MarketBeat
29 May 2026, 07:05 UTC · 13h ago
Source · https://www.marketbeat.com/instant-alerts/core-scientific-targets-ai-data-center-deals-as-build-costs-climb-2026-05-29/
Read source
Search tags
Story key points
5 claims · impact-rated
Core Scientific expects to reach approximately 450 megawatts of live capacity by the end of the summer, up from 245 megawatts currently commissioned. — Significant capacity growth directly translates to increased revenue potential and operational scale in the high-demand AI infrastructure market.
+0.60Data center build costs have increased from $8 million to between $11.5 million and $12 million per megawatt due to labor inflation. — Rising CapEx requirements per unit of capacity put pressure on margins and require more capital to achieve the same growth targets.
-0.40Core Scientific has secured sufficient power for its development pipeline for the next three years. — Power availability is the primary bottleneck for AI data centers; having a three-year runway provides a competitive moat and growth certainty.
+0.40Continue reading
6 related stories
The company is shifting designs toward 800-volt architectures and NVIDIA Vera Rubin cycle requirements to potentially lower future costs to $10 million per megawatt. — Technical adaptation to new hardware cycles can offset labor inflation and improve long-term efficiency.
+0.20Core Scientific maintains a 'hardline requirement' for credit backstops from investment-grade counterparties when dealing with non-investment-grade neoclouds. — This risk-mitigation strategy protects the company from the volatility and potential insolvency of smaller AI cloud providers.
+0.20Ticker attribution
Model heads
The company is aggressively expanding its AI infrastructure, securing power and land, and positioning itself for long-term contracts.
Mentioned as a primary driver of infrastructure demand through its chip cycles and potential role in supporting market winners.
Mentioned as building a data center neighboring Core Scientific, which is a neutral operational detail.
[a_to_b] Core Scientific builds data halls specifically for NVIDIA chips (GB300) and adapts designs to NVIDIA's Vera Rubin cycle.
[mutual] Core Scientific and Google are competing for labor resources at the Muskogee, Oklahoma site.
Impact vectors
11 dimensions · 9 clusters
Market reaction
10 bid · 0 offered
No stock impact ranking available yet.
Newsfile Corp
52m ago
