Voltar às notícias
IndústriaInstruções AI Understanding

Euclyd levanta US$ 230 milhões para um sistema de chip de inferência de IA

Pluang relata que a startup holandesa Euclyd levantou US$ 230 milhões da Samsung e de fundos de investimento para desenvolver um sistema de chip de inferência de IA com foco em energia e custos, destinado a competir com GPUs Nvidia.

4 min readRead the linked source
Source-provided image accompanying Euclyd raises $230 million for an AI inference chip system
Referência de fonteFonte registrada
Editora
pluang.com
Link da fonte
pluang.comhttps://pluang.com/en/news-feed/samsung-dukung-pesaing-chip-ai-nvidia-dengan-pendanaan-230-juta
Tipo de fonte
Fonte vinculada — o status da fonte primária não foi estabelecido.
ContextoEntenda isso em 60 segundos

Comece aqui

Termos-chave

Inferência
A fase de tempo de execução em que um modelo treinado gera previsões ou resultados.
Memória (memória do agente)
Contexto armazenado que um agente de IA usa em etapas ou sessões para melhorar a continuidade.
Referência
Um teste padronizado ou conjunto de dados usado para medir e comparar o desempenho do modelo.
Teste você mesmoQuestionário explicado sobre modelos de IA

O que aconteceu

Pluang reports that Dutch AI-chip startup Euclyd raised $230 million in a funding round co-led by Samsung and several investment funds. Euclyd is developing an alternative AI chip architecture, with physical systems planned for 2028 and a goal of serving thousands of enterprise customers by 2030.

Pluang reports that Dutch startup Euclyd has raised $230 million in a funding round co-led by Samsung and several investment funds. The company, founded in 2024, is designing an AI chip system with an architecture intended to reduce energy consumption and costs in AI data centers. Pluang says Euclyd plans to launch physical chip systems by 2028 and serve thousands of enterprise customers by 2030.

The source describes Samsung’s involvement as providing memory-manufacturing expertise and supply-chain support. It does not identify the other investors, disclose the financing structure or valuation, name customers, describe production arrangements, or provide a public primary document. These details have not been independently confirmed here.

Detalhes da fonte: pluang.com ↗

Por que isso importa

The reported round would make Euclyd a significant new entrant in the effort to diversify AI data-center hardware beyond Nvidia GPUs. A successful lower-energy, lower-cost system could affect the economics of deploying AI at scale, but the source provides no independent performance, efficiency, customer, valuation, or production evidence. Samsung’s reported participation may provide manufacturing and supply-chain expertise, although the exact scope of its involvement is unknown.

AI is the stage at which trained models generate responses or predictions, and it is increasingly driving data-center demand. Pluang’s report indicates that Euclyd is targeting a practical constraint in AI deployment: the cost and energy required to run models at scale. If the company’s approach works in production, it could give cloud and enterprise operators another hardware option and increase competitive pressure on incumbent accelerator suppliers.

The reported Samsung connection could matter because advanced AI chips depend on reliable memory and manufacturing supply chains. However, the source does not establish that Samsung will manufacture Euclyd’s chips, guarantee capacity, or provide commercial distribution. No independent shows that Euclyd’s design is more efficient or less expensive than Nvidia-based systems.

Interactive Mechanism

Mecanismo interativo: como realmente funciona

Explore a tecnologia subjacente a este desenvolvimento de forma interativa.

Document Size:128K tokens
Needle Placement Depth (Location in document):50% into text
Attention Context Buffer Map:
Target Fact (50%)
Equivalent Pages~320Standard book pages
Retrieval Accuracy99.9%Needle recall score
RAM / KV Cache5.1 GBMemory overhead
Prompt CachingActive~80% discount on reuse
Core takeaway: Million-token context windows allow querying whole codebases or legal archives in one prompt. However, KV cache memory scales with context length, making prompt caching crucial for real-time production.
Verificação de conceito interativo+10 Points
AI Models Explained Quiz

Which component of an AI application is the machine-learning model itself?

O que assistir a seguir

Key unknowns include Euclyd’s chip architecture, manufacturing partner, technical benchmarks, customer commitments, funding terms, valuation, and route to commercial production. The 2028 launch and 2030 enterprise-customer targets are plans reported by Pluang, not demonstrated milestones. Independent testing and evidence of working silicon will be needed to assess whether Euclyd can deliver a credible alternative to established AI accelerators.

The most useful next evidence will be a disclosed architecture, working silicon, independent performance and energy benchmarks, named design or manufacturing partners, and confirmed customers. Euclyd’s reported 2028 product target remains a future plan, and the 2030 goal of serving thousands of enterprises is a company ambition rather than a verified forecast.

The funding amount and Samsung’s co-lead role are reported by Pluang. The source does not provide independent confirmation, financing terms, valuation, product pricing, availability, or evidence that the chip system has entered production. Those unknowns limit what can currently be concluded about Euclyd’s competitiveness.

Guias e questionários relacionados

Modelos de IA explicadosTreinamento de IAFuturo da IATeste o que você sabe – experimente um teste gratuito de IAProcure um termo de IA em nosso glossárioSiga o rastreador de financiamento de IA
Achou isso útil?