About Terafab.com.au
This site offers AI infrastructure tools and calculators to help with planning, budgeting, and understanding AI compute. From training cost estimates to datacentre power budgets, it uses industry benchmarks and first-principles models for reference. Results are estimates only; verify important figures yourself.
AI infrastructure spend is projected to exceed $1T by 2030, yet planning tools remain fragmented. This personal project provides free calculators - training cost, GPU clusters, semiconductor fab, datacentre power, RAG vs fine-tune vs train - plus a token counter and AI glossary. No sign-up required.
AI infrastructure spending is projected to exceed $1T by 2030. Frontier models train on trillions of tokens with hundreds of billions of parameters. Power is increasingly the binding constraint-not chips.
Training costs scale with parameters × tokens × GPU hours. Beyond hardware: cooling, networking, storage, and facility overhead typically add 10–30%. Low utilisation multiplies effective cost per hour.
Contact Us
Questions or feedback? Get in touch via our contact form.
FAQ
Visit our FAQ Center for more information and commonly asked questions.
