Artificial intelligence stocks have taken the investing world by storm, with leading technology and semiconductor stocks having surged in 2023 after plunging in 2022.
Yet some are now comparing the AI boom to the dot-com bubble, warning that these stocks are due to fall after the recent hype is over.
While some AI stocks are no doubt now at higher and perhaps extended valuations, there is a fairly large difference between today's AI boom and the dot-com mania of the late 1990s. Unlike, say, an unprofitable or even pre-revenue start-up going public just because it has a "dot-com" in its name, many of today's leading AI companies are large, established, and profitable.
Moreover, the possibilities of AI should lead to outsized growth in the years ahead, even for large-cap companies like these two leaders.
Image source: Getty Images.
Microsoft
Software and cloud giant Microsoft (MSFT -2.45%) kicked off the latest AI surge when its investee OpenAI released ChatGPT last November. The advent of ChatGPT awakened the tech world, and really every business, to the transformative potential of artificial intelligence.
Following the earth-shaking debut of ChatGPT, Microsoft followed up its seed investment in OpenAI with another $10 billion in January, bring its total investment to $13 billion, or just under 50% of OpenAI's $29 billion valuation at the time.
Microsoft is now infusing all of its business infrastructure, platform, and software products with OpenAI's technology. Not only that, but Microsoft is also using its newfound AI capabilities to enhance its other products where it doesn't have an incumbent lead, such as Bing Search.
In February, Microsoft held an event outlining out how OpenAI and ChatGPT are helping improve Bing's functionality and accuracy, along with the Microsoft Edge web brower. This is important because the digital advertising industry is large enough and Microsoft's advertising market share is small enough that some improvement in this area could be impactful for Microsoft, despite the company already being so large. At the event, Microsoft CFO Amy Hood noted that for every one point of Search market share taken, that's an extra $2 billion in high-margin revenue. While Microsoft has made a little over $200 billion in revenue over the past 12 months, given that Bing has just 2.8% global search market share, the potential to take even a few points of share could be meaningful.
Even if that doesn't happen, Microsoft's investment in OpenAI seems likely to boost its competitive position in the cloud infrastructure wars and other enterprise productivity offerings where it's already an incumbent, such as the Dynamics enterprise resource planning suite, GitHub developer platform, the Power automation suite, and other cutting-edge tools for businesses.
Nvidia
Perhaps even more than Microsoft, Nvidia (NVDA 3.71%) has become the poster child for the AI revolution – and for good reason. In last week's earnings release, Nvidia's management blew analysts away with their July quarter guidance, forecasting $11 billion in revenue, up from $7.2 billion in the April quarter and about 55% higher than the analyst consensus.
It appears a wide swathe of businesses have decided to invest heavily in their AI capabilities, and Nvidia is today's dominant AI "arms" dealer, as its general-purpose graphics processing units (GPUs) form the backbone of today's accelerated computing workloads. While traditional data centers have historically been packed with central processing units (CPUs), GPUs' parallel-processing capabilities are needed to handle the massive computing workloads for artificial intelligence training and inference.
On the recent conference call with analysts, CEO Jensen Huang noted two important points driving the high future expectations for Nvidia. First, that there is about $1 trillion worth of installed global data centers, with most of that capacity built primarily with CPU architectures. Over time, Huang posits that this infrastructure will be mostly replaced by accelerated servers powered by GPUs, as generative AI and accelerated computing go from being the shiny new toys of today to the standard of tomorrow.
There appears to be some truth to this, as earlier this month, research firm Trendforce noted that despite the strong growth in AI servers, the overall server market could decline his year, because AI-related servers still make up less than 10% of overall server shipments today. That was before Nvidia's blowout guidance, so it's unclear if server shipments will do better-than-expected relative to Trendforce's mid-May update.
The second point Huang emphasized was Nvidia's advantage over emerging competitors. No doubt, with the rise of AI, more start-ups and large chip names will enter the field. However, Nvidia has a sizable lead and a moat due to its CUDA programming platform. CUDA is a collection of software, programming libraries, and APIs which Nvidia developed all the way back in 2006 to enable developers to program GPUs, initially built for computer graphics, for accelerated computing applications.
CUDA has now become a standard upon which AI gets programmed, not unlike Microsoft's Windows operating system for PCs. Moreover, Huang made the point that the ability to program Nvidia's general-purpose GPUs for any application means its expensive chips always will be utilized. On the flip side, some emerging Nvidia alternatives, which are perhaps specialized for a single task, might only be used for a specific purpose. That means a very expensive data center server wouldn't be fully utilized all the time, therefore increasing the total cost of ownership to the data center operator.
So with its CUDA software moat and technology lead, it's no wonder investors are bidding Nvidia up to a high valuation today.
