moves faster in the short term, but once European standards become international standards, they can create competitive advantages for European companies.
RISC-V: Trading Open Standards for the Right to Exit
The shift in processor architecture in the European space industry is the best case for understanding supply chain sovereignty. The European Space Agency chose the SPARC architecture about thirty years ago because it was open and had well-developed tools. Jiri Gaisler, co-founder of Frontgrade Gaisler, received a scholarship in 1997 and developed an independent VHDL SPARC processor model for space applications, later known as the LEON processor.
This was the first VHDL high-level chip, released as an open-source project by the European Space Agency.
The LEON processor became a standard component of the European aerospace industry, spawning a collaborative ecosystem in which multiple suppliers could manufacture chips based on the same core IP.
This model was very successful, but there was one problem: there are fewer and fewer engineers for the SPARC architecture, and upgrading to 64-bit requires huge investment.
Habinc said: "SPARC will be used forever, because once you use something in space, you will use it forever. But we had a 32-bit machine at the time. Upgrading to 64-bit was a huge investment for us."
Switching to RISC-V solved this problem. The open standard foundation of the RISC-V 64-bit architecture is readily available, and implementing 64-bit support is free.
Habinc pointed out that by switching to RISC-V, the company bypassed many development obstacles. RISC-V also allows engineers to remove outdated components and tailor processors according to the strict size, weight, and power constraints of space missions, as well as the needs of emerging applications such as neural networks.
Frontgrade Gaisler developed a processor that supports both architectures, called NOEL, which is LEON spelled backward. This processor can choose to run in either SPARC mode or RISC-V mode at startup, while having all legacy support and being backward compatible with all the code people have written.
Habinc explained that a company already using the old chip can easily start using the new chip directly. In the future, when they find they cannot find SPARC engineers, they only need to switch software, without replacing the circuit board or casing.
The key to this approach is keeping the ecosystem open and avoiding restrictive licenses.
Frontgrade Gaisler not only sells its fault-tolerant technology, but also released the basic processor code as open source. Habinc said: "We ensure that everything we do is compatible with RISC-V. We have no vendor lock-in. If you work with us now and find another vendor in the future, we will not stop you."
RISC-V has also been widely adopted in the United States. In 2022, Microchip received a NASA contract to manufacture the next generation of high-performance spaceflight computing processors. This new PIC64-HPSC is a processor with eight SiFive RISC-V cores and modern features such as Ethernet and PCI Express.
Tia-Martina Gauthier, project architect at NASA's Johnson Space Center, emphasized the importance of this new hardware.
She said NASA really likes the PowerPC-based RAD750 and often conducts flight tests on it, but needs to advance general computing capabilities.
Deep space missions require greater autonomy to process data on onboard equipment, rather than transmitting all data back to Earth through slow connections.
The new HPSC integrates high-performance computing, a 240 Gbit network switch, and powerful cybersecurity features into a reliable system. Gauthier emphasized the importance of using an open instruction set.
She said the PIC64-HPSC is designed entirely based on open standards.
RISC-V will help scientific missions reduce costs and accelerate the transmission of valuable data.
Andrea Gallo, CEO of RISC-V International, said at the 2026 European Summit that RISC-V is approaching a major commercial growth phase.
According to SHD Group forecasts, RISC-V will reach a 33.7% market share across all hardware sectors by 2031.
This growth is particularly significant in edge computing and data centers, with market sizes expected to reach $45 billion and $70 billion respectively by the end of this decade.
Therefore, the formal approval of the RISC-V Server Platform Specification 1.0 is a key moment. This progress standardizes hardware, bringing industry-standard boot systems and runtime services directly to RISC-V, ensuring that system software can run smoothly on different server hardware.
2026 is being called "the year of RVA silicon," with many companies launching new server-class processors.
SiFive's Performance P870D has up to 128 cores, and Akeana's Alpine test chip and NextSilicon's Arbel server-class CPU are also being launched.
Epic Semi launched its Contrail AIX, a superchip combining 32 RISC-V processor cores with 16 built-in AI cores, reaching up to 75 TOPS.
For hyperscale cloud service providers and data center operators, RISC-V is a powerful alternative to proprietary architectures, effectively reducing the risk of single-vendor lock-in. This market is very large, and everyone has an opportunity. So it is not that one side is squeezing the other, but that there is enough room for RISC-V to continue growing in the market.
For multinational companies and governments committed to achieving digital sovereignty, having choices is very important. RISC-V brings freedom of choice and freedom from a single vendor.
In the space sector, RISC-V is becoming a key component of next-generation space computers. At the end of 2025, a dedicated RISC-V Space Special Interest Group was established, chaired by representatives from the European Space Agency and E4 Computing.
The group brings together experts from NASA, Microchip, SiFive, and Frontgrade Gaisler to develop strict standards and white papers for specialized space missions.
NASA is working with Microchip and SiFive to test a high-performance spaceflight processor. The European Commission's COSMIC7 project is developing a 7nm RISC-V chip designed specifically for orbital operations. For established aerospace suppliers like Frontgrade Gaisler, they are moving from the SPARC-based LEON processor to the new RISC-V-based NOEL chip, and the open nature of RISC-V is the main attraction.
Aerospace groups need complete transparency in their hardware to obtain safety certification, because there must be public specifications for them to own the product and control their own destiny. RISC-V is the only alternative for natural evolution.
Europe's positioning on RISC-V is essentially trading open standards for the right to exit the supply chain.
SPARC is open but its ecosystem is aging; PowerPC is reliable but controlled by the United States; ARM and x86 are proprietary with strong vendor lock-in. RISC-V open standards allow European companies to audit, customize, and replace suppliers without being held hostage by a single vendor.
This is a strategic choice.
Europe does not have advanced process hegemony, but it can establish a voice in RISC-V space standards, radiation-hardened IP, and certification testing. The RISC-V Space Special Interest Group is chaired by representatives from ESA and E4 Computing, which is a contest for standard-setting power.
The risk is that if the United States extends export controls to RISC-V cores or EDA tools, open standards could also be weaponized. Europe must plan ahead for open-source EDA and European foundries.
But overall, RISC-V is Europe's smartest move in the space supply chain, because it trades openness for sovereignty and standards for security.
The foundation of the space economy is being rewritten
The European space industry seeks supply chain control, but ultimately it still has to answer an economic question: what is the underlying logic of the space economy? In the past, space was project-based. Each launch was unique, costly, long-cycle, and the market was small. Now space is becoming industrialized: standardized, repeatable, and scalable.
The core drivers of this transformation are falling launch costs and increasing in-orbit computing power.
Ariane 5's $10,000 per kilogram and Starship's $30 per kilogram differ by more than 300 times. When launch costs drop to $30 per kilogram, almost all assumptions of the space economy must be rewritten. Satellites can be made larger, heavier, and more complex because launch cost is no longer the main constraint.
Constellations can deploy more satellites because the launch cost of a single satellite can be negligible.
In-orbit servicing, manufacturing, and data centers all become feasible because the cost of sending things into orbit is greatly reduced.
SpaceX is taking everything to new heights, Habinc said, these companies are lowering launch costs and deploying large networks for communication and computing. They will build their own data centers and artificial intelligence in space. Habinc also pointed out that SpaceX covers everything from chips to final services. This vertically integrated model allows SpaceX to control the entire supply chain, from processor design to launch services, to in-orbit operations.
The Terafab project is the ultimate expression of this vertical integration, with Intel partnering with Tesla, SpaceX, and xAI to build a semiconductor manufacturing plant in Austin, Texas, with about 20% of capacity allocated to AI5 and AI6 architectures, processors focused on edge inference and energy efficiency, supporting Tesla's autonomous driving system, Cybercab fleet, and Optimus humanoid robot projects.
The remaining 80% of computing output is dedicated to space applications, particularly the D3 chip.
This radiation-hardened processor is designed for deployment in SpaceX satellite constellations, consistent with xAI's purpose of using the vacuum of space for thermal management and continuous solar power to establish orbital data centers.
The scale of this vertical integration is staggering, analysts at Yole Group calculated that achieving 1 terawatt of capacity would require $5 trillion to $13 trillion in capital expenditure over the project lifecycle, and noted that this would require processing 22.4 million advanced logic wafers per year.
Memory demand is equally staggering, to meet Terafab's needs, the eventual wafer volume required will reach 10 million wafers per month, for memory alone, equivalent to 5 times current DRAM production.
However, such vertical integration also has limits, ASML delivered only 48 EUV systems in 2025, and its order book is fully allocated to TSMC, Samsung, and Intel through 2027.
Terafab lacks confirmed ASML orders, and as a new market entrant, it faces procurement challenges, rather than building a brand-new fab from scratch and waiting years for equipment, Terafab can leverage Intel's existing 18A process technology, equipment allocation, and packaging infrastructure at its fabs.
In other words, Terafab is more likely an Intel fab expansion project with Tesla, SpaceX, and xAI as anchor customers, rather than an independent manufacturing enterprise.
Another bottleneck I think is probably in memory manufacturing.
I am currently skeptical about Europe independently conducting memory production.
Terafab is still years away from producing anything. Memory manufacturing requires completely different expertise, which Tesla, SpaceX, or Intel do not have. I doubt Terafab can produce memory on its own without cooperation from existing players. The emergence of concentrated demand from Terafab adds a demand signal to an already tight memory market, the real pricing risk is not Musk building his own DRAM, but the overcapacity that could eventually emerge when all these new fabs come online around 2028 to 2029.
There is also the issue of raw material dependence.
Establishing the facility in Texas provides specific supply chain advantages in terms of raw materials, particularly high-purity helium. Helium is a coolant used in advanced lithography and packaging processes.
Asian foundries rely heavily on imports from the Gulf Cooperation Council region, and that supply line experienced production disruptions in early 2026.
The United States produces a significant proportion of the world's helium from domestic reserves in Texas, Wyoming, Kansas, and Oklahoma. Building a plant in Austin might help with helium, since the U.S. produces nearly half of the world's helium, however, bare silicon wafers mainly come from Asian suppliers, and semiconductor manufacturing at such a scale also has non-trivial water and electricity demands.
This competition is unquestionably already changing the landscape of the contract foundry market.
Tesla currently purchases semiconductors from TSMC and Samsung. The Terafab project introduces a domestic alternative.
Stephen Ezell of the Information Technology and Innovation Foundation analyzed: "This announcement is a very significant win for Intel, as it is trying to build a complementary foundry business beyond its core chip manufacturing business and compete with industry leaders like TSMC. This will enhance Intel's ability to manufacture chips focused on AI and mobile applications."
Although the project may not pose a direct technical challenge to TSMC's 70% market share, it brings specific complications for Samsung Electronics.
Some market researchers point out that Samsung may receive a surge in orders from Tesla during the construction of the Texas plant to fill the supply gap. However, once Terafab is fully operational, Samsung could lose its high-volume core customer for advanced process nodes.
For Intel, the agreement serves as an anchor contract for its external foundry services currently under development.
The strategic value of this agreement for Intel is that Terafab is first and foremost a breakthrough for Intel's foundry business, Terafab now adds Tesla, SpaceX, and xAI as further anchor customers, this brings Intel's foundry business the credibility it has been pursuing for years.
Europe cannot replicate the Terafab model, the $5 trillion to $13 trillion in capital expenditure far exceeds Europe's fiscal capacity.
Europe's correct strategy is asset-light sovereignty: not building full-stack fabs, but anchoring mature-node foundry with IRIS² orders, reducing lock-in with RISC-V open architecture, controlling quality with ESA certification, and providing long-term capital with the European Investment Bank and pensions.
Financially, American space is shifting from project financing to platform financing.
SpaceX IPO, xAI merger, Terafab, all package space, AI, and robotics into infrastructure assets, Europe lacks this venture capital culture, but can use public procurement to create stable cash flow, turning IRIS² into securitizable long-term contracts.
In the next decade, the global space supply chain will diverge into three poles.
The United States is vertically integrated, capital-intensive, government equity plus commercial space, represented by SpaceX, Terafab, NASA HPSC.
China is state-led, self-controllable, system-integrated.
Europe is open standards alliance, rules certification, demand aggregation, represented by the RISC-V Space Special Interest Group, IRIS², Galileo, Ariane 6.
Europe's success criterion is not the localization rate, but the ability to switch to a second supplier within 18 months when the U.S. cuts off supply; to switch architecture through RISC-V when vendor lock-in occurs; and to protect its market with its own certification system when standards fragment.
On the investment map, I recommend focusing on four lines: RISC-V space IP and radiation-hardened chips, represented by the Frontgrade Gaisler, SiFive, Microchip ecosystem; satellite direct-to-device and 6G non-terrestrial networks, represented by the IRIS² supply chain, laser communications, ground stations; AI design tools and digital twins, reducing physical testing and shortening certification cycles; PFAS alternatives and green space materials, including seals, propellants, thermal control materials.
So, I believe the European space industry seeking stronger supply chain control is ultimately about building a substitutability sovereignty system.
It acknowledges its weakness in physical production capacity, and instead competes for architectural standards, rules certification, and demand aggregation rights, if Europe succeeds, its space supply chain will not be the cheapest, nor the fastest, but it will have the most second options.
In a world where supply cutoffs become the norm, second options are more precious than first-rate efficiency.
This article is from the WeChat public account: Dongzhen Shanglue , author: Dongzhen Shanglue
