Goldman Sachs' Bordlemay on AI Investment and Inflation Caused by Market Uncertainty

Goldman Sachs Asset Management highlighted equities at all-time highs as supported by upward earnings revisions and durable AI investment spending projected to reach $1 trillion over the next 3-4 years. The commentary is constructive for risk assets, with AI capex seen as a sustained tailwind for earnings and market leadership. Impact is primarily sentiment-driven rather than a direct catalyst for individual stocks.

Analysis

The market is treating AI capex as a demand shock, but the more important second-order effect is margin dispersion. Names with direct AI infrastructure exposure should continue to outperform, yet the real beneficiaries are the picks-and-shovels suppliers with pricing power, long backlogs, and limited customer concentration risk; those companies can convert spending intensity into visible earnings revisions long before end-demand is fully proven. By contrast, software and application-layer vendors without differentiated data access or workflow lock-in may underperform if buyers keep re-allocating budgets toward infrastructure rather than seats.

The biggest risk is that the current earnings revision cycle is being extrapolated too far into 2H and next year. If AI spending remains durable, it still creates a capital intensity problem for hyperscalers and large incumbents: free cash flow can lag revenue growth for several quarters, and any small slowdown in cloud growth or enterprise procurement could trigger multiple compression even while the AI narrative stays intact. That makes this a very asymmetric setup where high-quality enablers can work for months, but the trade can reverse quickly if capex guides flatten or depreciation catches up.

The contrarian view is that the market may be underestimating how much of the AI buildout is already crowded into valuation, while underpricing the breadth of beneficiaries outside the obvious megacaps. If the spend really reaches the stated scale over 3-4 years, there should be a second wave in power, networking, semicap equipment, and data-center infrastructure rather than just the largest platform names. The alpha is likely in relative winners that have not fully re-rated yet, not in chasing the most obvious AI proxies after a strong run.

Near term, the key catalyst is next earnings season: any upward revision to AI-related capex guides should support the group, but even a modest pause would hit the most crowded names first. Investors should also watch for signs of supply bottlenecks — those usually precede another leg higher in pricing power for equipment and infrastructure vendors.

Key Decisions for Investors

• Long a basket of AI infrastructure enablers versus broad tech: buy semicap equipment / networking leaders on any 3-5% pullback and hold for 3-6 months; the risk/reward is favorable because earnings revision momentum can compound faster than multiple expansion.
• Pair trade: long AI data-center power/grid beneficiaries and short a basket of over-owned software names with weak consumption metrics; target a 2-3 quarter horizon where capex flows matter more than narrative.
• Add call spreads on the most liquid hyperscaler/capex proxy names ahead of earnings; structure for a 6-9 month window so upside from revised spending guides outweighs theta decay.
• Fade crowded AI leadership after sharp post-earnings gaps: trim if forward capex commentary stops accelerating, because the trade can unwind in days if the market senses peak spend timing.
• Use pullbacks to accumulate quality infrastructure names rather than chasing megacap AI leaders; the cleaner asymmetry is in suppliers where backlog visibility can re-rate fundamentals over the next 12 months.