Artificial intelligence (AI) sounded rather convenient when it was helping someone write an email. It becomes a harder sell when the infrastructure behind that email moves down the road and raises questions about electricity bills, water use, and noise.
That tension is becoming an investing issue. In July, Hamilton councillors rejected a proposed temporary data centre moratorium, but the debate demonstrated that local approval can’t be taken for granted. Residents wanted answers about costs and community benefits, not another presentation explaining that the future was exciting.
For investors, the lesson isn’t to abandon everything connected to AI. It’s to distinguish businesses with dependable earnings from projects that still need permission, financing, and somewhere to plug in.
Two stocks I’d watch are Hydro One (TSX: H) and Emera (TSX: EMA).
Source: Getty Images
Follow the bill, not just the demand
A proposed data centre isn’t the same thing as a paying electricity customer. Before investing, I’d ask who funds the connection, who pays if construction is cancelled, and whether the infrastructure remains useful if demand disappoints. A utility spending heavily on speculative demand can face trouble even when its industry’s long-term outlook looks wonderful.
Regulation provides some protection, but it isn’t a blank cheque. Regulators still decide which costs customers should bear. That’s why my watchlist favours established Canadian stocks whose businesses don’t need every announced AI campus to materialize. Neither company automatically benefits from opposition. The attraction is having a broader investment case if the buildout becomes slower and more selective.
Hydro One
Hydro One operates Ontario’s electricity transmission and distribution infrastructure. Its role is moving power, rather than betting on which technology company eventually wins the chatbot competition.
Second-quarter earnings per share (EPS) increased to $0.62 from $0.54 a year earlier. Higher regulator-approved rates and peak demand supported the improvement, partly offset by higher financing and operating costs.
The company also placed $644 million of assets into service during the quarter. Its infrastructure work includes replacing aging equipment and connecting customers and generation sources, giving it reasons to invest beyond data centres.
However, investors shouldn’t confuse necessary infrastructure with unlimited shareholder returns. Financing costs have already increased, and project approvals and cost recovery remain important. I’d watch for investments supported by demonstrated demand and clear regulatory treatment.
Emera
Emera offers another approach through regulated electricity and natural gas utilities, including Tampa Electric and Nova Scotia Power. I’d watch it as a diversified utility investment, not as a confirmed winner from data centre restrictions. Its appeal should survive without assigning earnings to AI projects that haven’t arrived.
The latest quarter provides a useful reality check. Adjusted EPS fell to $0.69 from $0.79. Higher corporate interest expense, foreign-exchange losses, and changes elsewhere in the portfolio contributed to the decline.
Meanwhile, Emera invested more than $1.7 billion in infrastructure during the first half and remained on track for its $4 billion annual capital plan. That creates potential growth, but also funding requirements. Before buying, I’d want to see improving per-share earnings and evidence that investment is strengthening cash generation.
Bottom line
The backlash doesn’t mean AI infrastructure stops getting built. It means investors should become more demanding about which projects deserve their money.
Of these two, Hydro One would be my first watchlist choice for its existing grid role and recent earnings growth. Emera would need to show clearer earnings improvement before I became more enthusiastic.
I’d hold either alongside investments offering genuine portfolio diversification. Owning two utilities doesn’t remove their shared exposure to interest rates and regulation.
