Nvidia resolves a design flaw in its Blackwell AI processors and restarts mass production for Q4 2024 delivery.
Data center investments by hyperscalers to advance in the artificial intelligence race remain a focal point on Wall Street. A recent study by consulting firm McKinsey examined the European market in this regard, finding that data center power consumption in Europe is set to nearly triple by 2030, necessitating a significant increase in electricity supply, particularly from low-carbon sources, along with upgrades to grid infrastructure. According to the McKinsey report, the total IT load demand from data centers in Europe is expected to grow to approximately 35 gigawatts by 2030, up from 10 gigawatts presently, while data center power consumption in the region is anticipated to triple to over 150 terawatt hours by the end of this decade.
The McKinsey analysis reveals that data centers are expected to represent about 5% of Europe's total power consumption in the next six years, compared to roughly 2% at present. Meeting this growing demand will require a minimum investment of around $300 billion in data center infrastructure, excluding power generation capacity. The report highlighted that satisfying the increased electricity demand represents a significant shift for Europe, where overall power demand has remained largely unchanged since 2007. Kevin Restivo of CBRE recently noted that demand for data center space in Europe is set to outstrip delivery of new stock for the third straight year and that AI demand will exacerbate the issue.
For this analysis, AI stocks were selected by combing through news articles, stock analysis, and press releases, with a focus on stocks popular among hedge funds. A quarterly investment strategy selecting 14 small-cap and large-cap stocks has returned 275% since May 2014, beating its benchmark by 150 percentage points.
NVIDIA Corporation (NASDAQ:NVDA) provides graphics, computing and networking solutions. On October 23, Jensen Huang, the CEO of the firm, stated that a design flaw with the latest Blackwell AI chips, which had impacted production, had been fixed with the help of longtime Taiwanese manufacturing partner TSMC. According to news agency Reuters, Huang made this statement at the launch of a supercomputer in Denmark. Blackwell chips were originally scheduled to go into mass production in the second quarter of 2024 but were delayed by a few months, impacting hyperscale NVIDIA customers like Meta and Google. Huang indicated that mass production of Blackwell would resume in the fourth quarter, bringing in billions of dollars in revenue for the chipmaker. He emphasized that the Blackwell design flaw was "100% the fault of his firm," clarifying that seven different types of chips were designed from scratch and had to be ramped into production simultaneously for Blackwell.
NVIDIA ranks 5th on the list of AI stocks to watch. While acknowledging NVDA's potential as an investment, the analysis suggests that some AI stocks hold greater promise for delivering higher returns within a shorter timeframe.