Shares of semiconductor and cloud computing companies tumbled sharply on Thursday following reports that Meta Platforms is planning to launch a cloud business aimed at selling access to artificial intelligence (AI) computing power and models. The move would position Meta as a direct competitor to established cloud service providers such as Amazon Web Services, Microsoft Azure, and Google Cloud.
The news triggered a broad sell-off across Asian markets, with South Korean memory manufacturers Samsung Electronics and SK Hynix declining by 9 percent and 14.6 percent respectively. Similar declines were seen among U.S. chipmakers, including Micron Technology, AMD, Intel, and SanDisk, which experienced losses ranging from 4 to 14 percent. Cloud companies CoreWeave and Nebius also faced significant drops of 14 percent and 17 percent.
Investors appeared concerned that Meta's entry into the cloud sector might lead to an oversupply of AI computing resources, particularly graphics processing units (GPUs), thereby creating an industry-wide hardware glut. This perception contributed to heightened market volatility and steep declines in semiconductor stocks.
However, analysts at research firm SemiAnalysis characterized the market reaction as unwarranted. In a report released shortly after the sell-off, the firm argued that the fears of excess capacity were "erroneous." According to SemiAnalysis, Meta’s data center expansion and compute infrastructure investments are expected to accelerate rather than slow, with capital expenditures projected to be "shockingly high" in 2027.
The analysis highlighted that since early 2024, Meta has signed deals amounting to nearly 10 gigawatts of capacity, with most new additions facilitated through third-party providers. SemiAnalysis suggested this model is likely to continue, positioning Meta as a significant driver of future growth for cloud service vendors such as CoreWeave and Nebius rather than a threat.
Overall, the research firm concluded that the aggressive market sell-off and renewed concerns over AI hardware overcapacity did not reflect the underlying dynamics of Meta's cloud strategy, which is expected to expand demand for AI compute rather than suppress it.
