The optimism of CEOs of Wall Street’s leading artificial intelligence (AI) companies doesn’t always match their actions — and that could be a problem.
About thirty years ago, the Internet forever changed the growth trajectory of corporate America. The ability of companies to reach beyond their storefronts with the click of a mouse has fueled the growth of American businesses for decades.
However, Wall Street has been waiting for quite some time for the next big trend to rival the Internet. While a number of next-big-thing technologies, innovations, and trends have emerged since the mid-1990s, none have come close to what the advent of the Internet has done for businesses… until now.
Image source: Getty Images.
The rise of artificial intelligence (AI) has professional and everyday investors excited – and for good reason. The ability of AI-powered software and systems to become more proficient at their assigned tasks and possibly even learn new skills without human intervention gives this technology utility in most sectors and industries around the world.
Perhaps the most stunning estimate of AI’s usefulness comes from ‘Sizing the Prize’, a report released by researchers at PwC. Based on a combination of increased productivity and knock-on effects on consumption, PwC predicts a $15.7 trillion increase in the global economy by 2030, solely due to the artificial intelligence revolution.
But while most analysts and investors on Wall Street are overwhelmingly optimistic about AI, conviction seems to be lacking where it matters most: among the companies leading this game-changing revolution.
The sky is seemingly the limit for Nvidia, Broadcom and AMD
A quick look at the stock charts of Wall Street’s leading AI companies shows that they’re certainly not struggling. Nvidia (NVDA 0.78%) has gained more than $3 trillion in market cap since the start of 2023 Broadcom (AVGO -0.90%) And Advanced micro devices (AMD -0.18%) have achieved a gain of 216% and 141% respectively in 21 months and implemented change. AI undeniably ensures this outperformance.
Nvidia’s hardware has taken the lead. The graphics processing units (GPUs) are essentially the brains behind decision-making in AI-accelerated data centers. Demand for the H100 GPU (commonly known as the “Hopper”) has been so overwhelming that Nvidia has been able to command a 100% to 300% price premium for its chips, compared to its rivals. The end result is a double-digit improvement in the company’s adjusted gross margin.
Although Advanced Micro Devices entered the arena long after Nvidia, the cheaper MI300X AI GPUs are still in high demand. With Nvidia’s chips lagging behind, proven chip companies like AMD can be expected to take AI GPU orders away from the former.
Just over a week ago, AMD unveiled its next-generation AI GPU, the Instinct MI325X, which is poised to go head-to-head with Nvidia’s successor Blackwell GPU architecture. AMD plans to start production of its newly unveiled chip later this year, which wouldn’t give Blackwell much, if any, head start.
Meanwhile, Broadcom is the preferred choice of businesses for its AI networking solutions. While Broadcom is known for much more than just its AI bands, including the wireless chips and accessories used in next-generation smartphones, the company’s Jericho3 AI structure has played a key role in reducing tail latency and maximizing the computing potential of AI GPUs in high-computing data centers. Most of Broadcom’s current revenue growth can be traced to its AI solutions.
Seemingly there’s nothing to indicate that the arrow isn’t pointing higher for all three companies… unless you dig into their respective insider transactions.
Image source: Getty Images.
Wall Street’s leading artificial intelligence stocks have a persuasion problem
Based on comments from Nvidia CEO Jensen Huang, Broadcom CEO Hock Tan, and AMD CEO Lisu Su, demand for AI solutions is robust, as their companies’ respective double-digit growth rates would imply. Yet the actions of these CEOs and other insiders point to a decisive persuasion problem among Wall Street’s leading AI firms.
Whenever an insider of a publicly traded company (this could be a board member or a high-ranking member of the executive team) buys or sells shares of their company’s stock, they are required to report these transactions to the Securities and Exchange Commission on Form 4 . Exchange committee. Over the subsequent twelve-month period (October 17, 2023 through October 16, 2024), sales activity was lopsided, to say the least, at all three of these leading companies:
- Nvidia: 83 insider sales (25 from Jensen Huang) and 0 insider purchases
- Broadcom: 32 insider sales (eight from Hock Tan) and 0 insider purchases
- AMD: 23 insider sales (eight from Lisu Su) and 0 insider purchases
Collectively, this is 138 separate insider sales and not one open market purchase from a single insider at any of these artificial intelligence giants over the past year.
To be fair, there is a laundry list of reasons for insiders to sell shares, and not all of them are necessarily bad news. For example, insiders can sell shares to cover their tax bills. It is not uncommon for executives who receive a lot of stock-based compensation to have to sell some of their stock to cover federal and/or state taxes.
On the other hand, there is only one reason why company insiders buy shares: they believe they will rise. It’s been thirteen months since the last insider purchase at Broadcom, 46 months since the last insider purchase at Nvidia, and over a year. half decade since the last insider purchase at AMD! Simply put: actions speak louder than words.
If insiders at Nvidia, Broadcom and AMD lack the conviction to get their money’s worth, why should ordinary investors pay a premium for a technology that has yet to prove itself?
