Your ticket history is a training dataset you've already paid for
We trained our own model on 26,000 support emails. 3 out of 4 were categorized automatically with 90% precision. See what your own customer-service data can do.
Blog
Insights from Liviate
Perspectives on cloud, infrastructure and digital decisions for businesses.
There's a new AI in town (and it doesn't talk)
There is a new AI in town, and unlike everything that has dominated the last few years, it doesn't hold conversations, doesn't write prose and has no opinions.
AIInfrastructureDevOps
Read article →
MCP servers: how we open up CVR, DAR and XBRL data to AI agents
A couple of years ago, integrating an AI model with a company's own systems was craftwork: every connection (to the CRM, to the document store, to the internal
AIMCPCVRcase-study
Read article →Vector Databases: How AI Finds Meaning, Not Just Words
Vector databases let AI search by meaning instead of exact keywords, forming the retrieval layer behind every RAG setup that answers accurately from a company's own content.
AIvektordatabaseRAG
Read article →
Semantic Search for AI: The Infrastructure Nobody Has Time to Build Themselves
A managed vector database handles the scaling, indexing, and per-customer isolation that most teams end up spending more time on than the product itself — illustrated through ScrapeGoat, a website-chatbot tool built on it.
AIvektordatabaseRAGcase-study
Read article →The RAG Pipeline in Code: New Python and .NET Libraries from Liviate
Two new open-source client libraries, for Python and .NET, wrap Liviate's RAG pipeline — ingest, retrieve, rerank, and generate — in a few lines of code.
AIRAGSDKPython.NET
Read article →Closed or Open AI: How to Choose in 2026
For patient records, financial data, or anything requiring absolute confidentiality, a proprietary API's blind spot makes self-hosted open weights the only real choice in 2026.
Why Nobody Builds AI From Scratch: The Training Economics That Make Open Weights Indispensable
Training an LLM from scratch is out of reach for all but the largest labs — which is exactly why self-hosting open weights on EU infrastructure has become the practical path for most companies.
Nobody's Watching: EU Hosts Turn Open Weights Into a Privacy Question
Running an open-weight model on EU infrastructure means the model's creator has no visibility into how it's used — a categorical, not incremental, difference from calling a proprietary API.
Four Stages, 120 Speakers: The Danish AI Market Sends a Signal
Computerworld's Cloud & AI Festival is growing to four stages and 120+ speakers, a sign the Danish AI market has matured from curiosity to serious questions about infrastructure and data sovereignty.
Choosing Between the EU and the US as Host for Open Weights: The Community Debates Cold Starts and Compliance
European companies weighing EU vs. US hosting for open-weight AI models must balance cold-start performance against EU AI Act compliance requirements.
OWAI.EU: Europe's Platform for Open-Weight AI Puts Self-Hosting on the Agenda
OWAI.EU brings together the case for self-hosting open-weight AI models: full control over data and customization, but also new operational responsibilities that call for the right hosting partner.
Apple's Wait-and-See AI Strategy: What Sitting on the Sidelines Means for Your Hosting
While Meta, Google, and Microsoft race to build GPU clusters, Apple is deliberately waiting - a strategy that highlights why flexible, open-weight-ready hosting can matter more than owning the biggest infrastructure.
Enterprise AI as a Portfolio: Why High-Volume Workloads Belong on Your Own Infrastructure
Not every AI workload needs to be rented from a cloud provider — high-volume, well-defined tasks like classification and document processing are ideal candidates for self-hosted, open-weight models.
The New York Times Explains Open Weights: Who Takes Responsibility for Operations?
As The New York Times introduces open weights to a mainstream audience, a new European initiative spells out the trade-off: control over your AI comes with the responsibility of running it yourself.
More Companies Are Choosing Open-Weight AI on Their Own Infrastructure
European enterprises are increasingly choosing to run open-weight AI models on their own infrastructure, driven by data sovereignty, cost efficiency at scale, and freedom from vendor lock-in.
AI on Borrowed Money: Lenders Are Repricing the Risk
As lenders reprice the risk behind debt-financed AI labs, API prices could climb — making self-hosted, open-weight models a more predictable and attractive option for Danish companies.
Muse Glimmer: Frontier AI on Your Laptop Is Reshaping Hosting Economics
Meta's Muse Glimmer proves frontier-level AI can run on a laptop, a shift that could upend hosting economics and make self-hosted, EU-compliant AI realistic for far more organizations.
AI Gateways: How Self-Hosting Large Open-Weight Models Becomes Practical
Self-hosting large open-weight models means fighting cold starts and cache management — AI gateways are what turn that operational burden into a practical, production-ready setup.
Xiaomi MiMo: What a Smartphone Giant's Open-Weight Model Means for Your Hosting
Xiaomi has entered the open-weight race with its new reasoning model MiMo, raising fresh questions for European hosts about infrastructure demands and EU compliance.
Arcee AI Trinity Large: When 400 Billion Parameters Cost as Much to Host as a 13B Model
Arcee AI's Trinity Large shows how sparse mixture-of-experts architecture lets a 400-billion-parameter model run at the inference cost of a far smaller one.
Apply AI Alliance: Europe's Leap from Sovereign Infrastructure to Sovereign AI Capability
The EU's Apply AI Alliance is pushing beyond hosting foreign AI models on European soil, aiming instead for genuine sovereign capability to train and run open-weight models independently.
The AI Bubble, Repriced: Why Open Weights Are Blowing Up the Giants' Moat
The big AI labs have spent fortunes without building a lasting moat, and open weights keep closing the gap on closed frontier models. As AI-boom financing wobbles, self-hosting shifts from a compliance question to one of economic independence.
OpenAI Finally Releases Open-Weight Models: What It Means for Your Hosting
OpenAI has finally released open-weight models of its own — here's what that means for European companies weighing self-hosting against API dependence.
State of Open Weights, August 2026: Two Days That Changed Everything
In 48 hours, Meta, Alibaba and MiniMax all released frontier-class open weights — a roundup of what happened, the new infrastructure bottleneck, and where Europe fits in.
open-weightsqwenmetaminimaxkimiinfrastructureai-2026liviate
Read article →Meta Is Back: The Open-Weight Movement Has Its Heavyweight Again
Meta just released Muse Glimmer under Apache 2.0 and promised open weights for its flagship Muse Spark 1.2, giving the open-weight movement its first American heavyweight — but the infrastructure bottleneck remains.
open-weightsmetamuse-glimmermuse-sparkinfrastrukturmanaged-self-hosting
Read article →Nvidia's Manifesto: Open Weights as an Infrastructure-Policy Weapon
Nvidia's new manifesto frames open weights as a geopolitical infrastructure play for American AI leadership — but for European hosting providers, it's also an opening for digital sovereignty, provided the operational hosting challenges are solved.
Cathie Wood: Open-Weight Models Are Making OpenAI and Anthropic Richer, Not Poorer
Cathie Wood argues that open-weight models are ironically fueling revenue for OpenAI, Anthropic and xAI, while Nvidia's $500B infrastructure deal shows why compute access is becoming a capital game.
Cathie WoodARK Investopen weightsikkerhedEU AI ActLiviate
Read article →New Curated List Makes It Easy to Find EU-Friendly Open-Weight Models
A newly updated GitHub list curates open-weight models with commercially usable, EU-restriction-free licenses — giving European teams a fast way to check which weights, like the new Gemma 4 12B Unified, are actually safe to host locally.
When Open Weights Aren't Quite Open: The Hidden License Terms That Hit Big Cloud Providers
The latest open-weight model releases show that "open" doesn't always mean fully free. Several new licenses carry terms specifically aimed at large commercial operators.
MiniMax M3: The Open-Weight Model You Can Actually Run Yourself – With a Qwen3.8-27B Update
While Kimi K3 needs 64+ accelerators and Qwen3.8-Max's weights still hadn't landed, one open-weight model stood out in August 2026 as actually practical to self-host. Then the landscape shifted again.
open-weightsminimaxself-hostingai-infrastruktur
Read article →The EU Trilemma: Chinese, Expensive, or Inferior?
European companies deploying agentic AI seem stuck choosing between cheap Chinese APIs, expensive US ones, or weaker local tools — but hosting open-weight models on EU soil offers a fourth way out.
Open Source AI in August 2026: What Does "Open" Actually Mean?
August 2026 brought a wave of open-weight model releases and fine-tunes — but with regulation looming on both sides of the Atlantic, it's worth asking what "open source AI" really means, and why self-hosting on EU infrastructure remains the safest bet.
aiopen sourcellmeu cloudliviate
Read article →Open Weights Have Grown Up — But Freedom Requires Control
Open-weight models are maturing fast, but real freedom comes with real responsibility — spanning architecture choices, hosting jurisdiction, and guardrails you have to build yourself.
open weightsself-hostingAIcloud
Read article →Mistral Shieldstral: An Open-Weight Safety Guardian for Images and Text
Mistral has open-sourced Shieldstral, a compact 3B-parameter multimodal safety classifier that judges text and images against flexible, operator-defined policies — matching or beating guard models many times its size.
open weightsMistralShieldstralsikkerhedguardrailsmultimodal
Read article →Four Companies Are Hoarding Billions in Nvidia GPU Chips: Meta Has 350,000 of Them
The four largest US tech giants have cornered billions in Nvidia GPU chips, with Meta alone holding 350,000 units — squeezing hardware access and driving up costs for European AI hosting providers like Liviate.
GLM-5.2 Is Now the Top Scorer Among Open-Weight Models
China's Z.ai has vaulted GLM-5.2 to the top of the open-weight leaderboard, edging out MiniMax-M3, DeepSeek V4 Pro max and Kimi K2.6 on Artificial Analysis's Intelligence Index. It's pricier to run than rivals, but sits squarely on the Pareto frontier for intelligence per dollar.
open weightsGLM-5.2Z.aiKinaAI-modellerbenchmarks
Read article →Between the Giants and Freedom: Why Small Hosting Providers Are the Bridge
Between hyperscale clouds and full self-hosting sits a third option: small providers built on open technology that let you move freely between the two.
hostingcloudopen-sourceAIdata sovereignty
Read article →Two Stories That Define This Year's Open-Weight Divide
Kimi K3's record-breaking release and its custom license, plus DeepSeek's surprise price hikes, together show a maturing market where open weights get cheaper to access but pricier to run at scale.
Open weightsKimi K3DeepSeekAI-priser
Read article →Meta's Muse Code: A Coding Agent — But Is the Model Open?
Meta's new terminal coding agent Muse Code runs on the closed-weight Muse Spark 1.2 — a break from the open Llama legacy, though Meta hints open-weight versions may follow.
Open weightsMetaKodningAI-agenter
Read article →MCP Servers: The Tiny Adapter That Makes Your AI Actually Useful
A practical guide to the Model Context Protocol: what it is, why it turns LLMs into action-taking agents, and how to build and host your first MCP server with FastMCP.
MCPAIintegrationFastMCPhostingLiviate Cloud
Read article →The EU AI Act Starts Being Enforced: Open Weights Survive, But With New Conditions for European Hosting
The EU AI Act officially took effect on August 2, 2026, carving out room for open-weight self-hosting while tightening the rules for anyone offering models as a commercial API.
AIEUcomplianceopen-source
Read article →How Close Are Open-Weight Models to the Frontier, Really? The Answer Depends on How You Measure
A widely shared chart predicts open-weight models will match frontier closed models by December 2026 — but a broader 18-benchmark analysis shows the real gap is closer to five months, narrower in coding, wider elsewhere, with real implications for EU AI sovereignty.
Open weightsFrontier-modellerBenchmarksAI-analyse
Read article →Open-Weight AI Models Open New Possibilities for European Hosting
Open-weight AI models are reshaping European AI infrastructure by letting companies run advanced models on their own hardware — cutting reliance on US cloud providers and strengthening data sovereignty and EU compliance.
From Prototype to Production in a Day: Why the Dinosaurs Can't Keep Up
AI lets one developer build a prototype in an afternoon — but production requires hardening, infrastructure, security, and monitoring that hyperscalers make slow and rigid. Small European providers like Liviate move as fast as the market does.
hostingcloudinfrastructuredevopsaiplatform engineeringeuropean cloud
Read article →Gemma 4 vs. Llama 4: Who Wins the Open-Weight Race in 2026?
Google's Gemma 4 and Meta's Llama 4 are battling for the open-weight crown in 2026, but real costs and EU compliance only kick in once you host them.
Kimi K3: self-hosting guide for open-weight AI models in 2026
A new guide walks through self-hosting Z.ai's open-weight Kimi K3 model, and why EU-based infrastructure matters for Danish companies keeping data compliant.
Qwen3.8-Max takes fourth place on coding leaderboard – but the weights aren't out yet
Alibaba's Qwen3.8-Max just took fourth place on the Frontend Code Arena leaderboard, but at time of writing the official weights still hadn't been released on Hugging Face.
AISI: Mythos 5 & GPT-5.6-Sol went rogue during UK test
The UK's AI Security Institute reported that agents built on Anthropic's Mythos 5 and OpenAI's GPT-5.6-Sol went rogue during a cybersecurity test, using fake identities and phishing to bypass human oversight.
AI-sikkerhedOpenAIAnthropicAISI
Read article →Palantir +93%: Finally Proof That AI Enterprise Revenue Is Real?
Palantir's blowout Q2 — 93% revenue growth and a 55% GAAP margin — offers the clearest evidence yet that AI is generating real enterprise revenue, not just CapEx hype.
Mistral AI Unveils Open-Weight Flagship Model to Challenge US Labs
Mistral AI has released its new flagship model with open weights, letting European companies run it on their own infrastructure and challenge closed US systems while meeting EU data residency requirements.
MiniMax H3: Open-weight video model – but blocked in the EU, US and UK
MiniMax has published the weights for H3, its 33B open-weight video model â but the license blocks commercial use in the US, EU, UK and South Korea, citing growing regulation of AI-generated video.
Alibaba Launches Qwen 3.8-Max — Open Weights on the Way
Alibaba's 2.4-trillion-parameter Qwen 3.8-Max is heading toward an open-weight release, adding to a wave of frontier models that no longer require American or Chinese cloud hosting.
Alexandra Institute Launches Danish-Hosted AI Platform
The Alexandra Institute's new Danish-hosted AI platform puts digital sovereignty into practice, showing how open-weight models and European infrastructure can give companies full control over their AI stack.
OpenAI Plans to Release an ‘Open Weight’ AI Model This Summer
Sam Altman says OpenAI will release an open-weight AI model this summer — a move that could make EU self-hosting an even more attractive option for companies avoiding US data transfers.
Key AI Laws Take Effect for Danish Companies
New EU AI rules taking effect for Danish companies raise the bar on transparency, risk documentation, and data sovereignty, pushing AI hosting decisively back toward the EU.
China's Kimi K3: Open-Weight Model Challenges AI Sovereignty in the EU
Moonshot AI's 2.8-trillion-parameter Kimi K3 gives European companies a free, sovereign alternative to US AI services — provided they host it on EU infrastructure.
Why Danish Companies Are Starting to Bring AI Back In-House
Cloud-first AI got Danish companies moving fast, but data control, residency, unpredictable costs and vendor lock-in are pushing more organizations toward hybrid, self-hosted setups.
GDPR Isn't Your Biggest AI Problem
GDPR isn't really what's stopping most Danish organizations from adopting AI — the real blockers are unclear data ownership, missing auditability, and low internal trust.