In April 2026, Anthropic is exploring the development of its own AI chips—marking a new phase in the global race for AI infrastructure. The U.S. AI company wants to reduce its reliance on players like Nvidia, Google, and Amazon, turning up the pressure on the existing chip supply chain. For the Netherlands, and ASML in particular, the stakes are high.
Why does Anthropic want to build its own AI chips?
Anthropic’s main goal is to be less dependent on Big Tech. According to Reuters sources, the plan is still in its early days, with no finalized design or dedicated team yet.
The core reason is simple: there’s a global shortage of high-performance AI processors. Models like Claude demand massive compute, and companies are fiercely competing for hardware access.
Key drivers:
- Less dependence on Nvidia, which currently dominates the market
- Lower costs for training and inference
- More control over performance and optimization
- Strategic autonomy in a geopolitically sensitive sector
For now, Anthropic still uses a mix of chips from Google (TPUs), Amazon, and Nvidia. At the same time, it has already inked deals for massive compute capacity via Google Cloud.
What does this mean for the AI chip market?
Anthropic’s move underscores a broader shift: AI companies are increasingly verticalizing—building their own hardware, software, and infrastructure.
We’re already seeing this at:
- OpenAI (in-house chip ambitions and mega-compute deals)
- Meta (developing its own AI accelerators)
- Google (with TPU chips)
The impact is significant:
- Nvidia could lose market power over the long term
- New chip architectures tailored to AI workloads will emerge
- Cloud providers may become less dominant as intermediaries
The AI market is shifting from “software-first” to “compute-first.”
What’s the impact on ASML and the Netherlands?
Anthropic’s plans directly matter for ASML. The Dutch company supplies the machines required to manufacture advanced chips.
ASML stands to benefit from this trend:
- Demand for AI chips is exploding
- New entrants are boosting chip production needs
- Advanced EUV machines remain essential
ASML has already said AI is the primary growth engine for chip demand.
On top of that, the latest High-NA EUV machines are crucial for next-generation AI chips, enabling more powerful and efficient designs.
In short:
More AI companies + more in-house chips = more demand for ASML machines
Why this is geopolitically sensitive
Building in-house chips also taps into rising geopolitical tensions. The U.S. has imposed curbs on exporting advanced chips to China, making hardware access strategically vital.
For AI companies, this means:
- In-house chips = less exposure to export controls
- More control over supply chains
- Faster innovation without external constraints
As a result, AI is evolving from a software industry into a strategic one—more like energy or defense.
What does this mean for the Netherlands?
For the Netherlands, it’s a double-edged opportunity.
Opportunities:
- ASML cements its role as a key player
- The Dutch chip ecosystem scales with it
- More investment in European AI infrastructure
Risks:
- Rising geopolitical pressure on ASML
- Dependence on international customers
- Intensifying technological competition
The Netherlands is, quite literally, at the heart of the AI revolution.
Conclusion: AI companies want control over their future
Anthropic’s chip ambitions show AI companies no longer want to depend on existing infrastructure. They’re building their own foundations.
That leads to:
- More competition in the chip market
- A bigger role for European players like ASML
- A power shift across the tech sector
The AI race isn’t just about models—it’s about who controls the hardware.
Popular news
Apple Parts Ways with Tim Cook: What It Means for AI
AI “Grief Tech” Goes Mass Market: $3 Device Recreates the Dead
Anthropic faces backlash as powerful AI tool leaks beyond control
Dutch Companies Are Losing Their Grip on AI Data and Governance
Mozilla finds 271 Firefox bugs using AI tool Mythos from Anthropic
Meta pushes boundaries with AI training on employees, EU faces privacy dilemma
Latest comments
Loading