#Sweden #Amsterdam #Together AI #GPU Infrastructure #AI Workshops

Together AI Expands European Reach with GPU Infrastructure in Sweden

Together AI, a leader in the AI acceleration cloud space, has made a noteworthy stride in its European expansion by launching GPU infrastructure in Sweden. This move is crucial as it enhances the performance and accessibility of AI solutions across the continent. Alongside this development, Together AI will also host a series of AI-focused workshops, starting with its inaugural Model Shaping workshop scheduled in Amsterdam.

New Infrastructure in Sweden

The newly established infrastructure will leverage Together AI’s serverless inference API to support various models, including gpt-oss, DeepSeek, Meta Llama, and Qwen from OpenAI. Customers in Northern and Central Europe can now directly request GPU clusters and dedicated endpoints in Sweden, which will significantly improve their services with reduced latency and increased reliability.

Key Advantages for Clients

The Swedish infrastructure offers several key benefits tailored for AI businesses:
1. Regulatory Compliance: By housing GPU servers in Sweden, Together AI enables AI companies to meet European data governance and transparency standards.
2. Latency Reduction: By minimizing the distance between inference workloads and end-users, the solution can cut round-trip network times by 50 to 70 ms for many customers. This improvement translates to an overall increase in real-time AI application response times by 25-30%.

Mark McKenzie, founder of Caesar (caesar.xyz), noted, "We are utilizing the dedicated endpoints from Together AI to enhance our transformation with high concurrency and extensive context windows. With the new Together AI region in Sweden, we anticipate minimized latency and optimized data needs for our European clients. The combination of dedicated capacity and Together AI's serverless elasticity allows us to scale rapidly with growing demand."

Skill Enhancement Workshops

In tandem with expanding infrastructure, Together AI aims to bolster the skills of developers by organizing technical workshops focused on AI. The first event, happening on September 10th in Amsterdam, is set to cover upgrading and customizing open models. This workshop will delve into:

• - Adapting open models through post-training steps, such as supervised fine-tuning (SFT) with domain-specific data and optimization techniques.
• - Custom speculative decoding for faster inference, enhancing existing models like DeepSeek R1 by over 1.85 times.
• - Techniques for quantifying models to compress large LLMs, ultimately decreasing hardware requirements and inference costs, making them viable for smaller devices.

Interested participants can register for the Together AI training workshop online.

VIP Vipul Ved Prakash, CEO of Together AI, stated, "Europe is at the forefront of AI innovation, and we are dedicated to providing developers and researchers with the infrastructure and expertise needed for success. Our investments in Sweden demonstrate our commitment to promoting high-performance, reliable, and scalable AI throughout Europe."

Call to Action

This expansion calls upon developers and organizations to:

• - Attend the Amsterdam workshop on Upgrading and Customizing Open Models.
• - Request a GPU Cluster for their projects.
• - Seek a Dedicated Endpoint for Inference tailored to their needs.

About Together AI

Together AI is a leading cloud provider specializing in AI acceleration, enabling developers and AI-native companies to train, refine, and execute inferences for generative AI models. The platform supports a wide array of top open-source and customized models across multiple modalities while ensuring high privacy and security standards with flexible deployment options. Committed to advancing AI through open collaboration, innovation, and transparency, Together AI ensures that powerful AI systems remain accessible and adaptable, yielding optimal results for society.