Ketu Desai
By Ketu Desai
The AI boom is accelerating as a major source of growth for the economy. AI CapEx is already 6% of the economy but has contributed more to GDP growth this year. Data center construction will soon surpass office construction. In fact, cap tech spending now makes up over half of the total spending in the economy.
Cloud CapEx growth is running at 56%, after 65% growth last year. The major hyperscalers (Meta, Microsoft, and Alphabet) all increased their CapEx spend guidance during Q2 earnings to $100bn, $120bn, and $85bn, respectively. In aggregate, it is expected that $7 trillion will be spent on AI infrastructure over the next decade. Such numbers are so high it doesn’t seem real.
This spending has benefited companies throughout the supply chain. Taiwan Semiconductor, which sits at one of the earliest stages of the AI supply chain, reported an earnings growth of 60%. Nvidia, a four trillion-dollar company, grew 56% in sales with 72% margins. Vertiv, which makes cooling systems for data centers, reported 35% growth and guided to 23-25% growth.
The power necessary for these data centers is enormous. Proposed 2027 AI server rack designs require 50x the power of racks used for the internet today, according to Goldman Sachs. US data center power demand is expected to double by 2028 to 95 GW, per S&P Global. By 2030 demand is estimated at 117 GW, up 147% from 2024. Meta alone expects to have 1.3 million GPUs online for AI workloads by the end of the year. It is working on two major data center projects, including the 5 GW Hyperion data center in Louisiana, which is enough to power roughly five San Franciscos. Bain estimates that the demand for energy will outpace the supply by the end of the decade. This dynamic is starting to be reflected in prices. The latest auction of the largest US grid operator (PJM) saw clearing prices rise 833%.
While the AI economy is booming, the rest of the economy is falling. The labor market is clearly weakening as private payrolls only grew by 3k in June and 8k in July. Goldman estimates that job gains will slow to 30k per month in the fourth quarter. This is well below the 80k necessary to maintain full employment. Continuing jobless claims are at cycle highs reaching nearly 2 million.
It is a “no hire, no fire” environment as part of the labor market weakness is due to companies using AI to optimize their labor force. For instance, Broadcom now uses approximately 40 IT workers to support 50 thousand employees. Or Shopify in 2022 had 11,600 employees, it now has 8,100 employees while revenue is up 91%. The other part of the labor weakness is natural economic forces, led by housing. Housing is in a recession with prices falling in 39 of the top 50 metros in July. Residential building construction has also diminished for four months in a row along with the Case-Shiller home price index. US home foreclosures have jumped 11% since June. This underlying weakness will likely prompt the Fed to cut two or three times this year, starting in September.
As far as markets are concerned, you don’t want to fade the AI stocks. This is not the debt fueled tech bust in 2000. These companies are funding CapEx from cash flow and not trading at 100x. In fact, many of the stocks, including Nvidia, Meta, Google, Taiwan Semiconductor, Vertiv, Marvell, and Vistra trade below a 1.0 PEG ratio, meaning that they are growing earnings faster than their multiple. None of the large cap tech AI names are even in the top 50 in terms of highest multiple in the S&P. They have gotten cheaper this year, because returns have been driven by earnings. They grew 26% in the second quarter compared to just 7% for the rest of the market.
This earnings outperformance will continue for at least the next year. Positioning is not stretched either as Morgan Stanley reports that the large tech names are the most under-owned in 16 years. We are just at the beginning of the AI boom. It will take years to build out data centers and related infrastructure. AI agents, physical AI, AI drug discovery, autonomous vehicles, and robotics are in their infancy. Just as the economy is bifurcated between AI and the rest, so is the market, stick with the winners.
Looking forward, the market will focus on the latest out of Washington, a Fed meeting, and the latest economic data.
Ketu Desai is the Principal of i-squared Wealth Management Inc. (www.isquaredwealth.com), an investment management firm based in New Jersey. ketu@isquaredwealth.com
