Nvidia is nearing a record high. This is what’s driving it.
Jensen Huang, CEO of NVIDIA, arrives for the Inaugural AI Insight Forum in Russell Building on Capitol Hill, on Wednesday, September 13, 2023.
A steady flow of news has added up to a banner month for Nvidia, as investors seem to have let go of some of their trepidation around generative AI — at least for now.
By Wednesday’s market close, Nvidia’s stock was up nearly 25% since September, bringing it closer to once again being the most valuable company in the world. The climb contrasts with late summer when even solid earnings released by the company in August could not move investors. So what happened?
Since the start of October, investors and Nvidia watchers have gained a few new numbers to add to their models.
$6.6 Billion to OpenAI
Despite the recent high rate of high-ranking staff turnover at OpenAI, the company finalized a long-anticipated $6.6 billion funding round, which valued the company at $157 billion.
The funding is the clearest sign that investors are still willing to back the biggest bet on generative AI available. Most of OpenAI’s funds funnel down to Nvidia via the companies that operate the immense data center capacity OpenAI uses and will continue to use from Microsoft and others.
OpenAI announced the fresh funding last week, led by Thrive Capital with Microsoft, Nvidia, Softbank, Khosla Ventures, and others participating. Billionaire Vinod Khosla told B-17 that though the AI space is overhyped, OpenAI isn’t. Nvidia’s stock price climbed for six days before leveling out on Wednesday.
2 Blackwell deliveries
News that Nvidia’s next-generation Blackwell chips wouldn’t make it to market as planned due to manufacturing challenges shook markets back in August and cast doubt on the company’s ability to keep up its promised pace.
On his August 28 earnings call, Nvidia CEO Jensen Huang assured listeners several billion dollars in Blackwell revenue would ship in the fourth quarter, but the market didn’t respond.
Last week, Huang said on CNBC that the demand for Blackwell is “insane”. And on Tuesday, two key customers announced they had received their Blackwell systems in social media posts that were quickly shared by Nvidia as well. The Microsoft Azure team posted on multiple platforms that the company had the first cloud running Blackwell.
Additionally, the OpenAI team posted “Look what showed up at our doorstep. Thank you to Nvidia for delivering one of the first engineering builds of the DGX B200 to our office.”
$23.5 billion TSMC blowout
Though there’s a sea of component suppliers below Nvidia on the generative AI supply chain, the most important and coincidentally easiest one to watch is Taiwan Semiconductor Manufacturing Company.
TSMC reports earnings monthly instead of quarterly and therefore, as Nvidia’s largest supplier, offers more frequent confirmation that demand for graphics processing units remains robust. Though the company doesn’t officially announce its third-quarter earnings until next week, the monthly results show $23.5 billion in revenue for the third quarter, which is ahead of analyst expectations.
TSMC’s revenue gains bolster the view that we are still in the early days of AI computing demand.
1 underlying fear
Huang takes every opportunity to profess the money-making capabilities of the technology his GPUs enable. He told David Solomon last month that companies that rent GPU capacity make back $5 for every $1 they spend.
The market has shown however, that Huang isn’t the source investors need to hear it from. But, anecdotal evidence of companies finding value in generative AI may be starting to penetrate the investor class as well, according to a new survey from Morgan Stanley.
The bank conducted a survey of 400 companies across six industries in five countries and 90% of the companies adopting generative AI found that the technology met or exceeded their expectations.
The bank also predicted that spending on cloud capacity buildout in 2025 will be equal to the entire Apollo space program, nearing $275 billion.
Nvidia’s early October climb may be due to investors waking back up to the fact that Nvidia, though highly dependent on the growth of generative AI, is pretty close to the base of the supply chain.
More data centers mean more GPUs and more optimism about what companies can do with them only adds to Nvidia’s opportunity. Fears about Nvidia “are overdone,” Morgan Stanley analysts wrote.
