Apollo's Slok Warns AI Spending Isn't Lifting Profits for Most Companies

Apollo Global Management chief economist Torsten Slok has issued a stark warning about the disconnect between AI investment and profitability across corporate America. Profit margins are rising for the Magnificent 7 technology giants but remain flat for the other 493 companies in the S&P 500, according to Slok's analysis. This divergence isn't just a statistical curiosity; it represents a fundamental challenge to the valuation premiums investors have assigned to technology stocks.

The risk, Slok argues, is that if AI spending doesn't translate into broader productivity gains and profit growth, the market will face a painful repricing. Big Tech valuations assume that AI benefits will eventually spread throughout the economy. When that spread stalls, the foundation for those valuations weakens considerably. Bloomberg reports that Slok made these comments during a Tuesday appearance, emphasizing that marginal companies need to demonstrate they can convert their AI investments into actual returns.

Slok pointed to token optimization as a signal of the bumpier road ahead for AI adoption. This technical detail matters because it suggests companies are already searching for ways to reduce their AI costs rather than expanding their use of the technology. When businesses prioritize efficiency over expansion, it typically indicates that the initial excitement has given way to harder questions about return on investment.

The token optimization trend reflects a broader pattern in enterprise technology cycles. Companies invest heavily in new capabilities, then consolidate and optimize once the low-hanging fruit disappears. Yahoo Finance notes that this optimization phase often precedes a more selective approach to technology spending, where only applications with clear, measurable benefits receive continued funding. For AI specifically, this could mean slower growth in infrastructure spending than current projections assume.

The S&P 500's profit growth has become dangerously concentrated in a handful of technology companies, creating systemic vulnerabilities that extend beyond individual stock performance. When earnings growth depends on so few firms, any disruption to their trajectory, whether from regulatory pressure, competition, or simple saturation, ripples through entire market indices and the passive funds that track them.

Edward Conard's analysis of Mag 7 earnings convergence with the broader S&P 493 underscores this fragility. The narrowing gap between technology leaders and laggards isn't happening because the broader market is catching up; it's happening because the growth premium for tech is compressing as investors question how long the outperformance can continue. Seeking Alpha highlights that this concentration has made the index itself a bet on a single sector's continued dominance, rather than a diversified play on the American economy.

A repricing of AI valuations wouldn't necessarily mean a catastrophic crash, but it would represent a significant shift in how markets value technology investments. Slok's warning of a painful repricing suggests that current valuations embed assumptions about AI adoption that simply aren't materializing in corporate financial statements outside the largest technology firms.

The mechanism for such a repricing typically involves multiple compression rather than earnings collapse. Investors gradually demand lower price-to-earnings ratios as growth expectations moderate, particularly for companies whose AI narratives have outpaced their actual business performance. Bloomberg's coverage of Slok's comments indicates that this process may already be beginning, as the market digests the reality that AI's benefits are accruing narrowly rather than broadly across the economy.

The failure of AI spending to lift margins across the S&P 493 points to deeper challenges in technology diffusion. Historically, major technological innovations have taken years to reshape productivity across sectors, with early benefits concentrated in the companies that create and deploy the technology first. AI appears to be following this pattern, but at a scale that raises questions about whether the eventual diffusion will justify the current investment levels.

For non-technology companies, the barriers to productive AI adoption include data quality issues, organizational resistance, skills gaps, and the difficulty of integrating new systems with legacy infrastructure. News.futunn's coverage of Slok's remarks emphasizes that these practical constraints mean AI's economic impact will likely be more gradual and uneven than equity markets have priced in. The implication for investors is that patience may be necessary, but patience itself becomes costly when capital is deployed at today's valuations.

Companies outside the Magnificent 7 face a strategic inflection point. They can either demonstrate that their AI investments are generating measurable returns, or they risk investor pressure to scale back spending and return capital through dividends and buybacks instead. This dynamic creates tension between short-term financial performance and long-term competitive positioning, particularly for companies that fear being left behind if AI does eventually become transformative.

Slok's intervention adds intellectual weight to the skeptical case on AI valuations, coming as it does from a prominent figure at a major alternative asset manager. Apollo itself manages hundreds of billions in assets, giving Slok's views particular resonance with institutional investors who must decide whether to maintain or reduce their technology exposures. The coming quarters will reveal whether his warning prompts a broader reassessment, or whether the Magnificent 7 can once again prove the skeptics wrong by extending their profitability advantages.