Is the Semiconductor 'AI Boom' Really Over? Even Samsung Drops... What Really Matters Is Not Earnings but 'This' [Click e-Industry]

Concerns Mount Over Slowing AI Investment by Big Tech
DB Financial Investment: "ROI Management Takes Priority Over CAPEX"
Cost Burden Shifts in the AI Supply Chain
Q2 Earnings: Working Capital Metrics Take Center Stage Over Sales Growth

Concerns are mounting that the artificial intelligence (AI) investment boom may be cooling. Meta has announced plans to sell its idle computing capacity to external parties, and there have also been reports of some data center projects being delayed or canceled. The market has interpreted these developments as signs of a slowdown in the AI investment cycle.

However, is AI investment truly declining? Analysts suggest that what big tech companies are actually trying to reduce is not the investment itself, but rather the burden of return on investment (ROI) relative to the scale of those investments.

Investment in AI infrastructure is directly linked to competitiveness. For hyperscalers such as Amazon, Microsoft, Google, and Meta, it is not easy to cut back on these investments. It is also difficult to determine right now whether there is overinvestment. Only after data centers and other facilities are built and real demand and returns are confirmed can such a judgment be made.

As a result, big tech companies are responding in other ways. They are extending the useful life of their equipment, selling non-core assets, and reducing working capital. In particular, improving working capital efficiency is less likely to be perceived as a negative signal on the surface. This is because it allows them to enhance capital efficiency without appearing to be cutting back on investment.

The problem is that this burden is being shifted to the tail end of the supply chain. When hyperscalers reduce inventory and extend payment terms, their own cash flow improves. On the other hand, companies that supply servers, components, and equipment must accumulate inventory in advance, manufacture products first, and wait longer to receive payment. Jin Kyung Lee, a researcher at DB Financial Investment, stated, "Reducing working capital is an effective way to improve capital efficiency without sending a negative signal, unlike extending the useful life of assets," and added, "AI CAPEX is unlikely to slow easily, but the costs and funding burdens will be passed on to companies with weaker bargaining power, and the investment cycle will continue."

DB Financial Investment cited Super Micro Computer as a representative case. As demand for AI servers increased, its sales grew, but inventory and accounts receivable also rose rapidly. Ultimately, the company announced a $7 billion financing plan. What worried the market was not simply the dilution from a new share issuance. Rather, it was the structure in which cash generation could not keep pace with sales growth.

In the end, when evaluating AI beneficiaries, it is now important to look not only at sales growth, but also at whether those sales are actually being converted into cash. It is critical to check whether inventory is increasing too quickly, whether accounts receivable are accumulating, and whether operating cash flow is keeping up with profits. Researcher Lee emphasized, "As the AI investment cycle matures, the market is likely to place a higher premium on cash generation capability and capital efficiency rather than on sales or profit growth," adding, "In the future AI supply chain, the key variable for stock price differentiation will be which companies can absorb the working capital burden and which can pass it on to others, rather than simply sales growth."

This content was produced with the assistance of AI translation services.