Kalshi traders see high odds that more than 5,100 data centers will be planned or operating before 2027

Kalshi traders put a 75% probability on more than 5,100 proposed and operating data centers before 2027, up from 60% two weeks earlier; Data Center Map lists more than 4,700 operating and planned facilities in the U.S. Virginia leads with more than 670, followed by Texas with 537 and California with more than 200. Expansion faces community opposition over utility costs and water use, while Oracle’s proposed New Mexico project has encountered a rejected gas-pipeline application and a reported force majeure notice related to potential higher costs if it misses its anticipated 2028 start.

Analysis

The Kalshi probability is a weak proxy for monetizable capacity: a count of proposed and operating sites can rise even as projects lack permits, power, water, or committed customers. The key market distinction is not how many facilities are listed, but how many reach energization on schedule and at viable costs. Community opposition and utility constraints could shift scarcity value toward developers with secured power and permitted sites, while raising delay and cancellation risk for projects dependent on new pipelines or grid upgrades.

For Oracle, the New Mexico dispute is a project-level warning, not evidence by itself of consolidated earnings or balance-sheet damage. The second-order risk is that schedule slippage pushes customer workloads and associated revenue recognition out while development and power costs persist; the offset is that customers may migrate to other providers rather than abandon demand. Over the next 1–3 months, verify project milestones, power arrangements, and any change to Oracle’s capex or cloud growth guidance. Over 6–18 months, local permitting and utility affordability could become a durable constraint on industry capacity, favoring operators that can secure power earlier.

Contrarian read: a rising site count may be bullish for announced investment but bearish for execution quality. Avoid trading the Kalshi odds as a proxy for data-center demand. Oracle-specific downside is plausible but not yet quantified; a relative short is only attractive if project delays broaden into guidance or customer-start slippage.

Key Decisions for Investors

• No immediate trade on the 5,100-site probability: it measures listings, not energized capacity or revenue. Track cancellations, permitting status, and power availability instead.
• Put ORCL on a relative-underperformance watch versus larger cloud peers; consider a hedged position only if Oracle confirms delayed customer go-lives, raises infrastructure costs, or trims cloud-growth expectations. Falsifier: on-time project milestones and unchanged guidance.
• Over the next 1–3 months, monitor New Mexico pipeline and data-center approvals, Oracle’s capex commentary, and evidence of customer workload migration. Treat a single project dispute as isolated unless these indicators deteriorate.
• For a 6–18 month infrastructure theme, favor exposure to power-secured and permitted capacity over developers selected solely on announced site counts; reassess if utility upgrades and local approvals accelerate.