Together AI Launches GPU Infrastructure in Sweden

ByRicardo

Infrastructure arrives with a collection of AI upskilling workshops, together with a first-of-its-kind Model Shaping occasion in Amsterdam

Together AI at the moment introduced important progress in its European growth with the launch of GPU infrastructure in Sweden and a workshop collection targeted on upskilling AI engineers, beginning with its first-ever Model Shaping workshop in Amsterdam. These initiatives mark a key step in the corporate’s ongoing effort to supply the AI neighborhood with excessive‑efficiency, open, and customizable AI options.

Now Operational in Sweden
Sweden is the latest addition to Together AI’s world AI infrastructure community, powering its serverless inference API for fashions together with OpenAI’s gpt-oss, DeepSeek, Meta Llama, and Qwen. Customers also can request GPU Clusters and Dedicated Endpoints immediately in Sweden, benefiting from low‑latency and safe entry to Northern and Central Europe.

Infrastructure positioned in Sweden delivers these key buyer advantages:

  • Ensuring European regulatory and knowledge residency necessities – with the GPU servers positioned in Sweden, AI firms can efficiently meet necessities round knowledge governance and transparency.
  • Reduced finish‑person latency – Locating inference workloads nearer to finish customers can scale back spherical‑journey community time by 50–70 ms for a lot of Northern and Central European prospects, bettering general response instances for actual‑time AI functions by as much as 25–30%.

“Caesar is constructing a deep‑analysis AI platform designed for data professionals, builders, and establishments working on the frontier of innovation. We use Together Dedicated Endpoints at the moment—an 8×H200 Llama‑4 Maverick deployment—to energy our transformation stage with excessive concurrency and huge context home windows. As we transfer towards public launch, we’re enthusiastic about deploying our workloads in Together AI’s new Sweden area to ship decrease latency and meet in‑area knowledge wants for European prospects. Together AI’s mixture of devoted capability and serverless elasticity lets us scale rapidly as demand grows.”

— Mark McKenzie, Founder, Caesar (caesar.xyz)

Upskill on AI with Free Technical Workshops

Alongside this infrastructure growth, Together AI is launching a collection of AI upskilling workshops for builders. The first occasion, which can happen in Amsterdam on September 10, is about Upgrading and Customizing Open Models.

The workshop will cowl:

  • Tailoring open fashions utilizing publish‑coaching steps comparable to supervised positive‑tuning (SFT) on area‑particular knowledge, choice optimization, and coaching with verifiable rewards to attain efficiency on par with prime proprietary LLMs at decrease value.
  • Custom speculative decoding to speed up inference through the use of a website‑optimized draft mannequin to pre‑generate tokens, delivering speedups of over 1.85× on fashions like DeepSeek‑R1.
  • Model quantization methods to compress massive LLMs, decrease {hardware} necessities, and scale back inference prices, enabling deployment on smaller gadgets.

Click right here to register for Together AI’s AI upskilling workshop.

“Europe is on the frontier of AI innovation, and we’re dedicated to enabling its builders and researchers with the infrastructure and experience they should succeed,” mentioned Vipul Ved Prakash, CEO at Together AI. “Our investments in Sweden and the AI engineering neighborhood in Europe display our dedication to supporting excessive‑efficiency AI that’s each dependable and scalable in Europe.”

Call to Action

  • Attend the Amsterdam Workshop: Upgrading and Customizing Open Models
  • Request a GPU Cluster
  • Request a Dedicated Endpoint for Inference

The publish Together AI Launches GPU Infrastructure in Sweden first appeared on AI-Tech Park.

Similar Posts