Blog
Sovereign AI, LLM APIs and self-hosting.
Guides on sovereign AI in Australia, self-hosted LLMs and LLM API pricing, from the team building Xinference.
Start here.
The economics of sovereign vs cloud LLM inference: TCO analysis for enterprises.
How to compare the total cost of running models on your own infrastructure with paying for a managed API, including payback periods.
Self-hosted inference platforms compared: buyer’s guide 2026.
Ollama, vLLM, Xinference, and NVIDIA NIM compared across setup effort, multi-tenancy, model breadth, and total cost of ownership.
LLM inference cost: managed API vs self-hosted vs hybrid.
Which option costs least depends on traffic shape, model quality, utilization and operating cost. How to work it out for your workload.
News
Partnerships, events and company updates.
News
What we saw at The AI Summit Australia 2026
Oct 1, 2026 · 2 min read
Our recap of the first AI Summit Australia in Melbourne, 7 to 9 September 2026: what the event covered and what we showed on our stand.
News
What we heard at the AI & Data Governance Summit 2026
Oct 1, 2026 · 3 min read
Notes from ADG Summit 2026 in Melbourne, where Paul Towers moderated a panel on what AI governance looks like day to day.
News
Recap: Sydney Enterprise AI and Automation Summit 2026
Oct 1, 2026 · 2 min read
We sponsored the coffee cart at Clutch Events' Sydney Enterprise AI and Automation Summit on 12 August. Here is who was in the room.
News
Meet Xinference at Singapore FinTech Festival 2026, booth 6F07-08
Sep 30, 2026 · 1 min read
We will be in the AI & Quantum zone in Hall 6 at Singapore EXPO from 18 to 20 November 2026, booth 6F07-08.
News
Meet Xinference at the AI Infrastructure & Enterprise Systems Summit ANZ, 26 to 27 October 2026
Sep 30, 2026 · 1 min read
Find our stand in the Grand Ballroom at InterContinental Sydney Double Bay on 26 and 27 October 2026, at IQPC's AI Infrastructure & Enterprise Systems Summit ANZ.
News
Meet Xinference at DevOps Summit NSW on 21 October 2026
Sep 30, 2026 · 1 min read
We will be at DevOps Summit NSW at ICC Sydney on Wednesday 21 October 2026, with a stand for the whole day and one of the afternoon roundtables.
News
Xinference partners with BioQuest Advisory to deliver enterprise AI across Asia Pacific
Sep 30, 2026 · 2 min read
Xinference and BioQuest Advisory are now partners across Asia Pacific. Together we’re putting enterprise AI into production, on terms the customer sets.
News
Who controls your enterprise AI? A webinar with BioQuest Advisory on 15 October
Sep 30, 2026 · 2 min read
Join Xinference and BioQuest Advisory on Thursday 15 October for a free one-hour webinar on cost, security and model choice.
Sovereign AI in Australia
What it takes to run AI on infrastructure you control.
Sovereign AI in Australia
Can you use ChatGPT for confidential documents?
Oct 6, 2026 · 6 min read
Only on a business tier and only for approved documents. Business tiers skip training by default, but OpenAI can still store and screen your data.
Sovereign AI in Australia
Data sovereignty in Australia: what it means for AI
Oct 6, 2026 · 7 min read
Storing data in an Australian region is not the same as processing it there. What data sovereignty means for AI, and what to check before you buy.
Sovereign AI in Australia
The economics of sovereign vs cloud LLM inference: TCO analysis for enterprises.
Aug 7, 2026 · 10 min read
How to compare the total cost of running models on your own infrastructure with paying for a managed API, including payback periods.
Sovereign AI in Australia
Control is the moat, not model performance.
Jul 10, 2026 · 7 min read
Four controls an enterprise AI platform needs: versioned model lineage, tiered routing, consistent environments and private deployment.
Sovereign AI in Australia
From model selection to fleet management.
May 12, 2026 · 7 min read
When a company runs many models on separate stacks, AI becomes a fleet management problem. What platform teams do about it.
Sovereign AI in Australia
4 lessons from a 50-model private GPU cluster.
May 28, 2026 · 8 min read
What we learned consolidating 50+ open models onto one private cluster serving 6,000+ users.
Sovereign AI in Australia
Xinference brand refresh and platform update.
Jul 2, 2026 · 5 min read
What changed in the July 2026 Xinference brand refresh and platform update.
Self-hosted LLMs
Running open models on your own servers, and the platforms that do it.
Self-hosted LLMs
What is LLM inference? Engines, providers and GPUs explained
Oct 13, 2026 · 5 min read
LLM inference is running a trained language model on a prompt to produce output without changing its weights, which is what training does.
Self-hosted LLMs
A self-hosted alternative to the OpenAI API: switch without rewriting your code
Oct 1, 2026 · 2 min read
Keep your code and your OpenAI SDK, change the base URL, and run open models on your own infrastructure.
Self-hosted LLMs
Self-hosted inference platforms compared: buyer’s guide 2026.
Aug 4, 2026 · 10 min read
Compare Ollama, vLLM, SGLang, NVIDIA NIM and Xinference for running a self-hosted LLM: hardware, model support, scale and the work to run each.
Self-hosted LLMs
Xinference vs Ollama for enterprise: scaling local models.
Aug 7, 2026 · 12 min read
Ollama is built for one developer on one machine. Xinference adds the multi-user isolation, quotas and audit logs a team needs.
Self-hosted LLMs
Xinference vs vLLM: complete platform or raw engine?
Aug 7, 2026 · 11 min read
vLLM is an inference engine built for throughput. Xinference is a platform that adds multi-user access, quotas and model management for teams.
Self-hosted LLMs
Xinference vs NVIDIA NIM: platform flexibility or NVIDIA optimization?
Aug 19, 2026 · 4 min read
NIM packages NVIDIA-tested serving software for supported models. Xinference is a model-serving platform with several engine backends. How they compare on hardware, orchestration and cost.
Self-hosted LLMs
Deploy your first model on Xinference in 10 minutes.
Jun 10, 2026 · 6 min read
Install Xinference, launch a model and call it through the OpenAI-compatible API on your own hardware.
Self-hosted LLMs
How to choose an LLM inference stack.
Jul 22, 2026 · 9 min read
Six questions that narrow the choice between a managed API, a single-model server and a unified inference platform.
Self-hosted LLMs
Production LLM inference checklist: 8 tests before launch.
Jul 22, 2026 · 9 min read
Eight tests to run on a model endpoint before it takes production traffic.
LLM API pricing
Comparing the cost of LLM APIs and self-hosting.
LLM API pricing
Is self-hosting an LLM worth it? Self-hosted vs API cost
Jul 22, 2026 · 10 min read
Which option costs least depends on traffic shape, model quality, utilization and operating cost. How to work it out for your workload.
LLM API pricing
LLM model routing: how xrouter-llm chooses the cheapest capable model.
Jul 22, 2026 · 8 min read
How to send each prompt to the cheapest model that is likely to meet your quality bar.
Model APIs
Choosing and running open-weight models.
Model APIs
What is an OpenAI-compatible API? What it covers and what to test
Oct 15, 2026 · 5 min read
An OpenAI-compatible API accepts OpenAI-format requests for the endpoints it supports, so the OpenAI SDK can call it with a new base URL and key.
Model APIs
Open-weight AI models in 2026: a guide by modality.
Jul 22, 2026 · 12 min read
A shortlist of open-weight models by modality, and how to evaluate them before replacing a managed API.
Model APIs
Choosing an enterprise model: a workload-first decision framework.
Aug 21, 2026 · 4 min read
How to pick the smallest model that meets your quality, latency, governance and reliability thresholds.
Model APIs
When RAG isn’t enough: a field guide to fine-tuning.
Jun 19, 2026 · 8 min read
Four situations where fine-tuning an open model works better than retrieval alone, and a framework for deciding.
API alternatives
How Xinference compares with managed inference APIs.
Agents
Notes on building and governing AI agents.
