A 100MW Data Centre in a Country that Cannot Keep the Lights On; More than 1 Million Gallons of Pristine Water Needed Daily to Cool Equipment

The announcement of a 100MW AI data centre in Guyana sounds grand, but it collapses under the slightest scrutiny. This is a country where blackouts remain a weekly reality, water quality cannot be trusted from the tap, and basic infrastructure struggles to support the needs of everyday citizens. Yet we are told that Guyana is ready to host a hyperscale AI facility that will depend on steady power, sophisticated cooling, close monitoring, trained technical staff, and world class reliability. The contrast would be amusing if it were not so reckless.

The Government has committed itself to a project that carries enormous cost and dramatic overrun potential. Global data centres of this size routinely run into the billions and this one will be no different. They require advanced transmission lines, redundant grid connections, dedicated water or coolant systems, hardened security, and specialised engineering. The Wales location offers none of these elements today. Everything must be built. In a country where even road projects overshoot deadlines and budgets, the scale of this undertaking invites the kind of runaway spending that has plagued similar ventures across the developing world. Guyanese citizens will be expected to celebrate the ambition while absorbing the risk.  We will spend upwards of $4 billion dollars and what is the expected return on investment.  The Government of Guyana must inform the nation.

Even more baffling is the decision to place all hopes in a single monolithic data centre. Serious countries invest in redundancy. They spread risk across multiple sites. They diversify grid sources. They protect themselves from fires, grid failures, cyber incidents, or simple equipment malfunction. A lone 100MW installation in Wales is not resilience. It is fragility disguised as progress. One outage, one fire, one systems failure, one prolonged power fluctuation and the whole enterprise goes dark. For a country selling the idea of data sovereignty and national digital transformation, that is a dangerous way to begin.

Cerebras is a respected company in AI hardware and its wafer scale chip is impressive. But this is a firm known primarily for highly specialised supercomputing equipment, not for building and operating industrial scale data centres in developing contexts. Cerebras already owns multiple datacenters across the US, so this project sounds like the people of Guyana are paying to build another datacenter for Cerebras, so what is the expected return? The promise of training programs, startups, and glowing transformation narratives is familiar language in announcements like this, yet the global record of similar arrangements shows that benefits often concentrate abroad while the host country is left with the environmental and financial liability.

None of this means Guyana should avoid high tech investment or partnerships in emerging industries. It means leaders should anchor ambition in reality. Before declaring ourselves an AI-first nation, we must become a reliable-power nation, a clean-water nation, a competent-contract-management nation. We must solve the basics that our citizens have been pleading for. Without that foundation, a project of this magnitude looks less like foresight and more like performance.

Guyana deserves real innovation, not headline chasing. A 100MW centre in the middle of chronic blackouts is not a symbol of progress. It is a warning.