OpenAI takes Firmus’s contracted capacity above 900 MW—but not yet its asset value

By
CTOL News Desk
1 min read

OpenAI’s agreement with Firmus is a stronger customer signal than a capacity signal. Firmus said on Sept. 8 that OpenAI would contract dedicated AI compute from two of its Malaysian AI Factory sites under a multi-year partnership and become an anchor customer. The company said the deal lifted contracted capacity across all customers above 900 megawatts. Reuters reported that the announcement came ahead of a rumoured Firmus initial public offering and that the company was valued above US$10.5 billion in its latest fundraising. The contract improves demand credibility; its disclosed terms leave the amount of project-finance risk removed and asset value represented by the 900MW unresolved.

Firmus entered the announcement with significant platform capital. Blackstone funds and Coatue led a US$10 billion debt facility announced in February to fund Project Southgate. In April, Firmus announced a US$505 million Coatue-led strategic equity investment subject to closing conditions. In August, investors fully committed to a US$2 billion round involving Coatue, NVIDIA, Blackstone funds and Jane Street. Firmus said the round lifted total new equity over the prior year above US$3 billion and its post-money valuation above US$10.5 billion.

Those are platform-level financing facts. The Malaysia announcement gives no OpenAI MW allocation, price, contract value, minimum payment, take-or-pay, prepayment, guarantee, termination right or commencement condition. Reuters said Firmus declined to comment on contract value. An anchor customer can improve utilization underwriting, but bankable offtake requires a clear payment obligation, term, ramp and delivery condition. OpenAI raises demand credibility; the amount of financing risk transferred from Firmus remains unknown.

900MW has the same scope issue. It is Firmus’s total contracted capacity across customers, not OpenAI’s allocation or operating IT load. The portfolio spans seven factories across Australia, Singapore, Indonesia and Malaysia; two are operational and five are under development, targeting ready-for-service within 24 months. The Malaysian share, site capex, power delivery and service date are the inputs required to convert 900MW into revenue, EBITDA and enterprise value. It is a demand-book metric, not an infrastructure-output metric.

Malaysia is earlier in Firmus’s disclosure ladder

Firmus’s other projects show what a more mature disclosure looks like. Its Batam announcement specifies a 360MW campus, up to 170,000 NVIDIA accelerators through 2027 and 2028, a partnership running to 2034, and expected US$25 billion–US$30 billion committed offtake over six years. It also describes revenue-sharing and credit-support arrangements involving NVIDIA. Batam remains project-specific, but it shows the information available when a campus and commercial architecture are defined.

Project Southgate provides a different benchmark. Firmus disclosed a multi-billion-dollar contract for approximately 18,400 NVIDIA GB300 GPUs at Melbourne and linked that deployment to the US$10 billion debt facility. For South Australia, it later disclosed a 12-year Gunvor agreement for 600MW of firm electricity, supported by 1.2GW of renewable generation and 1.5GWh of battery storage. Malaysia has two sites, an OpenAI anchor and a multi-year relationship; its project-specific power, GPU, offtake and financing terms remain undisclosed. The comparison shows Malaysia is earlier in Firmus’s disclosure and underwriting cycle.

Malaysia has created a screening regime for that next stage. The Malaysian Investment Development Authority said data-centre and cloud-computing projects accounted for RM95.8 billion of approved investment in the first half of 2026, while the Data Centre Task Force clears projects with secured power, water and green compliance. The Edge reported actual data-centre electricity use of 1,102MW against 2,050MW of approved demand and water use of 28.68 million litres a day against 55.83 million litres. Prime Minister Anwar Ibrahim has said projected supply is sufficient while non-AI applications are restricted. For Firmus, the live question is whether its unnamed sites have secured approvals and utility connections.

That gap sits beneath the IPO narrative. Reuters described the listing as rumoured and cited the latest private valuation above US$10.5 billion. Separate Australian reporting said Firmus was seeking an IPO valuation of A$20 billion–A$30 billion after a July private valuation of A$15.5 billion. Those are reported expectations, not filed public-market guidance. The needed bridge runs from capital raised and available facility to named contracted MW, operating MW, revenue and EBITDA, remaining capex and retained project economics. Public disclosures stop at capital raised and aggregate contracted demand.

OpenAI strengthens Firmus’s order book and the story it can tell prospective public investors, while the Malaysian capacity remains a development-stage asset without a public cash-flow bridge. The disclosure that would change the valuation case is a site-level contract and funding package tying OpenAI’s minimum payments to a named MW block, committed power, capex funding and a service date. The prudent reading is demand validation layered onto a heavily capitalised platform—not 900MW of financed, operating infrastructure.

Sources

You May Also Like

This article is submitted by our user under the News Submission Rules and Guidelines. The cover photo is computer generated art for illustrative purposes only; not indicative of factual content. If you believe this article infringes upon copyright rights, please do not hesitate to report it by sending an email to us. Your vigilance and cooperation are invaluable in helping us maintain a respectful and legally compliant community.

Subscribe to our Newsletter

Get the latest in enterprise business and tech with exclusive peeks at our new offerings

We use cookies on our website to enable certain functions, to provide more relevant information to you and to optimize your experience on our website. Further information can be found in our Privacy Policy and our Terms of Service . Mandatory information can be found in the legal notice