Project in Batam will be developed with Firmus and DayOne, will have infrastructure based on Nvidia technology and aims to serve companies that need large computational capacity for artificial intelligence.
The global race for artificial intelligence has just gained a new pressure point on the map. Firmus Technologies announced a strategic agreement with Nvidia to develop a large AI data center campus in Batam, Indonesia, in partnership with DayOne.
The project will have a capacity of 360 megawatts and could accommodate up to 170,000 Nvidia artificial intelligence accelerators between 2027 and 2028. The forecast places Batam, an island near Singapore, at the center of an increasingly intense dispute for energy, chips, and advanced computing infrastructure.
The financial dimension also draws attention. According to Firmus, the already projected capacity contracts could yield between $25 billion and $30 billion in the first six years of the partnership.
-
NuScale proposes revised plan for six VOYGR SMR reactors in Poland after ending U.S. project
-
Holtec and EDF Submit Joint Proposal to Install Four SMR-300 Reactors in England with Combined Capacity of 1.3 GW
-
Brazil’s Angra 3 Nuclear Plant Reaches 66% Completion; Government Faces 2026 Deadline to Decide on $24 Billion Project’s Future
-
American company drills a 1,830-meter hole in Kansas using oil well technology, places a nuclear reactor inside, and claims it will generate electricity without any cooling tower.
In practice, the agreement shows how data centers have ceased to be just technical structures hidden in the backstage of the internet. Now, they have become strategic pieces of the artificial intelligence economy.
Batam enters the route of large AI data centers
The project in Batam was designed to serve native artificial intelligence companies, growing startups, advanced model developers, and clients who need quick access to large-scale processing.
The choice of the island is not random. Batam is close to Singapore, one of Asia’s main financial and technological centers. This proximity makes the region attractive to companies seeking computational capacity without relying solely on hubs already pressured by high demand.
With the new campus, Indonesia starts to compete for space in a race dominated by countries that can gather three factors at the same time: available energy, connectivity, and the ability to build infrastructure at scale.
The project also reinforces a trend that has been gaining strength in the sector. The expansion of artificial intelligence does not depend only on more advanced models, but on physical facilities capable of powering, cooling, and operating thousands of chips working simultaneously.
Up to 170,000 Nvidia accelerators are planned in the agreement
One of the strongest points of the announcement is the forecast of up to 170,000 Nvidia AI accelerators between 2027 and 2028. This volume helps explain why the project gained immediate weight in the market.
Accelerators are fundamental components for training and operating artificial intelligence models. The greater the demand for generative systems, autonomous agents, and corporate AI applications, the greater the pressure for specialized chips.
Firmus claims that the agreement involves Nvidia platforms from the Grace-Blackwell, Vera-Rubin, and Vera generations. This structure was designed to offer cloud services with high computational power to companies that do not want or cannot build their own data centers.
The model also changes the way Nvidia participates in this type of expansion. In addition to selling infrastructure, the company is now involved in agreements related to offering computational capacity to end customers.
Project could yield up to $30 billion in the first years
Firmus estimates receiving between $25 billion and $30 billion in contracted capacity agreements during the first six years of the partnership. The number shows the size of the bet on a market that continues to accelerate, even amid doubts about costs, energy, and financial returns from AI.
This forecast is linked to the demand for ready-to-use infrastructure. For many artificial intelligence companies, the bottleneck is not just developing models but gaining access to machines powerful enough to train, test, and deliver these systems to the market.
The agreement aims to fill exactly this space. Instead of relying solely on large global cloud providers, smaller companies and regional clients are now targeting specialized AI structures, with contracts focused on high performance.
Batam, in this scenario, emerges as a new player in the competition for computational capacity in Asia-Pacific.
Firmus bets on its own infrastructure for the new phase of AI
Firmus Technologies is positioning itself as an infrastructure company for artificial intelligence. The company uses the concept of an “AI Factory” to describe data centers specifically designed to process large volumes of computation.
This logic seeks to differentiate new projects from traditional data centers. Instead of just storing data or running conventional corporate systems, these structures are built to operate intensive AI workloads, with high energy consumption and high chip density.
The partnership with Nvidia expands this strategy. Firmus will have access to a significant number of accelerators and will be able to sell computational capacity based on this infrastructure to clients who need to scale quickly.
DayOne enters as a partner in the development of the campus in Batam, reinforcing the region’s role as an expansion corridor for new digital projects in Southeast Asia.
The race for chips has become a race for territory
The agreement shows an important shift in the technology industry. The battle for artificial intelligence is not only happening within laboratories or in companies that create models. It is also happening on the grounds where data centers will be built.
Each new project requires energy, connection, licensing, rapid construction, and access to equipment that is globally contested. Therefore, countries and companies that can enable large computing centers gain an advantage in the next stage of AI.
Batam enters this game as a strategic location. The island can benefit from its proximity to Singapore and the search for new areas capable of hosting large facilities.
At the same time, the project highlights the growing pressure on the physical infrastructure of artificial intelligence. The more companies adopt AI in products, services, and internal operations, the greater the need to expand the processing base.
Indonesia gains strength in the race for advanced computing
With a planned capacity of 360 MW, up to 170,000 Nvidia accelerators, and potential contracts of up to $30 billion, the project in Batam places Indonesia in a more visible position in the global race for AI data centers.
The bet by Firmus, Nvidia, and DayOne shows that the expansion of artificial intelligence is spreading new infrastructure hubs around the world. The movement is no longer limited to the United States, Europe, or the traditional major Asian centers.
Now, islands, industrial zones, and regions close to technological hubs are also competing for space in this new digital geography.
If the schedule progresses as planned, Batam could cease to be just a location near Singapore and become one of the most important points of the new artificial intelligence infrastructure in Asia-Pacific, allowing for rapid capacity contracting without waiting years for their own constructions or relying solely on the major global clouds.
With a 360 MW campus, billions of dollars in potential contracts, and up to 170,000 planned accelerators, the project in Batam places Firmus, Nvidia, and DayOne in an increasingly aggressive race for the next generation of data centers. If the forecast is confirmed, Indonesia will occupy a relevant position in this race for artificial intelligence infrastructure in Asia.
