GrowthTurtle
Data Science PhD & IEEE Fellow, Ex-Two Sigma
Equinix and Digital Realty are the real public play on the data-center permitting crisis, because they own permitted capacity that cannot be replicated for 3-5 years. A post the author replies to says new facilities cost $40-50B, take 3-5 years with no guaranteed permits, and that Anthropic and Google are instead renting SpaceX capacity at $1.25B and $920M a month. SpaceX is private, so these two landlords are how public investors get that exposure. x.com
Equinix's strengthening pricing power and widening competitive moat keep it among my favorite data center REITs. Q3 annualized bookings surged 35% to $533 million, AFFO per share jumped 18%, and management raised 2026 guidance for a third straight quarter. AI demand is real, and Equinix is a SWAN stock. x.com
Digital Realty's $3.5 billion Blackstone data center stake deal monetizes hyperscale capacity at peak demand rather than diluting shareholders, so the market's 5% selloff on the news is wrong. The AI power infrastructure trade remains structurally underowned. x.com
Digital Realty's core retail colocation business is a terrible one. It relies on short three- to five-year leases with enterprise customers such as banks, is operationally intensive with heavy customer churn, and runs a roughly two-decade-old portfolio built for air cooling and low rack densities. x.com
Equinix, Constellation Energy and Blackstone should collect as AI builders pay to get power on their own timeline. Dominion holds about 70GW of large-load requests against a 24.7GW peak, connecting about 10 a year with seven-year waits. Elon Musk personally bought New APR Energy's 1.1GW of trailer generation that runs in 60-90 days. Equipment makers gain less: 18 of GE Vernova's 20GW last quarter were slot reservations. Equinix owns the room, Constellation long contracts, Blackstone the debate. … Show more
Wide-moat investing remains one of the best ways to build long-term wealth, with Equinix, American Homes 4 Rent and Howard Hughes the latest picks. The conviction comes from hard experience: a real estate business worth tens of millions of dollars was mostly lost to a failed partnership, the financial crisis and a stroke. x.com
Digital Realty and Equinix face painful refinancing: the main credit they deserve over the past decade is for how much debt they piled up when rates were near zero, and rolling over their 2020–2021 issuance will not be pretty. x.com
Scarce power available for 2028 delivery and rising demand for that capacity favor Galaxy Digital, IREN and Cipher Mining. XCapitalMgmt highlights Galaxy management's report of delayed competing projects, stronger demand from hyperscalers and AI firms for 2028 capacity, and more selective financing. Galaxy has over 800 MW secured at Helios II but is still evaluating tenants and deal terms. x.com
Global resistance to data centers could constrain December cloud growth despite the longer agentic AI buildout. Some estimates say roughly 120 US projects worth over $130B face delays; opposition spans the UK, EU, Australia and Malaysia too. Microsoft, Alphabet and Amazon remain favored, though guidance is at risk. CoreWeave, Nebius and IREN may bounce hardest short term but look weak longer term as funding tightens; Elon Musk expects extensive neo-cloud consolidation. x.com x.com x.com x.com
Slower neoclouds such as CoreWeave face an uphill battle against SpaceX's faster data-center buildout. Demand may be effectively unlimited, but falling a year or two behind creates a $5-10B cash cost disadvantage they may not survive. Some could still pull off a miracle; SpaceX itself competes as a neocloud and has no reason to buy firms that cannot build data centers. x.com x.com x.com
Applied Digital's AI and high-performance computing data centers could expand substantially as demand grows. It is a high-risk upside bet: construction costs, financing needs and dependence on large customers could impede that expansion. x.com
