"Power is probably the forever bottleneck": Grid delays threaten the AI race.
The rapid development of artificial intelligence (AI) is facing significant challenges due to limitations in the U.S. electrical grid's ability to keep pace with growing demands. Goldman Sachs warns that energy bottlenecks could hinder the AI industry's progress, citing a projected 170% increase in global data center power demand between 2025 and 2030, with over 60% of that growth expected in the U.S. The report highlights that the U.S. currently has approximately 2,600 gigawatts of proposed power generation and storage projects waiting for grid connections, more than double the current installed capacity. However, many of these projects may not proceed due to lengthy permitting processes, which can take up to five years from application to commercial operation. This delay could lead to postponed financial returns and increased risks for companies investing heavily in AI infrastructure before securing reliable power sources. While the expansion of AI is driving substantial investment in utility sectors, with capital plans reaching $1.3 trillion between 2026 and 2030, slower deployment might also prevent overinvestment by allowing actual demand to align with spending.
Nvidia holds the top spot in the second annual ranking of America’s Growth Leaders, a list compiled by Time and Statista highlighting top-performing publicly listed U.S. companies. Artificial intelligence continues to drive American business expansion, with several firms benefiting from the surge in data center construction. However, a new report from Goldman Sachs warns that energy constraints pose a severe threat to this momentum, describing power as a permanent bottleneck for the industry. The ranking features Comfort Systems USA at number two. The firm provides mechanical, electrical, and heating, ventilation, and air conditioning services for commercial properties. Its data center segment has fueled significant recent growth. Chief Operating Officer Trent McKenna stated during an April earnings call that advanced technology, primarily data center work, now accounts for 56 percent of revenue. He added that this sector remains the largest driver of the company’s pipeline and backlog. Comfort Systems has benefited from establishing operations in Tier 2 cities, locations frequently chosen for new data centers. In July, the company reported a 50 percent revenue increase compared to the same period last year. Other ranked companies include Howmet at number six and Argan at number five. Howmet manufactures engine components for aircraft but faces rising demand for blades and vanes for gas turbines used in data centers due to the AI boom. Argan, an energy infrastructure provider, supplies power services and energy storage solutions to data center projects. Goldman Sachs issued a report indicating that global data center power demand will surge 170 percent between 2025 and 2030. More than 60 percent of this growth is projected to occur in the United States. Jason Tofsky, the bank’s global head of digital infrastructure investment banking, noted in the document that power is likely the enduring limitation. The bank estimates total capital spending across computing, data centers, and power infrastructure will reach approximately 7.6 trillion dollars between 2026 and 2031. Physical infrastructure development lags behind technological advancement. The U.S. currently has roughly 2,600 gigawatts of proposed power generation and storage projects waiting for grid connections. This volume exceeds twice the grid’s existing installed capacity, although many proposals may never proceed. The median duration from a grid connection application to commercial operation is about five years. Financial risks accompany these delays. Goldman Sachs forecasts 420 billion dollars in global hyperscaler debt issuance in 2027, representing an increase of more than 60 percent from its 2026 estimate. Debt is expected to finance roughly 35 percent of hyperscaler capital expenditures next year. Hyperscalers account for just 40 percent of AI ecosystem credit fundraising, indicating widespread borrowing. Since January, 86 projects have been canceled or significantly delayed. Of these, 76 were directly linked to local opposition. The report suggests the AI competition is fundamentally about securing sufficient power. This dynamic could lead to a longer, less predictable buildout and delayed financial returns for investors. Conversely, slower deployment might prevent over-investment by allowing actual demand to align with spending. Utilities are transforming from slow-growing sectors into major beneficiaries of AI investment. Utility capital plans are anticipated to hit 1.3 trillion dollars between 2026 and 2030. Technology giants are exploring direct funding arrangements for power generation, transmission lines, and electrical infrastructure required by their data centers. Recent market activity reflects investor caution regarding the sustainability of such capital expenditure. In October alone, three planned initial public offerings for data-center providers were shelved. These ventures had backing from AI powerhouses including Nvidia, Blackstone, and SoftBank.
Go to the primary sources (6)
The official sources this coverage is built on. Read them directly to bypass framing.
The article discusses the evolving challenges facing the AI industry, shifting from technical bottlenecks such as chips and power to financial constraints. Despite high-profile developments like Anthropic's potential $1 trillion IPO and rapid expansion of data centers, concerns are growing among lenders and financiers about the sustainability of current investments. Financial leaders warn of impending 'shortages of capital,' citing rising risks and tighter lending conditions. Examples include the difficulty in securing favorable terms for projects like Volta’s Norwegian data center and the increased cost of insuring against defaults by companies like SpaceX. While optimism persists in equity markets, the article draws parallels to historical financial bubbles, suggesting caution is warranted.
Bias read (Center): The article provides a balanced overview of the financial challenges facing the AI sector without overtly favoring any particular viewpoint. It cites warnings from financial institutions and highlights both the continued growth in AI investment and the emerging risks, presenting multiple examples of
Why factuality (80): The article accurately reflects the shift in bottlenecks from hardware to capital, citing a Singapore bank CEO and unnamed financiers. It provides concrete examples like Volta’s data center and CoreWeave’s bonds, aligning with the broader consensus on financial strain in the AI sector. Some details
Why objectivity (70): The piece maintains a relatively neutral stance, presenting both the growth of the AI industry and the emerging financial challenges. However, it uses phrases like 'Yikes' and quotes Elon Musk, which may subtly favor certain perspectives. The tone remains mostly objective but occasionally leans towa
In an interview with Democracy Now!, tech journalist Paris Marx discusses his new book 'Hyperscale: The Ambition and Excess of Big Tech's Data Empires.' Marx critiques the environmental impact of large-scale AI development by major technology firms, noting that companies previously committed to climate goals have abandoned those targets to prioritize generative AI and data center expansion. He highlights how some AI companies have resorted to constructing on-site gas plants due to difficulties connecting to existing electrical grids, thereby increasing reliance on fossil fuels. Marx argues that claims about AI potentially becoming so advanced that it might pose existential risks to humanity are more science fiction than reality, suggesting such narratives serve to divert attention from the tangible negative effects of current AI technologies.
Bias read (Progressive): The article presents a critical perspective on the environmental and societal impacts of big tech companies, particularly focusing on their shift away from climate commitments towards AI development. It frames concerns around AI's potential dangers as speculative and used as a distraction from real,
Why factuality (60): The article focuses on the political and social implications of data centers, particularly in northern Virginia, and critiques of AI regulation. However, it lacks sufficient detail on the specific event being discussed, relying heavily on the interviewee’s perspective without independent verificatio
Why objectivity (55): The tone is clearly critical of Big Tech and highlights environmental and regulatory issues. It frames the debate around data centers as a political struggle, which introduces a clear ideological bias. The article does not present opposing viewpoints or balance the narrative effectively.
AxiosIndependentCenterFactual: no official source document/info detectedObjective 90
The rapid development of artificial intelligence (AI) is facing significant challenges due to limitations in the U.S. electrical grid's ability to keep pace with growing demands. Goldman Sachs warns that energy bottlenecks could hinder the AI industry's progress, citing a projected 170% increase in global data center power demand between 2025 and 2030, with over 60% of that growth expected in the U.S. The report highlights that the U.S. currently has approximately 2,600 gigawatts of proposed power generation and storage projects waiting for grid connections, more than double the current installed capacity. However, many of these projects may not proceed due to lengthy permitting processes, which can take up to five years from application to commercial operation. This delay could lead to postponed financial returns and increased risks for companies investing heavily in AI infrastructure before securing reliable power sources. While the expansion of AI is driving substantial investment in utility sectors, with capital plans reaching $1.3 trillion between 2026 and 2030, slower deployment might also prevent overinvestment by allowing actual demand to align with spending.
Bias read (Center): The article discusses technological and infrastructural challenges related to AI development and does not present any overtly political stance or biased framing. It focuses on economic and technical factors affecting the AI industry without emphasizing political implications or taking sides in any争议
Why factuality: no official source document/info detected
Why objectivity (90): The article presents facts objectively, avoiding overt bias or emotional language. It explains the situation without favoring one perspective over another, maintaining a balanced tone throughout.
MarketWatchIndependentCenterFactual: no official source document/info detectedObjective 85
In October, three initial public offerings (IPOs) by data-center providers supported by major AI companies such as Nvidia, Blackstone, and SoftBank were postponed. The decision reflects investor skepticism regarding the long-term viability of significant capital expenditures required for these ventures. This trend highlights growing uncertainty in the global AI sector, where high costs and uncertain returns are causing hesitation among investors. The setbacks come amid broader concerns about the financial feasibility of expanding AI infrastructure.
Bias read (Center): The article presents factual information about the postponement of IPOs and cites investor concerns without taking a clear stance or using biased language. It does not favor any particular political perspective or ideology.
Why factuality: no official source document/info detected
Why objectivity (85): The article presents facts neutrally, focusing on market reactions and financial indicators without overt bias or emotional language.
TIME magazine, in partnership with Statista, released its second annual ranking of America's Growth Leaders, identifying top-performing publicly listed U.S. companies driven largely by artificial intelligence trends. Nvidia retained the number one position, while Comfort Systems USA ranked second due to significant revenue growth from data center construction, which now accounts for 56 percent of its total revenue. The report notes that Comfort Systems benefited from its presence in Tier 2 cities where new data centers are frequently located, reporting a 50 percent year-over-year revenue increase in July. Other highlighted companies include Howmet, which saw rising demand for gas turbine components used in data centers, and Argan, which expanded its role in providing power services and energy storage solutions for these facilities.
Bias read (Center): The article focuses exclusively on corporate performance metrics, market rankings, and industry trends related to AI infrastructure. There is no discussion of government policy, regulation, or political actors, rendering the content apolitical and neutral in tone.
How each side covered it
The same event, grouped by the political lean of the outlets covering it.
progressive
center
conservative
★
How each side covered it
Support independent, bias-aware news and unlock the social pulse, community voting, and every other Supporter feature.
United States
