Meet Hyperstack at RAISE 2026, 8th-9th July · Booth #14A · Scale your AI infrastructure with us.

Let’s Talk

Catch Hyperstack at ISC 2026, 22nd-26th June · Booth #A39 · Let's talk GPU-accelerated workloads

Book a Meeting

Reserve early access to NVIDIA B300s — arriving Q3/Q4

Register Your Interest

alert

We’ve been made aware of a fraudulent website impersonating Hyperstack at hyperstack.my.
This domain is not affiliated with Hyperstack or NexGen Cloud.

If you’ve been approached or interacted with this site, please contact our team immediately at support@hyperstack.cloud.

publish-date November 24, 2025

5 min read

Updated-date Updated on 18 Feb 2026

Hyperstack Weekly Rundown 47

Written by

Damanpreet Kaur Vohra

Damanpreet Kaur Vohra

Technical Copywriter, NexGen cloud

Share this post

Table of contents

summary

NVIDIA H100 SXM On-Demand

Sign up/Login

summary

We’re back with your weekly dose of Hyperstack updates!

This week comes with some seriously useful upgrades you’ll want to try. Whether you’re spinning up VMs, fine-tuning models or experimenting in AI Studio, these updates are here to make your workflow smoother and your builds faster.

Take a minute to explore what’s new and see how far you can push your next project on Hyperstack.

New on AI Studio

Here’s what’s new on our full-stack Gen AI platform, AI Studio this week:

Import LoRA Adapters Directly From Hugging Face

No more manual downloads or messy workflows. You can now import external LoRA adapters from Hugging Face straight into AI Studio and plug them into supported base models. Use them instantly in the AI Studio Playground or deploy them via API for inference. Learn more about importing LoRA adapters here.

Sample Datasets Now in the UI

No more hunting for starter data. You’ll now find a curated sample dataset directly inside the AI Studio interface, perfect for quick experiments, testing or getting hands-on without setup friction.

Export Your Fine-Tuned Models

You can now export any fine-tuned model you create in AI Studio and use it for external use, giving you more control and flexibility in your ML workflows.

Haven't tried AI Studio yet? Give Hyperstack AI Studio a spin and see how simple and faster AI building can be.

Try AI Studio →

New on Hyperstack

Here’s what’s new on Hyperstack this week:

Public IP Behaviour Change During Hibernation

You now have the option to retain your VM’s public IP address during hibernation. By default, the public IP is now automatically released to help reduce idle resource costs. Learn how to hibernate a Virtual Machine using the UI.

Latest Fixes and Improvements

A new retain_ip parameter has been added to the Hibernate VM API, so you can programmatically decide whether your VM's public IP stays attached during hibernation.

New on our Blog

Check out the latest tutorials on Hyperstack:

Integrate Hyperstack AI Studio with RooCode for Next-Gen Coding Support:

A Step-by-Step Guide

Modern developers are increasingly turning to AI-powered coding environments to accelerate development and improve code quality. One of the most exciting entrants in this space is Roo Code, a powerful AI-driven coding assistant designed to work directly inside Visual Studio Code (VS Code). In this guide, we’ll walk through how to integrate Hyperstack AI Studio with RooCODE to supercharge your development workflow.

Check out the full tutorial below!

Integrate Hyperstack AI Studio with Zed Code Editor for Powerful Coding Agents:

A Step-by-Step Guide

AI coding tools have evolved into intelligent environments that support code understanding, refactoring, and reasoning about complex systems. For developers who value performance and advanced AI-driven workflows, pairing Zed Editor with Hyperstack AI Studio delivers a powerful solution. This guide covers what sets Zed Editor apart, how Hyperstack AI Studio elevates its AI integration, and provides a step-by-step walkthrough for seamless setup.

Check out the full tutorial below!

Integrate Hyperstack AI Studio as a Provider in LiteLLM:

A Step-by-Step Guide

With the rapid evolution of AI-driven development tools, integrating large language models (LLMs) into software systems has become increasingly accessible and modular. Developers are no longer restricted to a single provider now, they can build hybrid AI systems by combining inference backends, model management tools, and application layer SDKs. Two such powerful tools that make this process seamless are Hyperstack AI Studio and LiteLLM. In this guide, we provide detailed steps to integrate Hyperstack AI Studio as a Provider in LiteLLM.

Check out the full tutorial below!

We’ve got more tutorials coming next week, so stay tuned.

Have an idea you'd like to see in Hyperstack? Let’s bring it to life.

At Hyperstack, we’re committed to continuous improvement and your ideas are a key driver of our innovation.

→ Is there a feature you’ve been waiting for?
→ Something that could speed up your workday?
→ Or a tweak that would make things feel effortless?

Tell us what would make your Hyperstack experience even better. Your feedback sets the direction for what we build next.

Share Feature Request →

That's it for this week's Hyperstack Rundown! Stay tuned for more updates next week and subscribe to our newsletter below for exclusive AI and GPU insights delivered to your inbox!

Missed the Previous Editions?

Catch up on everything you need to know from Hyperstack Weekly below:

👉 Hyperstack Weekly Rundown #45

👉 Hyperstack Weekly Rundown #46

Subscribe to Hyperstack!

Enter your email to get updates to your inbox every week

Hyperstack Weekly Rundown 47: Latest Edition

2:31

Related content

Stay updated with our latest articles.

Product Updates link

Hyperstack Weekly Rundown 55

Welcome to Hyperstack Weekly Rundown Your weekly digest of ...

Welcome to Hyperstack Weekly Rundown

Your weekly digest of the latest updates, tutorials and improvements from Hyperstack. This week features a Kubernetes version update, deployment guides for DeepSeek V4 and Kimi K2.6, and a chance to shape what we build next.

Take a few minutes to catch up on everything.

New on Hyperstack

Check out what's new on Hyperstack this week:

Supported Kubernetes Versions Updated

Kubernetes 1.35.1 is now live and set as the new default for all cluster deployments. Here's what you need to know:

Kubernetes 1.35.1 is the new default; all new clusters will be created on this version automatically.
Kubernetes 1.27.8 has reached end-of-life and is no longer supported for new cluster deployments.

Already running 1.27.8? No action needed. Your existing clusters are unaffected and continue to support adding and removing nodes and node groups. Learn how to upgrade your cluster.

New on our Blog

Check out the latest tutorials on new AI models on Hyperstack:

Deploy DeepSeek-V4 on Hyperstack:

A Step-by-Step Guide

DeepSeek-V4 is DeepSeek AI’s latest open-weight LLM family, optimised for efficiency and long-context reasoning. The lineup spans 284B–1.6T parameters with only 13B active per step, a native 1M-token context window, and FP4+FP8 mixed precision for efficient single-node deployment. This tutorial shows how to deploy DeepSeek-V4-Flash on Hyperstack using vLLM and Docker with an OpenAI-compatible API.

Read the full tutorial →

Deploy DeepSeek-V4 Pro on Hyperstack:

A Step-by-Step Guide

DeepSeek-V4 Pro is a 1.6 trillion parameter sparse MoE with 49 billion active per token, a 1M token context window, and an efficient hybrid attention stack. It delivers strong coding benchmark performance but exceeds single-node limits, making it ideal for multi-node deployment on 8x H100. This tutorial covers running it on Hyperstack using vLLM and Docker.

Read the full tutorial→

Step-by-Step Guide to Deploying DeepSeek - Blog post - 1000x600-1

Deploy Kimi K2.6 on Hyperstack:

A Step-by-Step Guide

Kimi K2.6 is an open-weight, native multimodal agentic model from Moonshot AI, built for advanced coding and autonomous agent workflows. Using a sparse Mixture-of-Experts architecture with 1T parameters and only 32B active per step, it matches or exceeds leading closed models on coding and agentic benchmarks while remaining practical to self-host. This tutorial covers deploying Kimi K2.6 on Hyperstack using a GPU VM, vLLM and an OpenAI-compatible API.

Read the full tutorial→

Help Shape the Future of Hyperstack

Great products are built with the people who use them. If there’s something you would like to see on Hyperstack whether it is a new feature, workflow improvement or integration that would make your work easier, we would love to hear about it.

Your feedback helps us prioritise what matters most and build a platform that works better for the community.

Share Feature Request →

That's it for this week's Hyperstack Rundown! Stay tuned for more updates next week and subscribe to our newsletter below for exclusive AI and GPU insights delivered to your inbox!

5 May 2026

•

Read More

Product Updates link

Hyperstack Weekly Rundown 54

Welcome to Hyperstack Weekly Rundown The GPU race doesn’t ...

Welcome to Hyperstack Weekly Rundown

The GPU race doesn’t slow down and neither do we. A new region, smarter inference, AI Image Playground and Auto Top-Up are all part of what’s coming up on Hyperstack.

Take a few minutes to catch up on everything.

New Hyperstack Region: EU-1 is Now Live

Built from the ground up for next-generation AI workloads, the EU-1 region delivers the power density, advanced cooling and physical capacity required to run the latest NVIDIA platforms at scale. With strong early demand, all Phase 1 capacity has already been fully reserved.

Expansion is already underway. In partnership with Glesys, we’re bringing additional NVIDIA Blackwell and Blackwell Ultra capacity online later this year to support growing customer needs.

If you're planning upcoming deployments, now is the time to get ahead of demand. We’re currently taking expressions of interest for future capacity. Reach out to discuss your requirements at sales@hyperstack.cloud.

Coming Soon

AI Studio Image Playground

We’re bringing image generation and editing to AI Studio. Create images from text, transform existing visuals and switch seamlessly between text-to-image and image-to-image workflows. All powered by state-of-the-art vision models.

Stay tuned for what’s coming next.

Auto Top-Up

Never run out of credits mid-workflow again. Auto Top-Up will keep your balance topped up automatically the moment it hits your set threshold. Set your limit, choose your top-up amount and keep building as your compute stays ready.

Launching beta customers next week, with GA to follow.

New on our Blog

Check out the latest blogs on Hyperstack:

Managed Kubernetes vs Managed SLURM:

Which Orchestrator Fits Your AI Workloads

Orchestration is not a deployment detail; it is the layer that determines whether your training runs hit theoretical throughput or loses half of it to scheduling contention, broken GPU allocation, and inter-node coordination failures. The Kubernetes vs SLURM decision compounds across every run you ship.

Read the full blog→

5 Inference Optimisation Techniques:

To Improve Performance

The model passes evaluation but latency is 3× too slow for production. This is where real inference work starts. Making a model accurate is a research problem; making it fast and efficient is an engineering problem with well-understood solutions.

Read the full blog→

How to Run Cost-Efficient Inference Workloads:

A Step-by-Step Guide

Cost-efficient LLM inference on Hyperstack or any GPU cloud means maximising tokens per dollar by optimising every layer of the stack. In practice, that includes right-sizing GPUs, applying quantisation, optimising KV cache, using modern runtimes such as TensorRT-LLM, vLLM and TGI, batching well, autoscaling dynamically and tracking the right metrics, often delivering 2×–10× cost savings in production.

Read the full tutorial→

When to Choose Dedicated Private Cloud:

A Decision Framework

Your model is ready and the team has delivered. Then InfoSec asks where the data lives and deployment stalls. Not because the infrastructure failed, but because you cannot prove it didn’t. This guide explains where these gaps come from, what compliance teams actually need to see and how to design infrastructure that holds up under scrutiny.

Read the full blog→

Distributed Inference Explained:

When Enterprise Teams Need It

Most teams only think about distributed inference when something breaks: latency spikes, GPUs max out or queues back up. By then, it’s no longer a design decision, it’s an incident. This guide explains when distributed inference actually becomes necessary, what signals to watch for and how to approach it as an architectural choice, not a last-minute fix.

Read the full blog→

Help Shape the Future of Hyperstack

Great products are built with the people who use them. If there’s something you would like to see on Hyperstack whether it is a new feature, workflow improvement or integration that would make your work easier, we would love to hear about it.

Your feedback helps us prioritise what matters most and build a platform that works better for the community.

Share Feature Request →

That's it for this week's Hyperstack Rundown! Stay tuned for more updates next week and subscribe to our newsletter below for exclusive AI and GPU insights delivered to your inbox!

Missed the Previous Editions?

Catch up on everything you need to know from Hyperstack updates below:

👉 Hyperstack March Update

27 Apr 2026

•

Read More

Product Updates link

Hyperstack March Update: New Features, Improvements and Tutorials

Hyperstack Monthly Update March brought a range of updates ...

Hyperstack Monthly Update

March brought a range of updates to the Hyperstack platform, with a focus on improving how you interact with your infrastructure. We also released exciting tutorials on Hyperstack MCP, OpenClaw and NemoClaw following its introduction at the NVIDIA GTC 2026.

Scroll down to explore the highlights.

New on Hyperstack

Here's what we released on Hyperstack in March:

Hyperstack MCP Server

We introduced the Hyperstack MCP (Model Context Protocol) Server, a new interface layer that allows you to manage infrastructure using natural language.

Instead of relying on manual API calls or navigating multiple dashboards, the MCP Server translates plain English instructions into secure and structured API operations. This makes common tasks like provisioning VMs, managing resources or updating configurations faster and more intuitive.

The MCP Server is compatible with MCP-enabled clients such as Claude Desktop and Open WebUI. This lets you integrate infrastructure control directly into the tools you already use.

No need to write or manage raw API calls
Faster execution of routine infrastructure tasks
Reduced friction in developer workflows

Node Group Firewall Management for Clusters

Firewall rules can now be defined and managed at the node group level within Kubernetes clusters. These rules can be configured directly from the console after cluster deployment and are automatically applied across all worker nodes in the selected node group.

Node Group Firewall API Support

Firewall management for node groups is now fully supported via API using the firewall_ids field. This capability is available across the following endpoints:

Create Cluster: Apply firewall rules during cluster creation.
Create Node Group: Apply firewall rules when adding node groups.
Update Node Group: Update firewall rules for existing node groups.

Across all cluster-related API responses that include the node_groups object, it will now return associated firewall details for each node group.

Latest Fixes and Improvements

We also made some improvements on Hyperstack to make your experience even better:

Kubernetes Cluster Deployment Defaults

The cluster deployment experience has been improved by automatically selecting the Kubernetes version and setting Full Deployment Mode as the default option.

User Role Management Improvements

The Create and Edit User Role experience has been enhanced with:

A new resource-based permission filter
Clear visibility of permissions by resource group with selection counts
Reduced horizontal scrolling and improved layout for better usability

Improved Resource Availability Error Messaging

VM and Kubernetes cluster creation failures due to temporary capacity shortages now display a clearer message, indicating resource unavailability and suggesting a retry.

VM Hibernation Request Handling Improvements

Improved lifecycle handling ensures instance states remain consistent during operations like hibernation, preventing duplicate requests while an operation is already in progress.

Firewall Attachment Improvements

Enhanced reliability and usability when attaching virtual machines to firewalls, with more consistent handling of VM states and firewall updates.

New on the Blog

Check out exciting blogs and tutorials on Hyperstack this month:

Deploying NVIDIA's NemoClaw on Hyperstack:

Step-by-Step Guide

In our latest tutorial, we cover what NemoClaw is, how it works as NVIDIA's open-source security stack for OpenClaw, and walk you through the full deployment process on Hyperstack GPUs including setup, configuration and verification steps.

Securing OpenClaw on Hyperstack: Safe AI Agent Deployment:

A Comprehensive Guide

In our latest tutorial, we cover how to securely deploy OpenClaw AI agents on Hyperstack, including network isolation, access controls, environment hardening and safety configurations to ensure your agentic workloads run reliably and safely on cloud GPUs.

Why Move Your AI Workloads from Public Cloud to Private Cloud:

5 Important Signs

In our latest blog, we walk through five concrete signs that public cloud is holding your AI operations back, including unpredictable billing, data sovereignty risks, limited GPU availability, compliance challenges and performance ceilings that private cloud can help you overcome.

Manage Cloud Infrastructure with Open WebUI:

Using the Hyperstack MCP Server

In our latest tutorial, we show you how to connect the Hyperstack MCP Server to Open WebUI, allowing you to provision VMs, manage GPU resources and control your cloud infrastructure through natural language conversations with AI assistants like Claude Desktop.

NVIDIA GTC 2026: Conversations on Secure AI Infrastructure

Collage Image ( abd ) - Blog

Our team attended the NVIDIA GTC 2026 in March and the conversations we had there reflected something a lot of teams are thinking about right now: how to build infrastructure for production AI the right way.

We had discussions around our Secure Private Cloud and why isolation, compliance and control need to be built in from day one rather than added on later. If you're deploying AI that needs dedicated infrastructure with full visibility and governance, book a meeting with the Hyperstack team:

REQUEST A CONSULTATION

Your Ideas Power Hyperstack

You know your workflow better than anyone. If there’s anything you wish Hyperstack did differently or better, now’s your chance to tell us.

Maybe it’s a feature you’ve been thinking about, a tool that could speed up your workflow, or a simple improvement that would make your project easier. Whatever it is, we’re listening.

Share Feature Request →

For any questions or suggestions, feel free to reach out at support@hyperstack.cloud. Stay tuned for even more updates and exciting tools next month.

2 Apr 2026

•

Read More

POWERED BY:

United Kingdom (Head office)

TechSpace.

9-13 St Andrew St, 3rd Floor,

London EC4A 3AF,

United Kingdom

Registered Office

6th Floor, 99 Gresham Street,

London, EC2V 7NG

United Kingdom

Spain

Ctra NACIONAL 340,

KM 176

Local C-12,

Marbella 29600

Malaga

Solutions

Resources

Site map

Products

Legal

® 2026 Hyperstack. All rights reserved. The Hyperstack logo is a registered trademark of NexGen Cloud Ltd. in the UK. Other company and product names be trademarks of the respective companies with which they are associated.