Why does Morgan Stanley expect BIP’s growth to accelerate?

  • Morgan Stanley forecasts a compound annual growth rate of 11.5% for BIP’s funds from operations (FFO) per unit in 2026–2028, above the 7.1% recorded in 2023–2025. Management hopes to raise its growth target from above 10% to above 12% over the next two to three years.
  • The company estimates that investing 2 billion dollars a year at a 15% return could generate approximately 5% growth in FFO per unit; investing 3 billion dollars at a 17% return could contribute approximately 8%. Together with inflation indexation and volume growth, these form management’s growth target.
  • The bank maintains its Overweight rating on BIP and its 48 dollar price target, forecasting potential total return of approximately 35% over one year, including a distribution yield of approximately 5%.

How large is the AI investment, and when will it contribute?

  • BIP’s disclosed AI investment framework of approximately 85 billion dollars covers power, AI data centers and computing capacity. Based on illustrative ownership stakes, BIP would need to invest approximately 2.2 billion dollars in equity capital over the next five years, or approximately 450 million to 500 million dollars a year. This is part of the company’s planned annual growth investment of 2 billion to 3 billion dollars.
  • Management targets returns above 17% on AI investments, compared with above 15% for traditional acquisitions and organic growth. As of the report, however, AI investments signed during the year were close to 100 million dollars, mainly from the Bloom power partnership. AI data center capital deployment is weighted toward the later part of the next three to five years and depends on commercial terms and return thresholds.
  • Some large projects remain at an early stage: the Kentucky campus requires definitive documentation and state regulatory approval, while the funding arrangements for the NAVER project in South Korea remain a non-binding term sheet.

What supports Data4’s expansion?

  • Data4 operates 10 campuses across six European countries, with approximately 1.5 gigawatts of contracted and development capacity combined, and plans to invest more than 20 billion euros in Europe by 2030. Its Paris hub had 120 megawatts of operating capacity in the second quarter of 2026; planned and available power capacity at the Paris Saclay campus could reach 505 megawatts.
  • The campuses connect multiple data centers to shared high-voltage electricity grids and fiber networks to share construction and operating costs. Revenue comes primarily from rent for data center space supplied with power; the key variables are leased capacity and contract prices. Data4’s long-term contracts are linked to inflation.
  • Expansion is constrained by power availability and the pace of contracting: projects above 20 megawatts in France typically face a wait of more than two years for grid connection. In the second quarter of 2026, 54% of capacity under construction in Paris was pre-leased, but almost all planned capacity had yet to secure customer commitments.

How do mature asset sales finance new projects?

  • BIP’s model is to develop data centers on land it owns with power connections secured, then sell partial interests in contracted assets with stabilized operations to fund subsequent construction. Data4 separates mature assets from its development business, and Arjun bought a 30% stake in StableCo, the mature asset portfolio, in 2025.
  • Data4 secured 3.3 billion euros in financing in 2025, comprising 2.2 billion euros of term debt for mature assets and a 1.1 billion euro capital expenditure facility for the development business. The construction plan depends on both debt and asset sales; construction delays or weaker asset prices would depress returns.
  • A previous exit provides a reference for returns: BIP reported an internal rate of return of 20% on its 2026 sale of a data center serving a major North American cloud customer. However, Data4’s customers are concentrated among a small number of large tenants, who can also build their own data centers.