The AI race among tech giants is heating up, but is it a sprint or a marathon? As we delve into the upcoming earnings season, one key metric will be under the spotlight: capital expenditure (capex) on AI data centers. Tech behemoths like Google, Amazon, Microsoft, and Meta have collectively committed to a staggering $700 billion in spending this year alone. But the question remains: is this an arms race or a costly standstill?
The Costly AI Arms Race
The race to build AI capacity is akin to a high-stakes game of chess. Each move, in the form of increased spending, is carefully calculated. However, the rising costs of memory chips, power equipment, and skilled labor are creating a challenging environment. Morgan Stanley's estimates reveal a 20% increase in the cost of building AI capacity for leading systems, with some setups now costing billions more per gigawatt.
This spiral of increased spending leads to further demand, which in turn drives up prices. Brad Gastwirth, head of research at Circular Technology, estimates that a significant portion of the next capex increase will be due to inflation, rather than genuine expansion. This distinction is crucial for investors, as it highlights the potential for a misleading perception of progress.
Inflation vs. Innovation
Inflationary pressures have a significant impact on the AI race. Research suggests that soaring memory prices account for almost half of the growth in capex this year. This raises the question: are these tech giants truly expanding their AI capabilities, or are they simply paying more to maintain their position?
As we look ahead to the 2027 earnings season, estimates suggest even higher capex plans. Google, Amazon, and Meta are expected to increase their spending significantly. However, without a focus on the details, such as power capacity, GPU deployments, and memory purchases, this increased spending may not translate into meaningful progress.
The Cautious Approach
In a highly competitive environment, no company wants to appear cautious. The fear of falling behind in the AI race is a powerful motivator. However, as Gastwirth points out, investors should pay close attention to the narrative accompanying any spending increase. A genuine expansion should be accompanied by details on the aforementioned metrics, indicating a strategic and thoughtful approach.
Without these details, the risk is that Big Tech is simply paying more to stand still. This raises the question: at what point does the cost of standing still become unsustainable?
A Broader Perspective
The AI race is not just a battle between tech giants; it's a reflection of the broader economic landscape. The impact of inflation, supply chain issues, and the ever-increasing demand for computing power are all factors that extend beyond the tech industry. As we navigate this complex landscape, it's crucial to consider the long-term implications and the potential for a shift in the balance of power.
In my opinion, the AI race is a fascinating and complex interplay of innovation, economics, and strategy. It's a reminder that, in the world of technology, standing still is not an option, but the cost of progress must be carefully considered.
