A Quick Look at Ontario's Data Centre Playbook
Some positives, some detail lacking, and some crucial things missing.
Yesterday, the Ontario Government released its draft Data Centre Playbook. I was pleasantly surprised by some aspects of it. It is not the AI-at-all-costs pitch that I might have expected from this Premier. It has some clear positions on issues like electricity costs, and ruling out financial incentives. However, it does not get into much detail, especially around enforcement, has some vague aspirations that are likely to be dashed, and leaves out some crucial pieces.
The Province’s approach is guided by three core principles:
Data centres must invest substantially, both financial and non-financial, in Ontario and local communities.
Data centres pay the full cost of electricity costs.
Ontario will only offer non-financial support to attract data centre investments.
These are then accompanied by three “Priority Assessment Pillars” that will inform whether projects are allowed to move ahead:
Advance economic development: projects that bring in significant investment, job creation and new tax revenue for the province and municipalities, support Ontario and Canada’s domestic supply chains and help businesses be more productive.
Protect data security and digital sovereignty: projects that can advance Ontario and Canada’s digital sovereignty to protect critical data and safeguard people’s information. Canadians’ data will stay in Canada.
Invest in communities and earn public confidence: projects that bring real benefits to local communities, significantly investing both financially and non-financially in host communities to drive local infrastructure developments, job creation and training, and future economic growth.
For me, the second of the principles is the most important of all of these. It is very welcome that the government has explicitly set out that ratepayers will not “bear the cost” of data centres operating. Indeed, the document explicitly states that data centres must pay their “full cost of electricity costs and invest in necessary generation and transmission upgrades for their operations to maintain the reliability of the grid.”
This is a huge issue. In places with a high number of data centres, such as Virginia, electricity prices have increased by 267% over the past five years. This has been “due to utilities needing to quickly deploy infrastructure, such as power lines and transformers, and pay extra for market-rate energy.” Setting out clearly from the get-go that these costs will not be passed down to locals is essential.
I am also pleased to see some explicit mention of the need to minimize environmental impacts such as minimizing water usage through closed-loop and waterless cooling as well as noise-reduction measures. These have been significant problems in communities with data centres, and issues where companies will just opt for the cheapest measures unless forced to reduce impacts.
Finally, it is significant that Ontario has ruled out providing any financial support to attract data centres. A large number of US states offer generous tax incentives to data centre companies, with Virginia alone losing $1.6 billion USD in tax giveaways annually. Financial incentives to attract investment are often a path that lines the pockets of companies for investments that would have happened anyway. It is important that Ontario will not go this route.
Beyond those pieces though, there is a lot to be desired. Foremost is what enforcement mechanisms will look like. While there are requirements to minimize impact, deliver significant investment to local communities, and help Ontario SMEs through access to compute, nothing is mentioned about how companies are going to be forced to deliver those commitments. History suggests that data centre companies will very quickly go back on their promises, as has happened in Chile and other countries.
Other aspects are vague or unlikely to deliver what they intend. This includes the whole pillar around digital sovereignty. As many others have argued at length, just having the data stored locally does not mean that it is “sovereign.” Wider governance, corporate ownership and many other factors are just, if not more, important.
Similarly, the intention to create high-quality jobs and apprenticeship opportunities might be disappointed. Recent Brookings research has explored the employment effects of data centres and found that while they certainly do create jobs, they are fewer than what advocates claim. Furthermore, they have found that the type of data centre (hyperscaler vs colocation facility) has a big impact on future employment. They also conclude that workers actually see little gain — wages are unaffected while home prices rise 2%-5%.
Finally, there are a few important things missing. While the Playbook talks up Ontario’s “ample land,” there is no mention of where these projects are preferred. Likely that will involve building over prime agricultural land unless there is a clear push towards brownfield sites. Already, Ontario has lost 2.8 million acres, or 18% of its farmland, over the past 35 years to non-agricultural land uses. At a time when food security is an ever bigger priority, failing to limit data centre construction on farmland could have long term consequences.
This also applies to how data centres invest in their power generation. In a North American environment that has pushed climate concerns down the priority list, data centres are increasingly turning to natural gas to power them. The US is adding 130 TWh of annual natural gas generation until 2030 to power them, almost as much as Ontario’s total annual energy use. By leaving out any reference to how data centres should be powered, it looks like Ontario will be following Alberta in prioritizing natural gas, to everyone’s ultimate detriment.
Overall then, while there are some real positives from this Playbook, there are many more questions and concerns. Fortunately, this is only the draft and the government is seeking input through a 30-day comment period. Hopefully, a future version of this will have addressed my points of concern.