Inaugurate First, Ask About the Environmental Permit Later: When Decisions Outrun Research

Let me tell you about an incident that, had it not been factual, I might have mistaken for a comedy sketch.

On September 22, 2026, West Java Governor Dedi Mulyadi attended the groundbreaking of a mega project: BDx’s AI Campus 2 in Jatiluhur, Purwakarta. The investment: Rp3.5 trillion, with a total capacity of 640 MW — the largest AI data center ever built in Indonesia. Ribbons cut, photos taken, everyone smiling.

Six days later, on Monday, September 28, the very same project was temporarily halted. By the West Java provincial government. Following the governor’s own surprise inspection of the site. The reason: its AMDAL (environmental impact assessment) and building permits were not yet complete.

Read that sentence again, slowly. The same man. Inaugurating, then halting. Six days apart.

For context: this case is about more than late paperwork. An AI data center is not an ordinary server warehouse. A facility like this uses direct-to-chip liquid cooling to support extreme-density AI racks — and liquid cooling means water. A lot of water. At its later stages, the Jatiluhur project is projected to gulp around 33 million liters of water per day, with a power demand of 640 MW — the equivalent of hundreds of thousands of households.

And the location? About five kilometers from the Jatiluhur Dam. The beating heart of irrigation for Karawang, Subang, and Bekasi, and a source of raw water for Jakarta. On land like that, the question “is there enough water, will it disturb the farmers?” is not an administrative question — it is a life-and-death question for thousands of farming families. A question that should have been answered before the ribbon was cut, not after.

What makes me shake my head even more: an AMDAL is not just about pipes and waste. By law, it must cover socio-economic, socio-cultural, and public health aspects. Meaning, had the assessment been done seriously from the start, every question now making headlines — about water, about farmers, about who actually benefits — would already have been answered on paper before a single foundation was poured.

Social observer Agus Sanusi put it perfectly: “We must not let the public get the impression that the project has already been decided, while the AMDAL merely follows… when a project has been introduced and construction has begun while the environmental assessment is unfinished, the question is whether the assessment still has enough room to correct the project.”

That is the heart of the matter. It is not about being for or against AI data centers — Indonesia does need digital infrastructure. The problem is the reversed order: decisions first, research later. Ceremony first, procedure to follow. If the assessment were to find a serious problem, would a Rp3.5 trillion project that has already been inaugurated truly be cancelled? Or would the assessment be “adjusted” so the project keeps rolling?

And this is not the first time. The pattern repeats like an old song. The food estate program in Central Kalimantan was touted as a national food barn and ended up as scrubland — until the Minister of Agriculture himself admitted its failure before parliament, blaming unmechanized, unprofessional execution. Further back, Sri Lanka in 2021: its president banned chemical fertilizers overnight and forced two million farmers to switch to 100 percent organic with no consultation whatsoever. The result? Rice production collapsed 20 percent within six months, and a country that had been self-sufficient in rice was forced to import it.

The disease is exactly the same: decisions preceding research. Only the scale differs.

There is a good term for the cure: evidence-based policy making. Before deciding, define the problem carefully, study the alternatives, calculate the economic, social, and environmental consequences based on data, listen to the public — then decide. In the regulatory world this is called Regulatory Impact Assessment. Its principle is simple: a policy is not worth adopting if its negative impacts outweigh the public interest.

The AMDAL is essentially a miniature version of that principle at the project level. Ironically, that is precisely where the failure is most glaring: the document that should serve as a space for correction is treated like a stamp that can arrive late.

Fortunately, the Jatiluhur episode was stopped on day six — not after failed harvests like the food estate, not after a food crisis like Sri Lanka. And credit is due: Dedi Mulyadi himself demanded the AMDAL documents be opened transparently to the public, including the identities of the consultants who prepared them. “An AMDAL is a public good, not a private good,” he said. The right stance — and it should be the standard for every project, not just the ones going viral.

Still, one lesson must not be lost in all the noise: every major decision made without deep research into its social, economic, and public-welfare dimensions will, sooner or later, send its bill. Sometimes the bill comes as a stalled project. Sometimes as dried-up rice fields. Sometimes as public trust that cannot be bought back, not even with Rp3.5 trillion.

There’s a simple principle I live by: measure twice, cut once. In the world of leadership, the principle is the same: research first, inaugurate later. Not the other way around.

“A leader who fears research actually fears reality. And the reality avoided today will collect its dues, with interest, tomorrow.”

— Kurniawan, October 2026