Cerebras CEO Andrew Feldman has identified energy infrastructure and regulatory fragmentation as critical factors giving China an advantage in the global artificial intelligence race. Speaking on Harry Stebbings’ 20VC podcast and in a subsequent interview with Business Insider, Feldman argued that China’s centralized government enables rapid deployment of massive power infrastructure projects essential for AI development.
Feldman explained that the United States faces significant challenges due to its decentralized regulatory structure. “Our decentralized form of government has left us with sort of a patchwork of power infrastructures where even if the federal government wants to support you, their local regulations like at the city and county level of towns that can interfere with a project and set a project back, billions of dollars,” he stated. This fragmentation creates obstacles for major AI infrastructure projects, including energy production facilities and data centers that power large-scale AI deployments.
In contrast, China has successfully built extensive power infrastructure through centralized decision-making, including massive dams, coal-burning facilities, and significant solar energy investments. Feldman emphasized that while the US shouldn’t simply copy China’s approach, it needs to find ways to support AI investments more effectively within its existing democratic framework.
The Cerebras CEO proposed a five-year moratorium on state-level AI regulations as a potential solution, similar to legislation Senator Ted Cruz unsuccessfully pushed earlier this year. Cruz’s original proposal, which was stripped from President Trump’s “One Bill Beautiful Bill,” would have blocked state and local AI laws for a decade. Feldman believes this would eliminate the burden on companies like Cerebras and OpenAI, which currently must navigate different regulations in each state—what he calls “a tax on innovation.”
As a chipmaker competing in the AI hardware space, Cerebras manufactures dinner-plate-sized chips specifically designed for AI applications, differentiating itself from Nvidia’s GPU-dominated market. The company recently withdrew its IPO filing, with Feldman explaining on LinkedIn that Cerebras wants to update its financials and strategy information to reflect the rapidly changing AI landscape before going public.
Regarding US-China tech policy, Feldman diverges from Nvidia CEO Jensen Huang’s position. While Huang seeks Trump administration approval to sell advanced chips to Chinese companies, Feldman advocates against selling America’s best technology to a political adversary. Instead, he recommends incentivizing European allies and Middle Eastern partners like Qatar, the UAE, and Saudi Arabia to strengthen the Western AI ecosystem.
Key Quotes
Our decentralized form of government has left us with sort of a patchwork of power infrastructures where even if the federal government wants to support you, their local regulations like at the city and county level of towns that can interfere with a project and set a project back, billions of dollars
Cerebras CEO Andrew Feldman explained how America’s fragmented regulatory structure creates obstacles for AI infrastructure development, contrasting it with China’s more centralized approach that enables rapid deployment of energy projects.
There’s no reason we should have a patchwork of AI regulations, meaning that each company, like Cerebras or OpenAI, has to think differently in each state; that’s just a tax on innovation
Feldman criticized the state-by-state regulatory approach to AI, arguing that forcing companies to comply with different rules in each jurisdiction diverts resources away from innovation and creates unnecessary competitive disadvantages.
I think they have been able to, at a national level, stand up vast amounts of power, and they’ve done it by building massive dams, by burning coal, by doubling down on solar. But they have put together an extraordinary power infrastructure, and I think the plan benefited from some form of central decision-making
The CEO acknowledged China’s success in rapidly building energy infrastructure through centralized planning, identifying this as a key advantage in the AI race that requires massive computational power and corresponding energy resources.
We should absolutely not sell our absolute best to a political adversary
Feldman broke with Nvidia CEO Jensen Huang’s position on chip exports to China, arguing against providing advanced technology to geopolitical competitors while suggesting the US should instead focus on strengthening partnerships with European and Middle Eastern allies.
Our Take
Feldman’s analysis introduces an essential but underappreciated variable in the AI competition: bureaucratic efficiency and energy policy. While the tech industry often celebrates American innovation culture, he’s pointing to structural disadvantages that money and talent alone cannot overcome. The energy demands of frontier AI models are exponential—training GPT-4 class models requires massive data centers consuming city-level electricity. If local zoning boards can delay power plant construction for years while China builds equivalent capacity in months, the innovation advantage shifts dramatically.
His proposed regulatory moratorium is provocative and will face resistance from states concerned about AI safety, privacy, and labor impacts. However, the core tension he identifies is real: can democracies move fast enough to compete with authoritarian efficiency in strategic technology sectors? This debate will intensify as AI becomes more central to economic and military power. The answer may determine not just which companies win, but which governance models prove most effective in the AI era.
Why This Matters
This story highlights a critical but often overlooked dimension of the AI race: energy infrastructure and regulatory efficiency. While much attention focuses on algorithms, talent, and computing power, Feldman’s perspective reveals that the ability to rapidly deploy power infrastructure may determine which nations lead in AI development. China’s centralized approach enables faster decision-making for massive energy projects essential to powering data centers and AI training facilities.
The regulatory fragmentation Feldman describes represents a significant competitive disadvantage for American AI companies. The patchwork of state and local regulations creates compliance costs, delays projects, and forces companies to navigate 50 different regulatory environments—resources that could otherwise fund innovation. This issue becomes increasingly urgent as AI models grow larger and more energy-intensive.
Feldman’s call for a federal moratorium on state AI laws also reflects broader tensions between innovation and regulation. As AI capabilities advance rapidly, the question of whether to prioritize speed or caution in governance becomes more pressing. His position suggests that excessive regulatory fragmentation could hand competitive advantages to nations with more streamlined decision-making processes, potentially affecting American technological leadership and economic security in the coming decades.
Recommended Reading
For those interested in learning more about artificial intelligence, machine learning, and effective AI communication, here are some excellent resources:
Recommended Reading
Related Stories
- US Government Pivots to Full AI Regulations: Uncertain Outlook
- How AI Can Help Make Buildings More Energy Efficient
- EnCharge AI Secures $100M Series B to Revolutionize Energy-Efficient AI Chips
- The Artificial Intelligence Race: Rivalry Bathing the World in Data
- Biden hails $20B investment by computer chip maker in Arizona plant
Source: https://www.businessinsider.com/cerebras-ceo-andrew-feldman-china-us-ai-race-energy-2025-10