At the "Dialogue with Omaha: The 11th U.S.-China Investors Reception" held on May 2 in Omaha, Nebraska, Sana Bao, Director of Alternative Investments, International Client Advisor, and Financial Advisor at Morgan Stanley, analyzed the drivers behind the U.S. stock market's high levels. She stated, "Regarding AI and AI infrastructure, we believe investment in this capital cycle is one of the key drivers."
Bao pointed out that capital investment in AI may far exceed expectations. For example, in 2022, when generative AI large language models first emerged, total capital expenditure by top global cloud service providers was approximately $100 billion. "By 2026, this figure is expected to reach $600 billion among leading companies," she noted, attributing this massive investment to the explosive growth in AI computing power.
She further elaborated on AI's rapid advancement: "By mid-2025, the most advanced AI models could only complete 24% of expert human tasks. Five months ago, that figure reached 71%. As of last month, it has surged to 83%." Bao described AI development as "non-linear and accelerating."
However, she highlighted a significant bottleneck emerging alongside this rapid growth: "With such tremendous growth, bottlenecks are inevitable. Each leap requires massive computing power, which is supported by energy and electricity. The bottleneck is shifting—it's no longer just about AI chips but increasingly about energy and computing resources, especially in the U.S." Bao added, "Of the projected $600 billion investment by top cloud providers, 60% is expected to go toward AI infrastructure, including data centers and power supply, rather than chips alone."
Regarding AI infrastructure's energy demands, Bao shared striking data: "In 2024, AI computing power, such as data center electricity consumption, accounted for 6% of total U.S. electricity usage. By 2030, just four years from now, we project this will rise to 18%—nearly one-fifth of the nation's total electricity demand will come from AI-related infrastructure." She candidly remarked, "Is the U.S. power grid ready for this? Far from it."
Based on this outlook, Bao emphasized investment opportunities in supporting infrastructure: "Addressing this bottleneck may not fall to AI software companies alone, but to less visible yet essential sectors enabling AI expansion—such as nuclear energy, natural gas, energy storage, grid equipment, and data center developers." She revealed, "While market consensus expects 5.6% growth for such assets by 2027, we project 22%. If this holds, these assets are significantly undervalued by the market."
As a fellow alumna of Warren Buffett, Bao shared her perspective on value investing: "Gathering in Omaha today, we discuss AI, energy, and infrastructure, but what we're truly exploring is value investing." She reflected, "Buffett’s lifelong practice has been seeking certainty amid uncertainty. In investing, what matters isn’t chasing trends but understanding the structural, irreplaceable demands behind them."
To illustrate, Bao drew a parallel to the Gold Rush a century ago: "Few struck gold—it was highly uncertain. But those selling shovels, jeans, or building railroads achieved far more predictable returns." She also cited Buffett’s investments in Coca-Cola and railways as examples of valuing assets with enduring pricing power and cash flow.
When asked about AI bubbles or stock recommendations, Bao responded frankly: "We don’t know who will win in the AI race—it’s highly uncertain. But the long-term demand for energy, computing power, and data centers is becoming increasingly clear."
In closing, Bao summarized her key insight: "Amid uncertainty, identifying long-term variables with sustained value is the hallmark of sound investing."
Comments