The Philippines is getting one of its biggest AI computing projects so far. Filipino digital infrastructure firm YCO Cloud and Singapore-based Aolani announced on September 28 that they will deploy more than 10,000 Nvidia Blackwell Ultra GPUs — graphics processing units, the chips used to train and run AI models — in the country. The first phase is targeted to go live in the first quarter of 2027.
What YCO Cloud and Aolani announced
According to the companies' joint announcement, the machines will arrive in Nvidia's GB300 NVL72 configuration, a rack-scale design built around the company's Blackwell Ultra chips. The partners describe it as one of the first deployments of its kind in the Philippines.
The split of work is clear. Aolani handles the financing, deployment and operation of the GPUs, while YCO Cloud develops the local site and digital infrastructure. Science Park of the Philippines Inc. (SPPI) will supply the site's power, water and connectivity. The companies say the facility will draw on dedicated renewable energy sources, follow water-use efficiency benchmarks, and connect through the country's existing submarine cable systems.
"The Philippines is now ready to scale on digital infrastructure with its robust submarine cable systems," said YCO Cloud CEO Nik de Ynchausti in the announcement. Aolani CEO Nicholas Chia added that "Southeast Asia is fast becoming a key linchpin for global AI infrastructure." Officials from the Department of Information and Communications Technology (DICT) and the Economy Department also welcomed the plan in the same release.
Who will get access to the compute
The cluster is not being built for a single customer. Telecompaper's report notes the infrastructure is designed to serve government organizations, domestic enterprises and startups, with capacity for both local and international AI workloads. The announcement specifically names local government units, Philippine businesses and startups, alongside global AI-native companies.
That detail matters. Today, a Filipino AI startup that needs serious GPU capacity usually rents it from data centers abroad, typically in Singapore or the United States. Renting Blackwell-class hardware inside the country could mean lower latency, and it keeps data on Philippine soil — a growing concern for government and financial workloads.
How it fits the regional AI race
The announcement lands while the Philippines' neighbors are moving quickly. OpenAI recently chose Malaysia for two Nvidia-powered AI data centers, while the Philippines was still mapping out its own data center strategy. A concrete, dated deployment of current-generation Nvidia hardware — not a memorandum of understanding — is a different kind of signal to the market.
It is worth being precise about the scale, though. More than 10,000 GPUs is large for the Philippines, but it is a first phase, and the companies have framed it as the start of a longer buildout rather than the finished project.
What has not been announced
Several important details are still missing:
- The investment amount. No peso or dollar figure was disclosed for the first phase.
- Pricing. There is no public rate card yet for renting the compute.
- The exact site. SPPI is named as the infrastructure provider, but the specific location was not identified.
- Customers. No anchor tenants or signed clients have been named.
Until those blanks are filled in, this is a firm plan with a date rather than a running facility. The Q1 2027 launch target gives the partners about five quarters to deliver — and gives Filipino startups, universities and government agencies time to figure out what they would actually do with Blackwell-class compute available in-country instead of overseas.
