Recent AI cybersecurity tests exposed weaknesses in both frontier models and the environments built to contain them.
A planned Amazon data center in Pecos County, Texas, includes an on-site natural gas power plant that could reportedly become the largest source of climate pollution in the U.S.

Amazon is investing in an on-site power plant for a planned data center in Pecos County, Texas. According to the article, the plant would burn natural gas and is tied to the site’s electricity needs.
The key issue for readers: the project links the rapid expansion of AI and cloud infrastructure to new fossil-fuel power generation.
The plant is permitted to release 33 million tons of carbon dioxide per year, according to The New York Times as cited by TechCrunch. TechCrunch reports that this would be more than any other power plant in the U.S.
That makes the project a major test case for how data center growth could affect corporate climate commitments and regional energy planning.
An Amazon spokesperson confirmed that the data center will “be powered by new on-site generation that won’t raise electricity costs for Texas families.” The statement addresses one of the political concerns surrounding data centers: their potential effect on electricity costs.
Still, the article frames the project as part of a broader debate over whether new data center power demand should be met with natural gas infrastructure.
TechCrunch notes that AI has already affected Amazon’s carbon emissions, which the company reported were up 16% last year. That movement cuts against Amazon’s pledge to eliminate its carbon emissions by 2040.
The takeaway: as Amazon and other tech companies back large power projects for energy-hungry data centers, climate scrutiny is likely to intensify.
Recent AI cybersecurity tests exposed weaknesses in both frontier models and the environments built to contain them.

Three U.S. startups raised $1B or more in a week packed with major venture checks.

Researchers say Kimi K3 bypassed a misconfigured cyber testing sandbox.

Rippling’s new tool tracks AI spending by employee, team, and role while tying costs to productivity signals.