Nvidia (NVDA +1.22%) has been a key player in the stock market over the last few years. With a market cap of over $4.4 trillion, its only rivals in history are the companies founded to run colonial ventures of European empires centuries ago, like the Dutch East India Company.
Nvidia dominates the artificial intelligence (AI) hardware market. But there's too much money on the table for other companies not to try and break into the sector.
Qualcomm's (QCOM +0.36%) new chips have the potential to compete with Nvidia in the data center space. And Alphabet (GOOG 1.34%) (GOOGL 1.43%) is entering the chip market with its Ironwood Tensor Processing Unit (TPU). All this means that cracks are beginning to emerge in Nvidia's facade of total market domination.
Image source: Getty Images.
From happy accident to near-monopoly
Nvidia made some of the best computer hardware on the market long before ChatGPT launched in 2022. Its chips were (and still are) used for everything from high-end video game graphics to crypto mining. But Nvidia graphics processing units (GPUs) were also fantastic for running AI programs. That, combined with its engineering prowess and CUDA software platform, gave Nvidia a serious competitive edge. Today, Nvidia controls between 85% and 90% of the $44.9 billion global AI chip market. Its Blackwell GPUs are the most coveted hardware in the world, and its only real competitor until now was Advanced Micro Devices (AMD +2.27%).
In late October, however, Qualcomm unveiled two AI chips aimed squarely at Nvidia. The AI200 and AI250 are set to launch in 2026 and 2027, respectively. Can Qualcomm come for the king and land a direct hit? It's more possible than you might think.
NASDAQ: QCOM
Key Data Points
In the short term, Qualcomm enjoyed a small leap in share price and some media hype with the announcement. However, Qualcomm started declining and is down about 1.5% in the weeks following its announcement, as I write this.
So the question is, why? My answer is that Nvidia's moat is large enough to intimidate investors, but it's not impassable.
A war on two fronts
To say that Nvidia has a lead in the AI chip industry would be an understatement. You don't come to control 90% of a market by accident. The main reason behind Nvidia's success is that, in many ways, its chips are simply better than the competition. However, Nvidia's chips aren't perfect, and there is room to compete in terms of capabilities, costs, and energy efficiency.
Enter Qualcomm's AI200 and AI250 chips. They don't have the same raw horsepower as Nvidia's chips, but they are more energy efficient and optimized for running AI in practical applications rather than training. According to Qualcomm, this focus on efficiency means the AI200 chip uses 35% less power than a GPU produced by Nvidia.
Qualcomm's chips are positioned to be cheaper to purchase and run than Nvidia's, which will be critical for data center applications and for companies that need to keep costs low and don't need all the bells and whistles, especially considering that AI infrastructure spending is set to exceed $2.8 trillion by 2029, according to some estimates.
On the high-end side of AI hardware is Alphabet's Ironwood chip. It's a TPU, which is optimized for training new AI, one of the most energy-intensive parts of the AI industry.
Ironwood is capable of matching Blackwell's performance for the same amount of power; it also has the potential to scale better with size than Blackwell.
Will it allow Alphabet to eclipse Nvidia? Probably not on its own, but it does provide a compelling alternative, and Meta Platforms might agree. It's reportedly in talks to spend billions on Alphabet TPUs.
So, Nvidia now has serious competitors on both the high and low ends of AI hardware. And, while it may take some time for Qualcomm, Alphabet, and AMD to narrow Nvidia's lead, the opportunity is there.
Moat, meet bridge
Nvidia's perennial rival AMD is also chipping away at its lead. AMD has been in the industry's No. 2 spot for years and controls 3% to 5% of the market, but it recently signed an agreement with OpenAI that will see its GPUs put to use running ChatGPT. While Nvidia has a head start, there is serious demand for competitors.
It's a tall order, but Nvidia's competitors are catching up faster than you might think. It's also worth noting that Alphabet's year-to-date stock return is 68%, more than double Nvidia's 30%.
I don't think either Qualcomm's AI chips or Alphabet's Ironwood will sink Nvidia's battleship, but all the new competition will add up and could make a dent.
Any one of these companies is a serious contender in the AI hardware market. Nvidia isn't a monopoly, and while Qualcomm's chips haven't changed the game on their own, it would seem that the game is afoot.