Chris Wood’s Warning: Why Malinvestment Could End the AI Boom

The artificial intelligence era has sparked the most dramatic capital expenditure (capex) cycle in history, but a significant structural risk looms on the horizon. Jefferies’ Global Head of Equity Strategy, Chris Wood, warns that the AI trade may not end due to a chip shortage or supply glut, but rather a realization that the massive investments are failing to generate adequate returns.

The Specter of Malinvestment

Unlike traditional semiconductor cycles that end when inventory gluts hit the market, Wood argues that the current AI boom faces a unique threat: malinvestment. He suggests the "endgame" for the AI trade will be triggered when hyperscalers and leading AI labs fail to earn a satisfactory return on the astronomical capex they are currently undertaking.

A particular concern is the existence of circular funding loops. Wood points to scenarios where major players, such as Nvidia, provide financing to companies like OpenAI, which then use those funds to purchase more Nvidia chips. While this creates a powerful short-term growth loop, it relies heavily on optimistic monetization assumptions that could unwind if investors lose confidence in long-term earnings visibility.

Massive Capex and the Taiwan Surge

The scale of investment currently seen in the sector is unprecedented. Wood describes the ongoing build-out as the most dramatic capex cycle he has ever witnessed. A prime example is TSMC, which has significantly increased its guidance, with projections for 2027 reaching as high as US$65–70 billion.

This concentrated spending is driving massive macroeconomic shifts, particularly in Taiwan. The region saw real GDP growth of 14.55% year-on-year in Q1 2026, fueled by an export order surge of 53.4%. Currently, AI-related demand is estimated to account for approximately 31% of TSMC’s total revenues for 2026, illustrating how deeply the global economy is becoming tethered to AI infrastructure.

Commoditization and the Rise of Cheap Models

Another layer of risk is the rapid commoditization of Large Language Models (LLMs). As the cost per token drops, the competitive advantage of premium Western providers is being challenged. Wood notes that new models, such as Hong Kong-listed Z.ai’s GLM-5.2, are reportedly approaching the performance of top-tier models like Anthropic but at just one-quarter of the cost.

Data from OpenRouter highlights this shift: in late June, top Chinese AI models processed 21.37 trillion tokens, a massive jump from 4.37 trillion in April. This volume significantly outpaced the 5.76 trillion tokens processed by leading US models, signaling a highly competitive and increasingly commoditized landscape.

Shifting Strategy: Focus on Memory and Hardware

Despite these warnings, Wood is not predicting an immediate collapse. Instead, he is repositioning portfolios toward the "picks and shovels" of the industry—specifically DRAM and memory suppliers. Due to the Jevons Paradox, as compute becomes more efficient and cheaper, total consumption actually increases, benefiting hardware providers.

Major memory makers like Micron are already locking in long-term stability, with Micron signing strategic agreements covering 20% of its DRAM volume. Consequently, Wood is increasing exposure to tech hardware names like SK Hynix, Kioxia, and Samsung Electronics, betting that they will remain beneficiaries even if the broader AI software and service layers struggle with capital discipline.

Key Takeaways

• The Primary Risk: The AI trade is more likely to end due to "malinvestment" and a lack of ROI for hyperscalers rather than a traditional semiconductor supply glut.
• The Commoditization Threat: Rapidly advancing and cheaper Chinese AI models are putting intense pricing pressure on premium Western AI providers.
• Strategic Pivot: Investment interest is shifting toward memory and hardware (DRAM) players who hold significant pricing power and long-term customer contracts.