Nvidia’s highly anticipated “Blackwell” B-200 artificial intelligence chip will reportedly be delayed, sending the near-term future of the entire AI industry into a state of uncertainty.
Tech news outlet The Information claims that a Microsoft employee and at least two other people familiar with the situation have stated that the new chip’s launch date has been pushed back by at least three months due to a design flaw.
While Nvidia hadn’t given a public launch date, CEO Jensen Huang recently announced that the company would begin sending engineering samples “this week” on July 31 at the SIGGRAPH event in Denver, Colorado.
It’s unclear at this time if the reported delay will affect the B-100 series as well as the B-200. While both share similar architecture, the latter is purported to have better performance.
Artificial intelligence chips
Analysts expect Nvidia to sell hundreds of billions of dollars' worth of B-series AI chips to the likes of Amazon, Google, Meta, Microsoft, and dozens of other AI firms. The company’s revenues for 2025 will likely hinge on its ability to produce enough chips to meet the demand. Arguably, the entire AI sector’s fortunes will as well.
Nvidia’s market dominance is startling. With a $2.6 trillion market capitalization, its next closest competitor in the semiconductor industry is TSMC, with a cap of $777 billion.
It may be a stretch to call TSMC a competitor, however, as the Taiwanese company fabricates the bulk of Nvidia’s chips through a strategic partnership.
In the US, Nvidia has next to no competition. A number of tech firms, including Microsoft and Google, are working to develop their own AI chips, but by and large the industry is currently dominated by Nvidia.
Rival chipmakers such as Intel and AMD have so far failed to find a foothold in the generative AI industry. Both are currently in the midst of a pivot to servicing the AI industry.
Related: Tech stocks plummet as Intel nosedives, Nvidia faces antitrust probe
Bubble or balloon?
The second quarter of 2024 wasn’t kind to big tech. Every technology company in the global top 20 by market capitalization is currently down month-over-month (as of Aug. 3, 2024) with the exceptions of Tesla and Apple.
Nvidia’s no exception, despite reaching record stock highs in 2024, it’s trending with the bulk of other tech companies. In the wake of Nvidia’s explosive growth since the launch of OpenAI’s ChatGPT (which was trained on Nvidia chips), investment analysts appear to be getting impatient waiting for the next big thing.
The Financial Times recently reported that hedge fund firm Elliott Management sent a letter to investors warning that Nvidia was “in bubble territory.” The letter also stated that current generative AI use-cases were “never going to be cost-efficient, are never going to actually work right, will take up too much energy, or will prove to be untrustworthy.”
Venture capital firm Sequoia Capital has also issued a recent negative forecast for the AI sector, stating that AI firms needed to make up about $600 billion in revenues to offset expenditures on Nvidia GPUs alone.