Datasection Builds Big AI Computer in Sydney With 10,000 Nvidia GPUs

Datasection is installing 10,000 Nvidia B300 GPUs in Sydney. This is a huge number of powerful chips for AI.
By Newsroom | Technology, Australia |

Datasection, a company positioning itself as a "neocloud operator," is pushing forward with significant plans to build large-scale artificial intelligence clusters, asserting its unique standing within Japan for such endeavors. The firm claims to be the only Japanese entity concurrently developing massive AI clusters while securing essential GPU capacity, power, and customers on a worldwide scale.

The company's financial outlook for the fiscal year ending March 2027 is explicitly tied to the timely operational launch of facilities in Japan, Australia, and Thailand, alongside the continued expansion of its primary client. Projects designated 'A' and 'B', located in Japan and Australia respectively, are already under contract.

Inventec Deal Fuels Australian Supercluster

In a move underscoring these ambitions, Datasection has contracted with Inventec for 1,250 servers destined to house 10,000 Nvidia B300 GPUs. This hardware will form the backbone of what the company describes as a "world-class, first-of-its-kind large-scale AI cluster" situated in Sydney, Australia. This deployment aims to deliver enterprise-grade AI computing services across Japan and the broader Asia-Pacific region.

Read More: Elderly Streamer 'GrammaCrackers' Swatted During Grandson's Fundraiser

The cluster, powered by Nvidia's latest Blackwell B300 architecture, is projected to exceed 200 ExaFLOPS in compute performance. A key element highlighted by Datasection is its proprietary TAIZA algorithm, which is said to optimize the performance of these extensive GPU setups under demanding AI workloads.

Global Aspirations and Market Positioning

Datasection's push into hyperscale AI infrastructure, particularly with this significant Nvidia B300 implementation in Sydney, positions it as a notable player in the rapidly expanding Asia-Pacific AI market. The company states that only a few Japanese firms possess the capability to construct such large-scale AI clusters. The success of Datasection's strategy hinges on the precise commissioning of its facilities, the continued growth of its anchor customer, and the conversion of its prospective client pipeline into concrete revenue.

Read More: Moreton Bay gets $1 Billion Quantum Computer Project

Frequently Asked Questions

Q: What is Datasection building in Sydney, Australia?
Datasection is building a large AI computer cluster in Sydney. It will use 1,250 servers and 10,000 Nvidia B300 GPUs.
Q: How powerful will the Sydney AI cluster be?
The AI cluster in Sydney is expected to have a compute performance of over 200 ExaFLOPS. Datasection will use its TAIZA algorithm to make it work better.
Q: When will Datasection's new AI facilities be ready?
Datasection plans to launch its new AI facilities in Japan, Australia, and Thailand by March 2027. The Sydney cluster is part of this plan.
Q: Who is supplying the servers for the Sydney AI cluster?
Inventec is supplying the 1,250 servers for Datasection's AI cluster in Sydney. These servers will hold the 10,000 Nvidia B300 GPUs.