Tech1 hr ago

AMD's $665M Silo AI Purchase Strengthens Edge as AI Market Nears $1 Trillion

AMD's $665 million acquisition of Finland's Silo AI aims to enhance its AI hardware stack amid a projected big data and AI market of $1.03 trillion by 2030, while Snowflake's Cortex Search claims over 11% gains on OpenAI embeddings.

Alex Mercer/3 min/NG

Senior Tech Correspondent

TweetLinkedIn
AMD's $665M Silo AI Purchase Strengthens Edge as AI Market Nears $1 Trillion
Source: OpenprOriginal source

AMD acquired Finland's Silo AI for $665 million in August 2024 to strengthen its AI hardware stack as the combined big data and AI market approaches $1.03 trillion by 2030. Snowflake's Cortex Search, launched in February 2025, claims to outperform OpenAI's embedding models by more than 11% on several benchmarks.

Context

The big data and artificial intelligence sector is projected to reach $1.0274 trillion by 2030, driven by a compound annual growth rate near 18% as enterprises seek real-time insights and automated analytics. Growth is fueled by the integration of AI into data platforms, rising demand for predictive modeling, and expanding investments in AI-optimized infrastructure. This expansion creates opportunities for chipmakers, cloud providers, and software firms to deliver tighter hardware-software integration.

AMD Acquisition

In August 2024, AMD completed the $665 million acquisition of Silo AI, a Finland-based firm that builds custom AI models tuned to AMD's CPU and GPU architectures. Silo AI's team brings expertise in large-scale model training, optimization, and deployment, allowing AMD to offer end-to-end AI solutions that run efficiently on its silicon. The deal is intended to accelerate AMD's go-to-market strategy for AI workloads in data centers, edge devices, and emerging AI-first platforms.

Snowflake Claim

Snowflake launched Cortex Search in February 2025 as a vector-based retrieval service that can query both structured tables and unstructured files such as PDFs and images. The company states that Cortex Search outperforms OpenAI's embedding models by more than 11% on multiple public benchmarks, including MTEB and BEIR subsets, measuring retrieval accuracy and latency. This performance gain illustrates how specialized retrieval techniques can enhance AI-driven applications without relying solely on larger, general-purpose language models.

What It Means

AMD's acquisition adds dedicated AI software talent that can tighten the coupling between its processors and AI frameworks, potentially improving performance per watt for training and inference tasks. Snowflake's claim signals that purpose-built retrieval systems can narrow the performance gap with generic large language models, influencing enterprise decisions about AI infrastructure stacks. Together, these developments point toward a broader trend of hardware-software co-design, where optimized silicon and tailored software combine to lower latency and cost for data-intensive AI applications.

What to Watch Next

Watch for AMD's upcoming announcements of AI-optimized Instinct accelerators and software stacks that leverage Silo AI's expertise, as competitors respond with their own AI-focused acquisitions and partnerships. Also monitor Snowflake's roadmap for expanding Cortex capabilities beyond search into broader AI agent frameworks and multimodal retrieval.

TweetLinkedIn

More in this thread

Reader notes

Loading comments...