THE APEX TIMES
NVIDIA and Firmus plan 360-megawatt, 170,000-GPU AI data center campus in Indonesia
The proposed Batam project is pitched at AI-native customers and is described as targeting up to $30 billion in committed revenue over six years.
NVIDIA is partnering with Firmus to build a large-scale data center campus in Indonesia designed for artificial-intelligence workloads, according to a report published Tuesday. The planned site in Batam is described as a 360-megawatt campus with capacity for 170,000 GPUs.
GPUs, or graphics processing units, are the specialized chips commonly used to run and accelerate training and inference for modern AI systems. Data centers built around large GPU fleets are increasingly viewed as a core requirement for customers deploying AI services at scale.
The Batam campus is expected to serve “AI-native” customers, the report said, reflecting a broader shift in demand toward infrastructure planned specifically for AI compute rather than being retrofitted from traditional cloud or enterprise workloads.
The report also said the project is expected to generate up to $30 billion in committed revenue over six years. “Committed revenue” typically refers to customer spending or contractual commitments that have been agreed in advance, though the report did not lay out the exact structure of those commitments or the timing of associated payments.
NVIDIA, whose business centers on accelerated computing platforms used to build AI data centers, has expanded its partner and customer ecosystem through long-running relationships with infrastructure providers and enterprises. Collaborations like this one, if executed as described, would add to NVIDIA’s reach beyond individual chip shipments into broader deployment of large AI-capable facilities.
Firmus, named in the report as the local partner behind the campus, is positioned in this arrangement as the developer that would turn AI infrastructure demand into a physical site. Such projects often require permitting, power agreements, and construction schedules that can materially affect the pace of hardware installations and revenue recognition.
As of now, the report does not specify key operational details such as which GPU configurations are planned, the timeline for construction phases, the identity of end customers, or whether the revenue figure is tied to specific contracts already signed versus expected take-rates from a pipeline. Without those particulars, investors and customers are left to wait for further disclosure or project documentation.
What to watch next is whether NVIDIA or Firmus provides additional information on the power and capacity ramp, the timeline for bringing the GPUs online, and the nature of the revenue commitments. Also relevant will be any disclosures about customer deployments or contracts that translate the campus headline numbers into measurable orders and installations.
Why It Matters
- Large GPU campuses are increasingly central to meeting AI compute demand, making site-scale projects a announcement of how quickly capacity may be added.
- If revenue commitments are tied to specific customers and timelines, they can indicate forward demand visibility for AI infrastructure providers and semiconductor suppliers.
- Power availability and deployment scheduling are major constraints in AI data center buildouts, so the disclosed megawatt target underscores the scale of the build requirement.
Sources
Key Facts
- NVIDIA and Firmus plan a data center campus in Batam, Indonesia.
- The campus is described as a 360-megawatt site with capacity for 170,000 GPUs.
- The project is described as targeting AI-native customers.
- The report says it is expected to generate up to $30 billion in committed revenue over six years.
Technology Related
Netflix faces investor jitters as content calendar and layoffs weigh on sentiment, says fund manager
A market report highlights mounting concerns around Netflix’s membership trends, its next slate of programming, and recent workplace changes, prompting fund manager Chris Versace to sell shares.
Apple’s stock marches to new highs as Bitcoin lags, fueling a debate on long-term positioning
A market comparison circulating today pits Apple’s steady rally against a Bitcoin move that has not recorded a new yearly peak, raising questions about what matters more for long-horizon investors.
Broadcom and Verizon join other firms in challenging tokenized stock listings tied to their shares
A group of nine companies, including Broadcom and Verizon, have publicly objected to the listing of tokenized versions of their shares on a cryptocurrency trading venue. The dispute raises questions about how far “tokenized securities” practices can go before running into issuer and regulatory boundaries.
Meta earnings in focus as investors weigh AI demand worries after Nasdaq weakness
With U.S. indexes breaking key technical levels, attention is turning to major technology and AI-adjacent companies, including Meta, as market participants search for evidence that spending on artificial intelligence is translating into durable growth.
Nvidia’s November track record is strong, but one variable could disrupt the pattern, market note says
A market recap points to Nvidia’s decade-long tendency to post gains in the 11th month, with an average return cited for the period. The note also warns that a single factor could break the historical rhythm.
Alphabet investors weigh AI-chip optimism after a Wall Street call for a potential 20% upside
A market commentator argued Alphabet’s stock could rise sharply if demand for the company’s AI chips expands. The case rests more on expectations than on new company disclosures.
Broadcom (AVGO) is likely to face tough competition for advanced AI chip manufacturing capacity as TSMC ramps in 2026, according to market coverage
A new market report frames Broadcom’s near-term outlook around how quickly Taiwan Semiconductor Manufacturing expands its most advanced production lines, and how crowded that capacity could become among AI chip designers.
Netflix cuts 1,000 jobs, leaving questions about what changes subscribers could actually see
A reported staffing reduction at Netflix highlights the company’s ongoing push for efficiency, but the specifics of which teams or product areas will be affected remain unclear.
Intel says Tesla will use its next-generation 14A foundry process as chip output ramps at 18A node
Intel reported progress on manufacturing output at its 18A process node and said Tesla has agreed to use Intel Foundry’s upcoming 14A technology as a customer for the foundry business.
Netflix releases trailer for Daniele Luchetti’s ‘You, Me, Us,’ set for Oct. 29 debut
The streaming company unveiled the trailer for the film starring Elio Germano and Maria Chiara Giannetta, with a Rome Film Festival premiere and a Netflix release date scheduled for Oct. 29.