Key Takeaways
- • Bill Ackman, in his third-quarter letter to Pershing Square shareholders, argues that Amazon (NASDAQ: AMZN), Microsoft (NASDAQ: MSFT), Alphabet (NASDAQ: GOOGL) and Meta Platforms (NASDAQ: META) are better positioned than smaller neocloud providers to clear the constraints on AI data center build-outs.
- • Ackman identifies capital, permitting, power, chip supply and political considerations as the major constraints on data center development.
- • Following Democratic gubernatorial wins in November 2026, Texas, the largest market for planned data center development, is expected to tighten rules, with Governor-elect Gina Hinojosa pledging to suspend approvals for new data centers until stronger community, electricity and water protections are enacted.
- • Developers may shift projects to states such as Louisiana and Oklahoma, while others may delay awaiting clarity on electricity pricing, water requirements and tax treatment.
- • McKinsey estimates global data center investment could reach $6.7 trillion by 2030, with about $5.2 trillion allocated to AI-workload infrastructure.
Summary
In his third-quarter letter to Pershing Square shareholders, Bill Ackman identified capital availability, regulatory approvals, electricity supply, semiconductor access and political considerations as major constraints. He argued that hyperscalers hold an advantage because their established businesses generate substantial cash flow and their strong credit profiles and high equity valuations secure external financing on relatively favorable terms. He added that financing alone does not guarantee capacity: "In addition to capital, standing up new data center capacity requires permitting, power, chip supply, and the navigation of political considerations." Ackman said that "the neoclouds and other secondary participants face meaningful challenges with each of these requirements." McKinsey estimates global data center investment could reach $6.7 trillion by 2030, with approximately $5.2 trillion allocated to infrastructure supporting AI workloads.
Neocloud providers, which supply cloud computing infrastructure for AI workloads, have benefited from capacity shortages, and major hyperscalers are among their customers, outsourcing certain workloads while keeping internal resources for operations requiring their own infrastructure. That demand supports the neocloud model, but continued revenue growth depends on bringing additional capacity online. Smaller providers generally have fewer resources to absorb construction delays, secure electricity or manage extended permitting, so their growth depends more on expanding available capacity than on raising prices for existing services.
Texas is expected to take a tougher approach after Gina Hinojosa defeated Republican incumbent Greg Abbott in the November 3 gubernatorial election. Hinojosa has pledged to suspend approvals for new data centers until lawmakers introduce stronger protections for communities, electricity consumers and water supplies. Thousands of people have also protested AI data center expansion across Europe, demanding construction moratoriums over electricity use, environmental damage and rising costs.
The consequences will depend on the scope and duration of new restrictions. Projects already under construction or holding permits are generally better placed to proceed than those still seeking approvals. The central risk, as framed in the coverage, is a slowdown in computing capacity deployment rather than an immediate halt to all construction, and decisions by incoming governors and state legislatures in 2027 will determine how much planned infrastructure proceeds on schedule.
Go deeper than the headline
Research every firm in this story, track the ones that matter and get the daily briefing — free for 7 days.