SoftBank, Grab, PETROS Commit $1.85B to Malaysian AI Infrastructure

SoftBank Group, Grab Holdings and Malaysia's Petroleum Sarawak Berhad (PETROS) have committed $1.85 billion in initial combined seed capital to build AI infrastructure in Malaysia, Grab announced on Monday.

The breakdown: $1 billion comes from SoftBank, $750 million from Grab, and $100 million from PETROS. The partners have signed joint development agreements for the push, which Grab said could rank among the largest AI infrastructure projects of its kind in Asia once fully operational.

How is the $1.85 billion split?

The money funds two separate projects.

• Sarawak project: $1 billion in initial investment — $800 million from SoftBank, $100 million from Grab, and $100 million from PETROS. Grab holds an option to increase its stake, which would reduce SoftBank's share but leave SoftBank as majority shareholder.
• Peninsular Malaysia project: $850 million in seed capital — $650 million from Grab and $200 million from SoftBank. Grab will lead development and bring in long-term strategic investors as the project scales.

The parties plan to build AI compute capacity across both regions with a long-term target of five gigawatts or more. They also intend to bring in additional investors and hyperscale customers as the projects grow.

What did the CEOs say?

Masayoshi Son, chairman and CEO of SoftBank Group, said AI infrastructure will power the next wave of growth, and that Malaysia has the talent and ambition to lead it.

Anthony Tan, co-founder, chairman and CEO of Grab, said the infrastructure would be built in Malaysia — where Grab began — to serve Southeast Asia with localized AI solutions and high-value jobs.

How does it fit Malaysia's plans?

The projects align with Malaysia's National AI Action Plan 2026–2030, branded AI Nation 2030. The facilities will operate under Malaysia's renewable energy guidelines and are designed for high efficiency, lower water use, and reduced reliance on the local power grid.

The deal lands as Grab posts strong numbers of its own: the company reported Q3 revenue of $1.1 billion, up 27% year on year, with record adjusted EBITDA of $237 million.

For SoftBank, the $1 billion commitment adds another compute bet to its AI portfolio. For Malaysia, the two projects anchor its AI Nation 2030 roadmap with capital from two of the region's most recognizable tech investors.