The ₱2-trillion plan and the ₱19-billion department

What the plan is

The Department of Information and Communications Technology published the Philippines AI+ Infrastructure Masterplan 2026–2033 in September 2026: 121 pages, six pillars — connectivity, AI compute and data centres, energy, workforce, policy, demand. Its anchor target is 1.5 gigawatts of AI data-centre capacity by 2033, from a baseline of roughly 50 megawatts, centred on a proposed Clark–Bataan hub. It cleared INFRACOM and the Economic Development Council in mid-2026.

Two things worth stating before any number below. The plan's headline moved from $30B in its June version to $34.4B in September, inside a single approval cycle, so every figure here is dated. And the Clark–Bataan hub is the plan's proposed anchor site — it is not a named, approved project anywhere outside the plan itself.

Around 38 of the plan's money figures are footnoted “Internal Analysis” with no published model. This page reports what the plan claims; it does not endorse the arithmetic. On the plan's own terms the total is also understated: the three-site option it adopts costs $17.871B where the headline assumes $14.647B, and the plan never restates it.

Who has to deliver it

The plan names six lead agencies. Five are top-level departments with pages on this site. This is what each is proposed to receive for FY2027.

DICT — co-chair, and lead on two of the six pillars — is 0.27% of the national proposal. Even the plan's residual public gap of $6.2B over 8 years works out to roughly ₱48.76B a year, about 2.5 times DICT's entire proposed budget. That last figure is the only one on this page that converts dollars to pesos: at 62.91 to the dollar, September 15, 2026. The plan publishes no exchange rate of its own.

2 of the five are proposed to shrink, in a national budget rising +5.99%. Whatever the reasons, this is not a budget leaning into a new eight-year technology programme.

What the budget actually says

Inside DICT, the lines closest to the plan's first two pillars are small, and they are not growing. These are enacted figures, every year the data covers.

Read the broadband series carefully: the API carries it as two program families, “National Broadband Plan” and then “National Broadband Program”. That is a rename, and this site's own checklist warns against reading a rename as a cut. The decline within each family is real; the comparison across the rename is ours.

Then the finding: no line in the FY2027 proposal names the masterplan, or Clark. That was checked four ways — this API's program index, DBM's own FY2027 NEP volume for DICT, the President's Budget Message, and the FY2027 BESF including its public–private pipeline. So the honest framing is not “the plan says one number and the budget appropriates another”. It is that nothing in the budget is labelled with the plan at all.

What the budget chose instead

The FY2027 President's Budget Message describes an Infrastructure Flagship Projects programme of 201 projects worth ₱10.277 trillion, 133 of them physical connectivity. Every flagship it names by title is a road, a bridge, or a flood-control project. Its only AI mentions are DICT workforce lines. The plan calls itself “a flagship public–private partnership opportunity”; the nearest thing to an AI hub in the FY2027 pipeline is a ₱2.50 billion fibre-duct scheme for New Clark City.

In fairness to the plan: it frames its public tranche as catalytic — land preparation and enabling works meant to draw in private money. It was never going to look like a ₱10-trillion road programme. Absence from the flagship list is a real finding about sequencing and priority. It is not proof of anything larger.

The gap, named carefully

A plan asking about $1.69 billion of public money a year, owned by five departments with ₱82.34B between them, led by a department at 0.27% of the national proposal, in a budget where no line carries its name.

Its limits, in the same breath. Most of the plan is private capital that would never appear in a budget, so a small DICT allocation is not by itself evidence the plan is unfunded. The peso conversion is ours. The FY2027 figures are a proposal, not a law. And the plan was approved after the FY2026 budget was already enacted, so FY2027 is the first real test — one cycle, a few weeks after publication.

The budget has not caught up with the plan. That the evidence supports. That the government does not mean it is a different sentence, and this does not support it.

What would make this checkable

There is a limit worth naming, because it shapes everything above. The year-by-year lines on this page are enacted budgets. For the FY2027 proposal the public data stops at programme level — the whole of DICT's information-infrastructure programme is a single block, and there is no way to ask what sits inside it. The year you would most want to read line by line is the one that is hardest to.

The plan promises quarterly agency reports and an annual report to INFRACOM. None of that is public data today, and neither is the part of the budget where a plan like this would actually show up: releases and disbursements, as opposed to proposals. This site is built from the proposal stage alone, which is why it can tell you the masterplan is not in the FY2027 proposal and cannot tell you whether a peso has moved toward it.

That is the CADENA Act argument, arriving from a new direction. The real test is FY2028: the plan's first-phase milestones start in 2026, so FY2028 is the first budget with no scheduling excuse. This is a story to open, not to close.