Prioritizing ROIC Over Speculative AI Growth Metrics

The current market is in a "show me" phase where initial excitement over AI spending is hitting the reality of return on invested capital (ROIC). While major indices look resilient, a divide is forming: high-quality hyperscalers are holding steady, while lower-quality tech stocks are being sold off. The main risk is not a lack of demand, but the possibility that circular financing among AI companies is hiding shrinking margins. Investors and operators should move up the capital structure and focus on companies that can prove actual earnings growth instead of speculative expansion. Shifting from "growth at any cost" to "proven efficiency" is the best way to navigate the next 12 to 18 months.

The Illusion of the "Earnings Bubble"

The market is questioning the sustainability of massive AI-related capital expenditures. Noah Weisberger of BCA Research notes that while there is widespread nervousness, the losses are mostly isolated to lower-quality tech stocks. The systemic risk is the circular nature of the AI trade, where companies like NVIDIA, AMD, and Broadcom are essentially financing their own customers' growth.

The market no longer rewards the scale of the build-out, but the efficiency of it. As Weisberger points out, we are in a "show me" phase. The turning point will not be the release of new models, but the reality of their earnings.

> "I think we were probably a little bit more confident a couple months ago, maybe a couple quarters ago... The path back on our return on equity is by growing revenues and expanding margins. We've seen some of that, but the market I think is questioning it."
>
> -- Noah Weisberger, Chief US Equity Strategist at BCA Research

The Hidden Cost of "Non-Urgent" Policy

In the macro sphere, the Federal Reserve’s communication strategy has created a gap between data and policy. Neil Dutta of Renaissance Macro argues that the Fed’s reluctance to act is a failure of risk management. By skipping rate hikes despite firm inflation data, the Fed is betting that financial conditions will tighten on their own.

The system is "routing around" the Fed’s lack of urgency. Because credit spreads have widened, the market is doing the heavy lifting. However, Dutta warns that this is a dangerous game: waiting for a good excuse to hike often forces the central bank to move more aggressively later.

> "It may not be urgent but it's prudent, right? I mean, it makes sense to go because it's about the balance of risk to the inflation outlook."
>
> -- Neil Dutta, Head of US Economic Research at Renaissance Macro

Where Immediate Pain Creates Lasting Moats

The political landscape, as analyzed by Jessica Taylor of The Cook Political Report, mirrors this "show me" market dynamic. Candidates who rely on party figureheads rather than local performance are increasingly vulnerable. The shift of races from "toss-up" to "lean Democrat" in places like Ohio shows a change in voter incentives: when the brand is damaged, the candidate must provide tangible, localized value to survive. This mirrors the corporate world, where hyperscalers with massive balance sheets are reliable, while speculative tech projects are being discarded as the cost of capital rises.

Key Action Items

• Move up the capital structure: Prioritize hyperscaler debt or high-quality equity over speculative, lower-rated tech debt. This provides a buffer if the current spending cycle contracts (Immediate).
• Monitor "Earnings Breadth" over "Token Growth": When companies like Anthropic report, look past the headline revenue numbers. Focus on who is buying and whether token price deflation is being offset by volume (Next 3 to 6 months).
• Prepare for a "Higher-for-Longer" Policy Reality: Given that inflation data remains firm, assume that rate hikes are likely in December. Position portfolios for a scenario where the Fed stops skipping and begins acting (Next 1 to 3 months).
• Shift from "Growth at Any Cost" to ROIC Metrics: Re-evaluate holdings based on Return on Invested Capital. The market is currently punishing companies that cannot prove their spending is generating actual profit (Ongoing).
• Ignore the "Circular Financing" Noise: Focus on the underlying demand from end-users rather than the inter-company financing deals among AI hardware providers. The true test of the AI trade is external adoption, not internal reinvestment (12 to 18 months).