OpenAI Signs Malaysian Data Center Capacity Deal
OpenAI Signs Malaysian Data Center Capacity Deal
Anonymous
Trending on XAI data centers electricity and water demands debate
Key Points
1
OpenAI Vice President of Compute Strategy Sachin Katti said the Malaysian facilities will help the company “serve growing demand for OpenAI’s products across the region and around the world.”
2
Firmus is also developing a 360-megawatt campus in Batam, Indonesia, designed to deploy about 170,000 Nvidia AI accelerators at the site.
3
Firmus’ HyperCube units combine liquid cooling, mechanical systems and electrification in prefabricated modules manufactured in regional New South Wales.
4
Firmus’ investors include Nvidia, Jane Street, Blackstone funds and Coatue Management; its most recent fundraising valued the company at more than $10.5 billion.
5
The OpenAI contract covers Firmus’ two Malaysian AI factories but does not include commitments for the company’s planned Australian facilities.
OpenAI has signed a multiyear deal to secure computing power from two data centers in Malaysia operated by Firmus Technologies, an Australian AI infrastructure company. E-Week reports the agreement makes OpenAI an anchor customer and pushes Firmus's total contracted capacity above 900 megawatts, expanding operations to a fourth country across Asia-Pacific.
The Malaysian facilities will use Nvidia's next-generation Vera Rubin systems and Firmus's liquid-cooled HyperCube infrastructure to support demand for OpenAI's products across the region and globally. Quartz notes the partnership strengthens Firmus ahead of a potential Australian stock listing valued at up to A$15.5 billion.
OpenAI Secures Regional Compute Capacity
Pulse2 reports Firmus has entered a strategic partnership with OpenAI to provide dedicated AI compute capacity from two Malaysian AI Factory sites. The contract covers Firmus's Malaysian operations but excludes planned Australian facilities. Financial terms were not disclosed.
OpenAI Vice President of Compute Strategy Sachin Katti said the Malaysian facilities will help the company serve growing demand across the region and worldwide. Malaysia offers affordable electricity, political stability, and proximity to major undersea cable routes that reduce latency for Asian users.
Firmus Expands Across Asia-Pacific With Nvidia Tech
The deal pushes Firmus's committed customer capacity beyond 900 megawatts and marks entry into a fourth country. Yahoo Finance reports Firmus operates two AI factories currently and is developing five more across the Asia-Pacific region, including a 360-megawatt campus in Batam, Indonesia designed to deploy about 170,000 Nvidia AI accelerators.
Firmus's HyperCube units combine liquid cooling, mechanical systems and electrification in prefabricated modules manufactured in regional New South Wales. The facilities will deploy Nvidia's Vera Rubin systems alongside the HyperCube infrastructure to maximize cooling efficiency and computing density.
Strategic Anchor Bolsters Firmus's Path to IPO
OpenAI's role as an anchor customer provides Firmus with long-term revenue visibility ahead of its potential Australian stock listing. Firmus's investors include Nvidia, Jane Street, Blackstone funds and Coatue Management. The company was valued above $10.5 billion in its most recent fundraising round.
The undisclosed contract value leaves investors uncertain about pricing, profitability and customer concentration risks. Reports have suggested a prospective Australian listing valued around A$15.5 billion, though timing remains uncertain. The OpenAI deal demonstrates strong regional demand for premium AI compute capacity.
Malaysia's Data-Center Boom Draws Environmental Scrutiny
Malaysia's rapid expansion as an AI infrastructure hub comes with growing concerns about electricity and water consumption. Large-scale data centers require continuous power and cooling resources that strain regional resources. The OpenAI-Firmus partnership exemplifies how AI companies are locking in capacity years ahead of demand to secure competitive advantage.
Access to power, specialized chips, cooling and connectivity is becoming a strategic constraint alongside software capability. This infrastructure race can accelerate model development and service delivery, but it also raises execution, energy, supply-chain and asset-utilization risks if demand or chip economics shift unexpectedly.
Sentiment
Mostly PositiveNEGATIVENEUTRALPOSITIVE
Publishers
16
Articles
90
Reach
106
Topics
A
AnonymousAuthor
