DALLAS - s4story -- CoreSignal Research has published a new research report examining whether the physical infrastructure behind some of today's fastest AI capability timelines can actually scale by 2030.

The report, AI Extinction by 2030? The Infrastructure Reality Check, looks at a part of the current AI risk discussion that receives less attention: the physical systems required to support rapid growth in frontier AI.

The research compares public AI capability timelines with evidence on electricity demand, grid development, data centers, semiconductors, HBM, advanced packaging, networking, cooling and deployment timing.

One of the clearest findings is the difference between the AI development clock and the infrastructure clock. Major grid infrastructure can take roughly 5 to 15 years while several prominent frontier AI capability milestones are publicly targeted before 2030.

The report also examines the difference between announced capacity and capacity that is actually contracted, under construction, energized, delivered or operational.

More on S For Story

• Homestead Harmony Brings Friendship, Stewardship, And Self Awareness To The Sunny Stella® Series
• Regional Magazines Shine at IRMA Awards
• Matt Zavala of Zavala Law, PC Selected to the 2027 Super Lawyers List
• Sawgrass Ford President Attributes Dealerships Success to Long-Term Employees
• New Children's Book "Frolly Found a Dolly in the Park" Celebrates Empathy and the Joy of Help

Among the findings:

• ERCOT was tracking more than 438 GW of large load requests by June 2026, nearly 89% associated with data centers, but its own process applies maturity, commitment and verification filters.
• NVIDIA's GB300 NVL72 reference rack can require up to 142 kW while most operators surveyed by Uptime Institute reported no racks above 30 kW.
• Current EPRI scenarios translate to approximately 43.7 GW, 68.0 GW and 90.5 GW of average U.S. data center electrical load in 2030, compared with about 20.1 GW in 2023.

The research does not estimate the probability that AI causes human extinction and does not assume that a specific amount of compute or electricity creates AGI or superintelligence.

The 8-page report and a public 67-record evidence table are available at:

https://coresignalresearch.com/#ai2030

CoreSignal Research is an independent research publisher focused on semiconductors, AI infrastructure, data centers, power and the physical supply chains behind emerging technology.

https://coresignalresearch.com

Media Contact:
James McIntosh
Founder and Publisher
CoreSignal Research
james@coresignalresearch.com