Poolside is an AI-focused private company, and this weekly recap highlights its latest strategic and product developments. Over the past week, the company emphasized its commitment to scalable, open AI infrastructure and continued to advance its Laguna model line for secure, enterprise-grade adoption.
Co-founder and Co-CEO Eiso Kant is set to speak on two panels at the RAISE Summit, focusing on training at scale and the role of open source in AI innovation. This participation underscores Poolside’s positioning in high-performance AI systems and its advocacy for open ecosystems that support developers.
The summit involvement highlights the importance Poolside places on robust infrastructure, rigorous evaluation, and scalable training as key enablers of AI progress. By engaging with leading researchers and infrastructure providers, the company is reinforcing its role within the broader AI ecosystem.
On the product side, Poolside recently introduced Laguna XS 2.1, a lightweight upgrade to its Laguna XS.2 model aimed at improving multilingual coding and terminal-style tasks. The model is designed for easy local deployment, reflecting the firm’s focus on flexible, on-premise and edge-friendly AI solutions.
Laguna XS 2.1 supports multiple frameworks, including vLLM, SGLang, NVIDIA TensorRT-LLM, HF Transformers, and Ollama. It is also offered in several checkpoint formats such as BF16, FP8, INT4, and NVFP4, allowing users to optimize performance and cost across diverse hardware setups.
Poolside also released DFlash speculator models that, based on internal tests, can roughly double tokens per second in local inference. This focus on higher throughput and lower latency is aimed at enterprises and developers who prioritize responsiveness, privacy, and reduced reliance on cloud-based inference.
Distribution of Laguna XS 2.1 spans Hugging Face, OpenRouter, the Poolside API, Ollama, and pool, the company’s terminal-based coding agent. This multi-channel approach is intended to integrate Poolside’s tools into existing AI and developer workflows, broadening its reach across the ecosystem.
In addition, Poolside highlighted a session at an Amazon Web Services Summit on bringing AI-powered software development into secure environments within an organization’s AWS boundary. Company representative Ivo Pinto is expected to discuss deploying frontier AI models and agents closer to development teams, particularly in sensitive or constrained settings.
The AWS-focused messaging stresses the benefits of owning AI models for security, governance, and cost management at scale. This aligns Poolside with enterprise and potentially regulated-industry customers that require robust control over their AI infrastructure and deployment.
Taken together, the week’s developments show Poolside strengthening its technology stack, deepening ecosystem partnerships, and reinforcing its focus on scalable, secure, and open AI infrastructure. These moves could support the company’s long-term prospects as organizations increasingly seek flexible, high-performance AI solutions.