DICT’s AI infrastructure plan estimates $34.4 billion and targets 1.5 GW by 2033

What was launched

DICT and the Asian Development Bank launched the Philippines AI+ Infrastructure Masterplan for 2026–2033 on 8 September 2026 at Crowne Plaza Manila Galleria.

The plan ties government AI services, business-process work, data centers, connectivity, power, water, and training into one roadmap. A chatbot, a reskilling class, and a power line show up together because the services need computers, and the computers need electricity and people who can use them.

What $34.4 billion means

Chapter 5 of the masterplan estimates about $34.4 billion across the program through 2033. About $13.5 billion, or 39 percent, is framed as public. About $21 billion, or 61 percent, is framed as private.

The same chapter says about $18.2 billion is already partly covered by existing national programs. That leaves an estimated gap of about $16.3 billion. This is a funding model. It is not $34.4 billion already secured, contracted, or spent.

The original flagship compute and data-hub project is estimated at about $14.647 billion, with a split near 37 percent public and 63 percent private.

50 megawatts today, 1.5 gigawatts on paper

The plan uses a baseline of about 50 megawatts of installed AI-ready data-center capacity. It targets 1.5 gigawatts by 2033. The country does not have 1.5 gigawatts of this capacity now.

Phase 1, from 2026 to 2030, targets about 400 megawatts. Phase 2, from 2030 to 2033, targets about 1,100 megawatts more.

Clark is the Phase 1 anchor. The document later studies a distributed layout of 800 megawatts in Clark, 350 in Batangas, and 350 in the wider Calabarzon area outside Batangas. That option raises the modeled cost to about $17.871 billion. Later sites still depend on demand, investment, and site checks.

Power, water, and workers

The plan sets a target of 40 percent renewable energy for AI clusters by 2033, with dual grid feeds and storage. Its own analysis says Luzon congestion is rising, and that a 1.5 gigawatt load could exceed safe grid margins without new generation and transmission.

It also names compute supply, investor concentration, climate and environmental limits, and the pace of AI adoption as risks.

Workforce targets include reskilling 1.3 million IT-BPM professionals for AI-enabled services by 2033. The document projects more than 500,000 AI-related jobs and 175,000 jobs from infrastructure projects. Those are projections, not jobs already created.

What to watch

• Ask whether a new data-center claim names a signed contract, a power connection, or only a plan.
• Treat job and GDP figures from this roadmap as targets until a funded program publishes results.
• If you work in IT-BPM, look for a named reskilling course, not only the 1.3 million target.
• For a site near you, the practical questions are power, water, and who pays.

How this was checked

The figures above follow the Pinoy AI Works News Desk check of the 121-page DICT masterplan, recorded on 24 September 2026. That pass is the basis for calling the row verified. Launch coverage from Newsbytes.PH and the Philippine News Agency matches the public announcement. A figure here is a target, an estimate, or a projection unless a later contract says the work is funded or built.

Try itCheck a claim against its original source