OpenAI targets Canada for data centres as energy costs derail UK expansion plans

OpenAI Shifts Focus to Canada for Data Centre Expansion

OpenAI is actively exploring data center partnerships in Canada, with the company's head of countries George Osborne citing abundant energy and land availability as primary attractions1

1

2

2

. Speaking at Prime Minister Mark Carney's Canada Investment Summit in Toronto on 14 and 15 September, Osborne emphasized that Canada represents "a very important market for OpenAI" with significant potential for investment2

2

. The summit brought together approximately 100 of the world's largest investment firms, with Carney aiming to catalyze a trillion dollars of investment over five years1

1

.

Osborne, who joined OpenAI in December as managing director and head of OpenAI for Countries—the arm responsible for taking the Stargate model abroad—praised Carney's approach to AI adoption1

1

. The former British chancellor, who served from 2010 to 2016 and notably hired Carney to run the Bank of England in 2012, described Carney's medium and long-term plan as "very impressive" and noted he has articulated "one of the clearer plans for AI adoption of a western government"1

1

2

2

.

UK Project Stalls Over Energy Costs and Regulatory Uncertainty

The Canadian exploration comes after OpenAI paused its Stargate UK project in April, citing concerns over regulation and energy costs that prevented long-term infrastructure investment1

1

. The British project was designed to run 8,000 Nvidia GPUs in northeast England, with plans to scale to 31,0001

1

. British industrial electricity costs roughly four times what it does in the United States, Finland, Norway and Sweden—a stark differential that makes AI infrastructure development economically challenging1

1

.

The UK faces additional infrastructure constraints, with grid connection requests jumping dramatically from 41 gigawatts in November 2024 to 125 gigawatts by June 20251

1

. Further complicating matters, the Guardian revealed that OpenAI had never actually visited the site of its flagship British project, and the government had publicized a GBP 1.9B contract that was never signed1

1

. Regulatory uncertainty persists around copyright issues, with a proposed text and data mining exception meeting opposition from creative industries and no law materializing1

1

.

Norway Success Demonstrates Viable Model

OpenAI's successful Stargate Norway deployment illustrates what works when energy and land align. The Norwegian facility runs on Narvik hydropower with 230 megawatts to start, targeting 100,000 Nvidia GPUs by the end of this year1

1

. This represents the same two factors Osborne highlighted in Toronto: energy and land1

1

. The contrast between Norway's renewable hydropower infrastructure and Britain's expensive grid demonstrates how AI infrastructure strategy increasingly depends on access to affordable, reliable energy sources.

Implications for AI Development and Global Competition

OpenAI's pivot toward Canada signals how energy economics and land availability are becoming decisive factors in where AI companies build infrastructure. Osborne noted that Carney can look beyond recent trade disruptions between Canada and the United States to articulate clear medium-term and long-term plans2

2

. Meanwhile, the EU is attempting to address its competitiveness gap through a gigafactories call closing on 12 November, expecting more than EUR 20B of private investment across up to seven sites, with France already submitting a $10B bid1

1

. Watch for how Canada's investment climate and energy resources position it against European competitors, and whether OpenAI investment opportunities materialize into concrete projects that avoid the missteps of the UK announcement.