According to BlockBeats, Tendrils has completed a 30000000 USD seed round and is developing a general-purpose processor architecture that supports thousands of cores working in parallel.
Article
💰 Financing | Tendrils completes 30000000 USD seed round, plans to build a general-purpose processor capable of running thousands of cores in parallel
BlockBeats news, November 11, Cambridge semiconductor startup Tendrils completed a 30000000 USD seed round, approximately 26000000 EUR, led by Benchmark, with participation from Visionaries Club, the UK sovereign AI fund, FR8, Antimatter, Cloudberry VC, Conviction, Embassy Ventures, and SV Angels, as well as multiple angel investors. The company was founded in 2025 by CEO Nils Cremer and co-founders T6 and Hugo McNally, and is developing a general-purpose processor architecture that allows thousands of cores to work in parallel.
The new architecture is based on the interaction net computation model, representing programs as interconnected nodes where independent operations can execute simultaneously, attempting to move away from traditional shared mutable memory and linear instruction sequences. The company is also developing an open-source programming language called Vine, with syntax close to Rust, making it easier for software to adapt to this parallelism-first architecture. Tendrils said that GPUs already handle large-scale numerical computation, but irregular tasks such as AI agents executing code and retrieving documents still rely heavily on CPUs. Benchmark general partner and Lattice co-founder Jack Altman will join the board.
Original link https://m.theblockbeats.info/flash/376789
Key points
01
BlockBeats reported on November 11 that Cambridge semiconductor startup Tendrils completed a 30000000 USD seed round, approximately 26000000 EUR.
02
The round was led by Benchmark, with participation from Visionaries Club, the UK Sovereign AI Fund, FR8, Antimatter, Cloudberry VC, Conviction, Embassy Ventures and SV Angels, among others, as well as several angel investors.
03
Tendrils was founded in 2025 by CEO Nils Cremer together with co-founders T6 and Hugo McNally.
04
The company is developing a general-purpose processor architecture that enables thousands of cores to work in parallel, based on the interaction net computational model, which represents programs as interconnected nodes and allows independent operations to execute simultaneously.
05
The company is also developing the open-source programming language Vine, whose syntax is close to Rust, to adapt to this parallelism-first architecture.
06
Tendrils says GPUs already handle large-scale numerical computing, but irregular tasks such as AI agents executing code and retrieving documents still rely heavily on CPUs.
07
Benchmark general partner and Lattice co-founder Jack Altman will join the board.
AI-assisted interpretation
The following is analysis, separate from reported facts. Verify important claims independently.
Tendrils wants processors to handle more independent operations simultaneously: representing programs as interconnected nodes rather than following the traditional shared mutable memory and linear instruction sequence. The accompanying language Vine is used to help software adapt to this design. The original text describes an architecture under development, not a delivered product.
Why it matters to readers
This design targets what Tendrils calls the irregular tasks of AI agents that still rely heavily on CPUs. It proposes processing requirements different from large-scale numerical computing, but the funding news itself cannot prove that the architecture has achieved the expected performance.
Parallelism can be understood as allowing multiple independent operations to proceed simultaneously. When reading this news, one should distinguish between "obtaining funding," "under development," and "performance verified": the original text reports the first two and provides no performance test results.
Risks and unknowns
The evidence provided consists only of media reports, with no independent confirmation materials from the company or investors attached.
The goal of thousands of cores working in parallel is not yet supported by prototype or performance test data.
The original text does not state a product delivery timeline, commercialization arrangements, or specific use of funds.
The judgment that AI agent tasks still rely heavily on CPUs comes from Tendrils, and the materials provide no independent verification.
Related Developments
Loading event timeline…
Related concepts
Seed round funding
This term is not in the glossary yet. Browse related concepts in the glossary.
