
Investors should prioritize exposure to Nuclear Energy and Small Modular Reactors, as they are identified as the primary long-term solutions for powering the massive global surge in data center demand. Natural Gas remains a vital bridge fuel for the energy transition, offering a stable investment theme due to decades of low-priced domestic supply and its role in reducing carbon emissions. The rise of Agentic AI is rapidly eroding competitive moats, creating a "golden age" for small, agile startups that can now challenge large incumbents with significantly less headcount. To mitigate extreme geopolitical risk, investors must account for the U.S. economy's "staggering" dependency on TSMC, where any loss of access to Taiwanese chips could trigger a depression-level 8% drop in GDP. Finally, shift focus toward "visionary" stock picking that targets long-term transformative products rather than short-term earnings beats, as Alternative Data has made quarterly predictions increasingly crowded and less profitable.
• Ken Griffin views AI as a component of a broader digital revolution that has been accelerating for a decade. • Citadel has used machine learning (specifically TensorFlow) for over 10 years to price and manage risk. • Agentic AI systems are significantly reducing the time for high-level tasks; for example, reproducing academic finance papers has dropped from 6-8 weeks of human labor to 2-3 hours. • Griffin notes that while AI is the current buzzword, many corporate productivity gains are actually coming from "optimization, digitization, and technology" rather than pure AI.
• Productivity over Headcount: Expect high-end firms to use AI to increase the volume of problems they can solve rather than simply cutting staff. • Erosion of Moats: AI is filling in competitive moats at "breathtaking rates." This creates a "golden age" for entrepreneurs who can now challenge incumbents with much smaller teams (3-5 people instead of 30-40). • Retraining Necessity: Specific sectors like translation (English to German) are at immediate risk, requiring urgent national focus on skills retraining.
• There is a massive, "breathtaking" demand for compute power, which is currently fully utilized globally. • Griffin argues that data centers are a critical national security asset and must be built within the United States to avoid dependence on foreign countries. • The rise of AI is driving a massive need for new power generation.
• Nuclear Energy: Griffin identifies nuclear (and Small Modular Reactors) as the primary solution for carbon-free, reliable power for data centers. • Natural Gas: Viewed as a vital bridge fuel that has already helped the U.S. reduce carbon emissions; decades of low-priced supply remain. • Investment Strategy: Data center providers should be required to build their own corresponding power generation rather than passing costs to consumers.
• The U.S. economy has a "staggering" dependency on Taiwanese semiconductor chips. • Griffin warns that a loss of access to TSMC would cause U.S. GDP to fall by 8% within six months, triggering a "Great Depression" level event. • Key industries (Boeing, automotive, consumer electronics) would cease production almost immediately.
• Geopolitical Risk: There are "no winners" in a Taiwan conflict; investors must recognize that a blockade would cause a global economic tailspin. • China’s Innovation: China is no longer just a low-margin manufacturer; they lead in 67 out of 74 critical technologies (solar, EV batteries, quantum). • Competitive Threat: The U.S. must focus on education and STEM degrees to out-innovate China rather than relying solely on tariffs.
• Griffin defines the industry's cost of capital as the Risk-Free Rate + 4%. • Citadel has returned $25–$30 billion to investors recently to avoid "diluting alpha" (excess returns) by being overcapitalized. • The rise of Alternative Data (e.g., credit card spending tracking) has made short-term earnings predictions more transparent and competitive.
• Shift in Stock Picking: The market will increasingly reward "visionary" investors who can predict transformative products over years, rather than those simply predicting quarterly earnings beats. • Alignment of Interest: Investors should look for "Performance-based" firms where the General Partners (GPs) are the largest investors in their own funds. • Risk Management: Griffin advises managing portfolios for "tolerable losses." You cannot hedge every tail event, but you must ensure the "worst-case scenario" allows you to stay in business.
• China's Oil Demand: China has shown a surprising "elasticity of demand," significantly curtailing crude oil needs, which has kept global oil prices lower than expected despite Middle East conflicts. • Compute as a Commodity: Compute is now a major operational expense, with leading firms spending hundreds of millions of dollars on it. High-margin businesses will survive this "inflation in compute," while low-margin businesses may struggle.
• Energy Shield: The U.S. is currently shielded from the Middle East energy crisis due to domestic production and China's reduced demand. • Compute Costs: Investors should monitor the "compute spend" of technology-heavy firms, as it has become a necessary utility similar to jet fuel for airlines.

By @goldmansachs
The sharpest insights, timely analysis, and out-of-consensus thinking from Goldman Sachs. Founded in 1869, our firm delivers world-class execution for clients across our leading Global Banking & Markets and Asset & Wealth Management businesses. We aspire to be the world’s most exceptional financial institution, united by our shared values of partnership, client service, integrity, and excellence.