Newfoundland Opens Door to AI Data Centres as Electrification Reshapes Energy Demand

Companies have approached the province about building AI data centres, CBC News reported on 29 September. Energy and Mines Minister Lloyd Parrott told the House of Assembly during a special session on the Churchill Falls agreement in mid-September that the government could consider selling excess power to a data centre operator.

"We said our door is open for business and we're looking forward to doing business with the world," Parrott said at the time, according to CBC. A department spokesperson, Brodie Thomas, told CBC by email that the government cannot comment on specific proponents. He said data centres may be considered where proposals align with policy objectives and deliver clear benefits to Newfoundlanders and Labradorians.

Electricity, land and cold weather

TechNL chief executive Andrea King told CBC that Labrador appears to meet what data centre proponents want: low-carbon hydro electricity, cold weather and a lot of land. She also urged caution. "It's not just 'does this make sense?' by itself. You have to look at economic costs and opportunity costs of what else you could do with that electricity," she said.

Tech journalist Paris Marx, author of Hyperscale, a book about the rise in data centres, raised a similar question in the same CBC report: whether giving electricity to data centres makes sense when better uses may exist. Marx noted some operators use closed loop cooling that recirculates water, an approach aimed at limiting water use.

The context is a Canadian building boom. CBC reported that Meta and Bell have announced billions of dollars for data centres in Alberta, Ontario and Saskatchewan this year. Nova Scotia Premier Tim Houston has set five starting conditions for any proponent, and EverWind Fuels has expressed interest in a Nova Scotia project. The Meta data centre planned for Sturgeon County, Alta., is expected to cost $13 billion and come online in two to three years, CBC reported.

Electrification cuts demand in one model

Separately, analysis published on 29 September by Hannah Ritchie, citing work by Oxford professor Nick Eyre, estimates that global final energy demand would fall from 416 to 247 exajoules in a post-transition system where suitable sectors are electrified and the rest run on hydrogen. Electricity demand would rise from 110 to 189 EJ in that model, according to the Substack post.

The piece credits the International Energy Agency with the framing that "electrification is efficiency." Road transport and buildings do most of the work in the model: electric cars and vans are assumed to use about a quarter of current energy demand, while heat pumps replace 90% of gas boilers and 10% comes from hydrogen. High-temperature industrial processes remain the hard part.

The numbers are a thought experiment, not a forecast. The author states the model does not assume efficiency gains beyond electrification or a move to hydrogen. It does not account for energy growth as countries develop, and it assigns every non-electrified sector to hydrogen. Ritchie writes that these choices likely understate the reduction in demand.

Computing's own power problem

Efficient Computer, a chip company, said on 29 September that it raised a $97 million Series B round led by TQ Ventures, taking its total raise to $173 million at a $650 million valuation. In a company blog post, CEO Brandon Lucia argued that fixed-function AI accelerators trade programmability for near-term speed. He cited Amdahl's Law to claim that the unaccelerated part of a computation sets the ceiling on system efficiency.

The company says its technology brings 10 to 100 times better energy efficiency than traditional CPU architectures, and that it is launching its Electron E1 chip at volume for embedded systems including robots, drones and wearables. Those are company claims, not independently verified figures.

The three stories point in different directions. Newfoundland is weighing new electricity demand from data centres while research suggests electrification will shrink total energy use. Efficient Computer is betting that the constraint on AI growth is energy per useful computation, not raw generation.

For European industry, the backdrop is not settled either. Oil prices moved from $66 a barrel at the U.S. crude hub in late February 2026 to $101 a barrel on 13 April after U.S. and Israeli attacks on Iran closed the Strait of Hormuz, according to an analysis published in The Conversation on 29 September. The article, by an energy economist and an international trade economist, traces where the windfall goes: largely to oil companies, then to shareholders, state budgets or sovereign funds depending on ownership.

Older material is background only. The U.S. Energy Information Administration's photovoltaic timeline, last updated on the agency's site, records that modern solar cells can reach a 40% conversion rate, against less than 1% for the first PV cells. David MacKay's 2008 book "Sustainable Energy Without the Hot Air" remains online, but its last site modification date is 29 August 2015.