Together AI Drives European Growth with New GPU Infrastructure in Sweden
Together AI Drives European Growth with New GPU Infrastructure in Sweden
Together AI, a leader in AI acceleration, has made significant strides in its European expansion by launching GPU infrastructure in Sweden. This new initiative not only enhances their capacity to provide AI services but also aligns with a broader goal of upskilling engineers in the AI community.
Sweden: A New Hub for AI
As part of their strategy, the infrastructure in Sweden is now operational, marking an important milestone in Together AI's commitment to delivering high-performance AI solutions. This facility supports a variety of models including OpenAI’s gpt-oss, DeepSeek, Meta’s Llama, and Qwen. Users in Northern and Central Europe will benefit from improved accessibility, as they can now request dedicated GPU clusters and endpoints directly from this location.
Key Benefits
- - Regulatory Compliance: The establishment of GPU servers in Sweden enables compliance with European data governance and residency requirements. This ensures companies can operate within legal frameworks while utilizing AI technology.
- - Reduced Latency: By positioning the infrastructure closer to end-users, Together AI can significantly decrease round-trip network time, resulting in faster response times for real-time applications. Reports suggest that latency could be reduced by 50 to 70 ms, which is a substantial improvement for businesses reliant on rapid AI interaction.
Mark McKenzie, founder of Caesar, a key client, expressed excitement about deploying workloads in the new region, anticipating enhanced performance with lower latency. He stated, “We use Together Dedicated Endpoints to power our transformation stage with high concurrency. The new infrastructure will allow us to meet in-region data needs better and quicker.”
Workshops for AI Engineers
Complementing the infrastructure launch, Together AI is also focusing on expanding the skills of AI engineers through technical workshops. The first event in this series will occur on September 10 in Amsterdam, covering the topic of upgrading and customizing open models.
Workshop Highlights
- - Post-Training Customization: Attendees will learn how to refine open models using techniques like supervised fine-tuning on domain-specific data and preference optimization. This approach is designed to improve performance metrics while maintaining lower costs compared to proprietary models.
- - Speculative Decoding: The workshop will also introduce custom speculative decoding methods, which can accelerate inference and streamline operations.
- - Model Quantization: Training on model quantization techniques to compress large models, making deployment feasible on smaller devices, is another essential focus of the workshops.
Participation in these workshops aims not only to enhance individual competencies but also to push the boundaries of AI capabilities across Europe. As articulated by Vipul Ved Prakash, CEO of Together AI, “Europe is at the frontier of AI innovation. Our commitment to this market exemplifies our dedication to supporting AI builders with robust infrastructure and expertise.”
Conclusion
The recent developments from Together AI highlight its commitment to both infrastructure development and educational initiatives aimed at the AI community in Europe. As Together AI expands its operations, it continues to play a pivotal role in shaping the landscape of AI technology and development in the European market.
Topics Consumer Technology）
【About Using Articles】
You can freely use the title and article content by linking to the page where the article is posted.
※ Images cannot be used.
【About Links】
Links are free to use.