AI Growth Is Outpacing Moore's Law: $500 Billion Needed Every Year To Build Data Centers, Says The Kobeissi Letter
News Summary
Market commentator The Kobeissi Letter warns that the global artificial intelligence (AI) boom has broken Moore's Law, with compute demand growing at twice the historical rate, creating a massive shortage. To meet current demand, $500 billion must be invested in data centers annually until 2030, with global data center spend projected to hit $900 billion by 2028. The analysis states that compute has become the world's "most valuable commodity," and data centers face an $800 billion revenue shortfall. The value of U.S. data centers under construction has surged 400% since 2022 and will soon exceed that of office buildings. AI-driven power demand is projected to quadruple over the next decade, with data centers consuming 1,600 terawatt-hours of electricity by 2035, representing roughly 4.4% of global supply. The Kobeissi Letter characterizes the AI boom as a "generational opportunity," with AI-related projects making up nearly 40% of capital expenditures across the S&P 500, suggesting early adopters are poised to profit. The commentary arrives amid an intensifying debate over whether the AI boom resembles the dot-com bubble of the early 2000s. While GQG Partners, OpenAI CEO Sam Altman, Amazon founder Jeff Bezos, and Meta CEO Mark Zuckerberg have warned of potential overvaluation, Goldman Sachs argues this cycle is driven by fundamental growth in AI adoption and corporate investment rather than speculation.
Background
Moore's Law, an empirical observation by Intel co-founder Gordon Moore in 1965, posits that the number of transistors on an integrated circuit doubles approximately every two years, leading to exponential growth in computing power and a corresponding decrease in cost. This trend has been a driving force behind rapid advancements in information technology for decades. Presently, the accelerated pace of artificial intelligence (AI) technology, particularly the rise of large language models and generative AI, has created unprecedented demand for computational power. These AI models require immense computing resources for both training and inference, leading to a surge in demand for high-performance chips, data center infrastructure, and energy supply. Investor enthusiasm for the AI sector is robust, yet it has also sparked a debate regarding the sustainability of current valuations and investment levels. Some observers draw parallels to the dot-com bubble era, while others argue that AI's transformative potential and the tangible corporate investments and applications differentiate it fundamentally.
In-Depth AI Insights
What are the true investment implications of AI compute demand outstripping Moore's Law? - This signifies that AI innovation and growth will no longer be solely constrained by chip-level advancements, but increasingly by infrastructure (e.g., data centers, power) and supply chain bottlenecks. Capital-intensive investment will become a critical determinant of AI leadership, rather than purely algorithmic or chip design superiority. - The availability of traditional infrastructure resources like energy and water will become a core strategic consideration for AI development. This could catalyze significant investment opportunities in energy production, storage, and transmission, and potentially trigger geopolitical competition over these critical resources. - The slowing of Moore's Law coupled with doubling demand will accelerate the rise in computing costs, potentially making AI technology more expensive to adopt broadly and favoring large tech companies with substantial capital and existing infrastructure to further consolidate their market dominance. What does the analogy "data centers are the new oil" signify for the global economy and geopolitics? - Compute power and data centers will transition from purely commercial assets to national strategic assets, akin to historical oil reserves. Governments may increase subsidies and protection for domestic data center construction and semiconductor supply chains to secure national data sovereignty and technological leadership in the AI era. - The surging demand for resources like land, electricity, and water required for data center construction and operation could intensify competition and environmental pressures in resource-scarce regions. This will drive investment towards renewable energy, smart grids, and efficient cooling technologies. - Nations and regions with advanced AI infrastructure will see a significant boost in their economic and strategic influence. This could lead to a redistribution of global technological power and the emergence of new forms of "compute resource nationalism," impacting international trade and technological collaboration. Despite Goldman Sachs' support, what are the fundamental similarities and differences between the AI boom and the dot-com bubble, and how should investors weigh them? - Similarities exist in the massive capital expenditures and fervent enthusiasm for emerging technology, alongside potential valuation bubbles in certain segments. Over-optimism about the future and irrational exuberance are common historical patterns. - The fundamental difference is that many dot-com companies lacked clear revenue models and physical assets, whereas current AI CapEx (e.g., data center construction) is investing in tangible, high-value physical infrastructure with clear utility and commercial demand. AI technology is already demonstrating real productivity gains across multiple industries, underpinned by robust technological foundations. - Investors should weigh that while AI's long-term potential is immense and supported by real-world applications, the market may exhibit short-term over-exuberance for certain companies or tech niches. The key is to identify AI infrastructure providers (such as chip manufacturers, data center operators, power equipment companies) with core technology, sound business models, and sustainable profitability, rather than merely concept-driven or highly valued "story stocks."