68.8% of ACR Respondents Support a Large AI Data Centre Near Their Community
New ACR primary research of 6,415 respondents finds strong support for nearby AI data centres, but also broad support for operators funding grid infrastructure.
American AI news and analysis focused on what is being built, bought and regulated: foundational models, compute, enterprise adoption, investment, productivity and the policy framework around them.
New ACR primary research of 6,415 respondents finds strong support for nearby AI data centres, but also broad support for operators funding grid infrastructure.
New ACR primary research finds productivity is the leading measurable AI benefit while 45% report reduced planned hiring and entry-level roles face the greatest pressure.
American Commerce Review survey data show electricity availability has become a decisive factor in AI and data-centre investment, sharpening competition between US states.
A 14,726-person ACR audience survey finds deep AI deployment and growing interest in proprietary systems, while Census data show adoption remains far less uniform across the wider US economy.
GoodVision AI's new Texas agreements pair a proposed 15MW modular AI factory with a 100-node inference pilot. The model tests whether AI capacity can scale in smaller increments than hyperscale campuses.
How Michael Intrator, Brian Venturo and Brannin McBee transformed a crypto-mining operation into a major US AI cloud company.
US senators are negotiating AI duty-of-care rules for advanced models, including risk mitigation, national-lab testing and potential federal authority to block unsafe releases.
Former OpenAI and Anthropic researcher Jacob Coxon has warned about frontier AI risk. We examine the evidence, counterarguments and implications for US AI companies and policy.
Why do frontier AI labs keep accelerating despite safety concerns? We examine the coordination problem driving OpenAI, Anthropic and the wider US AI race.
The Pentagon is reportedly in talks over a $5 billion Fluidstack loan to strengthen the US AI data-centre supply chain. Here is why it matters.
Positron raised $875 million and reached a $5 billion valuation as investors pour capital into AI inference chips and alternatives to Nvidia.
Amazon and Qualcomm are collaborating on AI data-centre chips. An SEC filing ties warrants to purchases of up to $60 billion. Here is what the deal actually commits to.
OpenAI released GPT-6 Astra on September 3, 2026. We examine its computer-use, professional-work and cybersecurity capabilities and what they mean for US businesses.
A project-level guide to Texas AI and data-center investment, including Google's $40 billion commitment, hyperscale construction, power constraints and the difference between announced projects and operating capacity.
A data-led guide to Virginia's data-center economy, including EIA electricity-demand evidence, Northern Virginia's hyperscale cluster and the policy trade-offs created by AI compute growth.
A detailed profile of California's AI economy, including its concentration of leading AI companies, data-center electricity demand, venture capital and the state's 2026 workforce-policy response.
A data-led guide to where US businesses are adopting AI, how use differs by industry and why adoption surveys need to be read carefully.
How to compare business AI adoption across US states using Census evidence, company mix and local economic structure.
A guide to measuring AI-related employment in the US using BLS occupations, job demand and industry data without treating every software role as an AI job.
How to evaluate whether AI is lifting US productivity using BLS productivity data, BEA output and business investment rather than anecdotal time savings alone.
A framework for comparing AI automation exposure by US industry using task content, occupations and business processes rather than predicting whole industries will disappear.
Nvidia's latest results show AI infrastructure spending remains exceptionally strong, with data-center revenue up 117% year on year and a $108bn quarterly forecast.
A guide to 15 major AI companies in the United States, from Nvidia, Microsoft, Alphabet and Amazon to OpenAI, Anthropic, Palantir, Databricks and CoreWeave.
Census data show US business AI use remains concentrated among larger firms, with 37% of companies employing 250 or more people reporting use in spring 2026.
Eight US cities with the strongest concentration of AI jobs, led by the San Francisco Bay Area and supported by major hubs including Seattle, New York and Boston.
How Aravind Srinivas and Perplexity built an AI-native search company around cited answers, agents and a direct challenge to Google.
How Runway's founding team turned research in generative media into one of New York's most prominent AI video companies.
How Winston Weinberg and Gabriel Pereyra built Harvey by applying generative AI to the highly specific workflows of law firms.
The Stargate Project plans up to $500 billion of US AI infrastructure investment over four years, beginning in Texas. Here is what it means for data centres, power and industrial policy.
Meta's 49% investment in Scale AI valued the company at $29 billion and brought founder Alexandr Wang into Meta's superintelligence effort. The deal gave Scale capital but complicated its position as a neutral supplier to rival AI labs.
CoreWeave's downsized March 2025 IPO raised $1.5 billion at $40 per share, below the original range. The offering gave investors their first direct chance to price a specialised GPU-cloud model built on heavy debt and concentrated customers.
Anthropic's $3.5 billion Series E arrived just after Claude Code and Claude 3.7 Sonnet. The valuation reflected a shift from conversational AI toward software-development workloads with clearer enterprise willingness to pay.
By December 2024, xAI was no longer raising capital for an abstract frontier-model ambition. It had built Colossus with 100,000 Nvidia Hopper GPUs in 122 days and was preparing to double the system.
Databricks' 2024 Series J valued the company at $62 billion and provided enough capital and employee liquidity to keep scaling privately while public software valuations remained selective.
Perplexity's late-2024 financing discussions showed how quickly AI search had moved from product experiment to strategic threat. The valuation increasingly depended on browsers, distribution and publisher relationships as much as model quality.
Cerebras' 2024 S-1 showed revenue had more than tripled in 2023 but losses remained large and customer concentration was significant. The filing turned a hardware narrative into a testable public-market business model.
xAI's Series B gave Elon Musk's AI company the capital to move from model development into a full-stack infrastructure race. The money mattered because frontier AI was becoming a contest over compute procurement as much as algorithms.
Scale AI's 2024 financing valued the company at $13.8 billion and showed that labelled data, evaluation and human feedback were becoming infrastructure for frontier models rather than a low-margin services layer.
CoreWeave's record private debt facility showed that GPU infrastructure was beginning to be financed like capital-intensive industrial capacity rather than ordinary venture-backed software.
Perplexity's $73.6 million Series B gave investors an early pure-play bet on answer-first search. The company was small, but its source-linked interface challenged the assumption that Google would own the discovery layer of generative AI.
Amazon's decision to invest up to $4 billion in Anthropic was more than a venture cheque. It tied Claude's future to AWS infrastructure and showed that frontier-model competition would be financed partly through cloud partnerships rather than independent balance sheets.
The first Condor Galaxy system paired Cerebras' wafer-scale chips with G42's capital and demand. The partnership mattered because it gave Cerebras a reference customer willing to deploy its architecture at supercomputer scale.
Databricks' $1.3 billion MosaicML acquisition moved the company upstream from data infrastructure into model training, setting up the broader data-and-AI platform strategy that followed.
Palantir's Artificial Intelligence Platform was built around a simple premise: enterprises would not deploy large language models at scale unless permissions, operational data and human oversight were integrated into the same system.