OpenAI Signs Multi-Year Malaysia Compute Deal; Firmus to Deploy Nvidia Vera Rubin at Scale in Asia-Pacific
AI Podcast
On September 8, Sydney time, Australian AI infrastructure company Firmus announced a multi-year partnership with OpenAI, making OpenAI the anchor customer for two new AI factories in Malaysia. Backed by Nvidia, Firmus will deploy Nvidia’s next-generation Vera Rubin NVL72 platform at scale. While contract values and specific procurement scales remain undisclosed, the deal extends OpenAI’s computing procurement into Southeast Asia and provides Nvidia with another major Asia-Pacific commercial deployment. Investors should monitor actual procurement volumes and the translation of this demand into Nvidia’s Data Center revenue as Rubin enters deployment.
TradingKey - OpenAI is extending more of its compute procurement to Southeast Asia. On September 8, Sydney time, Australian AI infrastructure company Firmus announced a multi-year strategic partnership with OpenAI. Under the agreement, Firmus will provide dedicated compute to OpenAI across two AI factories in Malaysia, with OpenAI becoming its anchor customer.
[Source: Firmus]
Following the partnership, Firmus's contracted compute capacity from all customers will exceed 900 megawatts. Neither party has disclosed the contract value or the exact amount of compute OpenAI is purchasing.
Malaysia is a key destination for Firmus's current expansion. According to Reuters, Malaysia has become the fastest-growing data center market in Southeast Asia, attracting growing AI infrastructure investment as AI companies scramble to lock in power and data center capacity.
What Kind of Company Is Firmus?
Firmus is an Australian AI infrastructure company, and Nvidia (NVDA) has both participated in its equity financing and served as its technology partner.
In August this year, Firmus completed a $2 billion strategic equity financing round with a post-money valuation of over $10.5 billion. Other participating institutions included Jane Street, Coatue, and Blackstone's investment platform.
Firmus currently has 7 AI factories planned across Australia, Singapore, Indonesia, and Malaysia, of which 2 are already operational and the remaining 5 are under development. The company plans to bring these projects into service over the next 24 months.
Firmus's IPO plans have also recently drawn market attention. According to media reports, Firmus is preparing for a listing on the Australian Securities Exchange this year, though the company has not yet officially announced an IPO timeline.
For Firmus, which is rapidly expanding its AI infrastructure, locking in long-term clients like OpenAI in advance helps reduce the risk of newly added computing power sitting idle and makes future revenue more predictable. However, because key terms such as contract value and procurement scale have not yet been disclosed, it is currently impossible to estimate how much revenue OpenAI will bring to Firmus.
What OpenAI's Partnership With Firmus Means for Nvidia
Nvidia is also a major beneficiary of this partnership, primarily adding another case of large-scale commercial deployment for Vera Rubin in the Asia-Pacific region.
Firmus confirmed that the AI factory built in collaboration with OpenAI will deploy the Vera Rubin NVL72 rack-scale system, utilizing Nvidia's DSX AI Factory platform and Firmus's proprietary HyperCube infrastructure. Firmus also stated that it will deploy Vera Rubin at scale in the Asia-Pacific region.
Vera Rubin is Nvidia's next-generation AI computing platform following Blackwell. The NVL72 consists of 72 Rubin GPUs and 36 Vera CPUs, with the Rubin GPUs equipped with 288GB of HBM4 memory.
Firmus has not yet disclosed the specific procurement quantity, order value, or delivery timeline for the NVL72. Therefore, at this stage, what is confirmed is a large-scale deployment plan rather than a hardware order of disclosed scale.
The Firmus project is also not OpenAI's only infrastructure deployment plan involving Vera Rubin. In February this year, OpenAI announced a further expansion of its partnership with Nvidia. Nvidia participated in OpenAI's new funding round with a $30 billion investment. On the infrastructure front, OpenAI will use 5GW of Vera Rubin computing power, including 3GW of dedicated inference capacity and 2GW of training capacity. This arrangement represents a further extension of the previous collaboration between the two parties.
In September 2025, OpenAI and Nvidia announced plans to deploy at least 10GW of Nvidia systems. Nvidia stated at the time that it would invest up to $100 billion in OpenAI as each 1GW system is progressively deployed, with plans to begin deploying the first 1GW Vera Rubin system in the second half of 2026.
In addition to directly advancing large-scale AI infrastructure development with Nvidia, OpenAI is also procuring external computing power through AI infrastructure providers such as Firmus.
This signing further extends OpenAI's computing power procurement to Malaysia, while adding another large-scale deployment project for Vera Rubin in the Asia-Pacific region.
For Nvidia investors, what is more worthy of attention going forward is how many Vera Rubin systems Firmus will actually procure, and how quickly the related demand can translate into revenue for Nvidia's Data Center business once Rubin enters large-scale deployment.
This content was translated using AI and reviewed for clarity. It is for informational purposes only.
Recommended Articles
Comments (0)
Click the $ button, enter the symbol, and select to link a stock, ETF, or other ticker.