SoftBank Group, Grab, and PETROS signed a Joint Development Agreement on November 9, 2026, committing US$1 billion in seed capital toward a five-gigawatt AI compute campus in Sarawak, Malaysia [1]. The deal positions the consortium as a first-mover in Southeast Asian sovereign AI infrastructure, directly addressing the power and land constraints that now determine how fast new AI capacity comes online globally [2]. With alignment to Malaysia's AI Nation 2030 mandate and a mandate to attract hyperscaler customers, the project targets a rapidly expanding APAC market that generated US$78.0 billion in 2024 and is projected to reach US$180.1 billion by 2028 [3][1].
What is Covered in this Article
- JDA structure and capital allocation: US$800M SoftBank, US$100M Grab, US$100M PETROS [1]
- Sarawak's energy advantage as a solution to global AI power constraints [2][1]
- Wholesale platform strategy targeting hyperscalers and sovereign customers [1]
- Regulatory alignment with Malaysia's AI Nation 2030 and Grab's demand channel [1]
- Market timing against Futurum's AI data center deployment growth forecast [2]
The News: SoftBank Group, Grab Holdings, and Petroleum Sarawak Berhad (PETROS) signed a Joint Development Agreement on November 9, 2026, announcing plans for US$1 billion in initial seed capital to develop AI infrastructure in Sarawak [1]. The investment comprises US$800 million from SoftBank Group, US$100 million from Grab, and US$100 million from PETROS, to be deployed in stages tied to development and customer contract milestones [1]. The long-term target is five gigawatts of AI compute capacity, with SoftBank as majority owner and lead developer [1]. The JDA advances a Framework Agreement signed in Kuching on October 8, 2026 [1]. Once operational, the facilities are expected to rank among the largest AI infrastructure of their kind in Asia [1].
SoftBank Bets $1B on Sarawak as Southeast Asia's AI Power Hub
Analyst Take: This deal is less about capital and more about geography. SoftBank has identified Sarawak's energy surplus as a structural advantage at precisely the moment when power, land, and grid access determine how fast new AI capacity comes online globally [2]. By locking in a majority ownership position now, SoftBank captures the steepest part of the AI infrastructure deployment curve before construction and power lead times begin to bind capacity additions.
Power Scarcity Makes Sarawak a Strategic Asset
The single biggest constraint on global AI infrastructure expansion is no longer chips or capital, it is power. Futurum's Always-Live Data Center Market Model projects annual AI data center power deployments rising from 11.7 GW in CY2025 to 76.1 GW in CY2030, with growth peaking at 113.8% in CY2026 and 87.8% in CY2027 before slowing sharply to 29.5% in CY2028 [2]. Sarawak's abundant energy resources, PETROS's role as Master Developer of the Kuching Low-Carbon Hub, and the state's active Sarawak Gas Roadmap directly address this constraint [1]. North American and European markets face saturated grid queues and permitting delays. Sarawak offers an alternative: established power generation capacity, low-carbon infrastructure, and a government actively seeking to attract AI investment [1]. That combination is rare and increasingly valuable.
A Wholesale Platform, Not a Captive Asset
SoftBank is not building this campus for internal use. The JDA explicitly targets additional investors and hyperscaler customers to support long-term expansion [1]. The intended customer base spans governments, multinational companies, and small businesses across Southeast Asia [1]. This wholesale positioning is strategically sound. Futurum data shows that AI Infrastructure, the accelerator capacity that labs and enterprises rent, is projected to peak at 78.2% growth in CY2027 as silicon shipped in 2026 becomes live capacity [2]. A five-gigawatt campus coming online in that window, with no single anchor tenant dependency, maximizes optionality. SoftBank's majority ownership and lead developer role also aligns with where value accrues in the AI stack: the two hardware submarkets account for 83.0% of AI stack revenue, concentrated among those who build data centers rather than those who merely operate them [2].
Policy Tailwinds and a Built-In Demand Channel
The project carries meaningful de-risking from its policy alignment. Development proceeds in close collaboration with Malaysia's Federal Government and the Sarawak Government, anchored to the National AI Action Plan 2026–2030 (AI Nation 2030) and Sarawak's Post COVID-19 Development Strategy 2030 [1]. That alignment reduces permitting friction and signals sovereign backing, a material advantage for a project of this scale. Grab's role as strategic partner adds a demand-side dimension that pure infrastructure plays lack. Operating across over 900 cities in eight Southeast Asian countries [1], Grab brings an existing distribution network for AI services and a track record deploying AI tools across millions of users and merchant partners. The APAC Data Intelligence, Analytics, and Infrastructure market generated US$78.0 billion in 2024 and is projected to reach US$180.1 billion by 2028 [3], giving the consortium a large and growing addressable market from day one.
What to Watch
- Hyperscaler commitments: whether a named anchor customer signs before the end of Q1 2027, validating the wholesale platform thesis [1]
- Milestone-gated funding: whether the staged deployment of the US$1 billion seed capital proceeds on schedule as development and customer contract conditions are met [1]
- Additional investor close: which sovereign wealth funds or infrastructure investors join the cap table in Q1-Q2 2027, signaling broader confidence in the Sarawak thesis
- Power infrastructure delivery: whether PETROS's Sarawak Gas Roadmap and Kuching Low-Carbon Hub timelines stay on track to support the compute buildout [1]
- APAC competitive response: how rival AI infrastructure developers in Singapore, Indonesia, and Japan reprice or reposition as Sarawak's scale becomes apparent [3]
Sources
1. SoftBank Group, Grab and PETROS Plan to Invest US$1 …, Group, November 2026
2. 2H 2026 AI Platforms Market Sizing & Five-Year Forecast, Futurum Research, November 2026
3. 2H 2025 Data Intelligence, Analytics, & Infrastructure Market Sizing & Five-Year Forecast, Futurum Research, July 2025
Disclosure: Futurum is a research and advisory firm that engages or has engaged in research, analysis, and advisory services with many technology companies, including those mentioned in this article. The author does not hold any equity positions with any company mentioned in this article.
Read the full Futurum Group Disclosure.
Author Information
This content is written by a commercial general-purpose language model (LLM) along with the Futurum Intelligence Platform, and has not been curated or reviewed by editors. Due to the inherent limitations in using AI tools, please consider the probability of error. The accuracy, completeness, or timeliness of this content cannot be guaranteed. It is generated on the date indicated at the top of the page, based on the content available, and it may be automatically updated as new content becomes available. The content does not consider any other information or perform any independent analysis.
