CPP Investments commits $1.75B to EQT’s AI infrastructure strategy, betting big on data center boom
Canada's largest pension fund is pouring billions into data centers as AI demand reshapes the infrastructure investment landscape, with implications for crypto mining operations riding the same wave.
Canada Pension Plan Investment Board, one of the world’s largest institutional investors, has committed US$1.75 billion (C$2.4 billion) to EQT’s AI Infrastructure strategy. The investment, announced on July 3, centers on EdgeConneX, a data center operator that has become the backbone of EQT’s push into AI-ready computing facilities.
EQT launched its dedicated AI Infrastructure strategy in April 2026, fully seeded by EdgeConneX through transfers of minority stakes from EQT Infrastructure IV and V. EdgeConneX operates across more than 20 countries. Since EQT acquired the company in 2020, its operational capacity has scaled nearly 20 times.
The company has plans to develop more than 10 gigawatts of additional data center capacity.
CPP’s commitment represents the fund’s latest in a series of aggressive bets on digital infrastructure. In July 2025, the pension fund made a C$225 million construction financing commitment for a hyperscale data center in Ontario. Then in June 2026, it entered a partnership worth up to approximately $741 million with Indian data center operator CtrlS.
For a pension fund with a multi-decade investment horizon, data centers generate stable, long-term cash flows from creditworthy tenants and leases typically include inflation escalation clauses. EQT has positioned EdgeConneX specifically for the AI workload segment, distinguishing between traditional cloud hosting and the far more power-intensive requirements of training large language models and running inference at scale.
Several publicly traded Bitcoin miners have already pivoted portions of their data center capacity toward AI and high-performance computing contracts. Companies like Core Scientific and Hut 8 have recognized that their power infrastructure and cooling systems are valuable to AI customers willing to pay significantly more per megawatt than mining operations typically generate.
High-density computing facilities require the same core ingredients regardless of whether they’re training neural networks or hashing SHA-256: cheap power, advanced cooling, and locations with favorable regulatory environments. The risk for crypto-native infrastructure players is that institutional money flowing into purpose-built AI facilities could bid up the cost of electricity contracts and suitable real estate.