Google's Quiet Entry Into Fiji Raises a Loud Question: What Is the Water Cost?

Google's planned FJ$200 million project in Fiji has been publicly framed as a major digital and economic win, tied to the South Pacific Connect subsea cable initiative and a new data centre. The government's announcement and subsequent reporting describe the investment and its connectivity benefits, but the public record found does not spell out the facility's detailed cooling design or water budget. That gap is exactly why the project deserves sharper scrutiny, especially in a country where water reliability is already a live issue.

The comparison with the Vuda-Saweni waste-to-energy incinerator is revealing. That proposal generated organized public opposition and was later rejected after its environmental assessment failed legal and technical standards, with reporting noting unresolved concerns around hazardous ash, foreign waste, public health, water, infrastructure, and cultural impacts. In other words, one project was seen as a visible environmental threat; the other arrived as a cleaner, more abstract kind of infrastructure and did not trigger the same sustained public alarm.

That difference matters because data centres are not water-neutral. Google's own sustainability material says data centres must be cooled and that the company is trying to balance energy and water trade-offs, while also announcing new water stewardship commitments in 2026. Academic and government-backed research shows that cooling choices can materially change water use: a widely cited LBNL report says medium-to-large data centres using cooling towers can consume significant on-site water, and gives an example range of 80 to 130 million gallons annually for a 15 MW facility. A more recent review in npj Clean Water notes that a 15 MW data centre can use as much water as three average-sized hospitals.

That does not mean Google's Fiji facility will automatically have a huge freshwater footprint. It means the public has not yet been shown enough information to know. The crucial unanswered questions are simple: will the site use air cooling, closed-loop cooling, evaporative cooling, recycled water, or some hybrid design; what will the daily water demand be; and from what source will that water come? Without those disclosures, claims about “draining Fiji's water” go beyond what is currently proven. The better journalistic question is whether Fiji has been asked to host a major industrial user without a clear public accounting of its water needs.

Fiji's water picture is more complex than a simple scarcity story. The World Bank's data show Fiji has substantial renewable internal freshwater resources per person, and the FAO's AQUASTAT framework makes clear that “available water resources” do not mean fully exploitable water. Fiji's Meteorological Service reported that national average rainfall in 2023 was 2,190 mm, 97% of the long-term average, making 2023 the 17th driest year in 66 years of record; 2024, by contrast, was much wetter. The lesson is volatility, not comfort: Fiji can be wet in the aggregate and still vulnerable to supply shocks, storage limits, and distribution bottlenecks.

The Water Authority of Fiji's own strategy says customers already experience intermittent and disrupted water servicing because of freshwater constraints, including drought and high demand. Current and recent disruption notices for parts of the Western Division and other areas have cited low reservoir levels, low raw water inflows, maintenance, and temporary supply interruptions. That means any major new industrial user should be assessed not only as an investment opportunity, but as a potential pressure point on an already strained system.

Internationally, this is no longer a fringe concern. Reuters reported in 2024 that a Chilean environmental court partially reversed Google's permit for a planned data centre and ordered a reassessment of climate-change impacts on an already stressed aquifer. AP reported in 2026 on concerns in Virginia about data centres and groundwater decline. Reuters also reported in 2026 that an OpenAI-linked Australian project dropped a recycled-water cooling plan because the required water infrastructure and permissions were not in place. The pattern is clear: in wealthy countries, data-centre water use is increasingly a public-regulatory issue, not just a private engineering choice.

That is why Fiji should not let this conversation be swallowed by celebration alone. The issue is not whether Google is good or bad, or whether Fiji should reject all digital investment. The issue is whether the country is being given enough transparency to weigh the trade-offs honestly. Compared with the louder fight over the incinerator, the data-centre debate has been quieter, but quiet does not mean harmless. A serious public conversation would ask for the cooling method, the water source, the expected consumption, the mitigation plan, and the legal conditions attached to approval. Anything less is not development policy. It is guesswork dressed up as progress.

If Fiji is to be a genuine champion for climate action in the Pacific, we must start acting like one. Climate leadership cannot be reduced to a branding exercise or a label used on the international stage. It must be reflected in the decisions we make at home — including how we assess major infrastructure projects, protect our natural resources, and demand transparency from investors and decision-makers.

Editorial Note: While this article is not about cannabis, Cannabis Fiji is committed to more than cannabis law reform. We advocate for environmental stewardship, transparency, public accountability, and evidence-based policy on issues that affect the health, wellbeing, and future of all Fijians. Protecting our natural resources is a responsibility that extends beyond any single cause.