Article by Edward Sheldon
AI Chip Stocks: 4 Companies Seeing Prolific Growth in 2026
April 22, 2026 | Research Insights
The explosive rise of artificial intelligence (AI) has turned semiconductors or “chips” into a critical resource. From Generative AI apps like ChatGPT and Claude to data centers and humanoid robots, the technology depends on chips for computing power.
In this article, we are going to highlight four chip stocks that currently stand at the heart of the AI theme. As the data center buildout accelerates and the demand for computing power increases, these companies are seeing prolific growth.
Nvidia: Expecting Huge GPU Sales Through 2027
It’s impossible to discuss AI chips and not mention Nvidia. To date, it has almost single-handedly powered the AI revolution with its high-powered GPUs, and this is reflected in its revenue, which has soared from $17 billion to $216 billion1 over the last five fiscal years.
In March, investors got a look at Nvidia’s AI roadmap at its GTC conference. Here, the company revealed2 its new Vera Rubin platform. This platform, which is now in full production, features seven different Nvidia chips operating together as one AI supercomputer and is designed to power every phase of AI from pretraining, post-training, and test-time scaling to real-time agentic inference. In terms of demand, CEO Jensen Huang said that he expects $1 trillion in cumulative revenue through 2027 between Rubin and current generation Blackwell chips.
Note that at GTC, Huang also introduced the Groq 3 Language Processing Unit (LPU). Part of the Rubin platform, this is an ultra-high-speed inference chip designed to meet the demands of agentic systems. Alongside this product, the company launched the Groq 3 LPX platform – a server rack powered by 128 individual Groq 3 LPUs. With these new products in its portfolio, Nvidia looks well placed to continue playing a major role in the AI revolution despite an increase in competition.
Amazon: Triple-Digit Chip Growth
One company that could potentially become a major competitor to Nvidia in the future is Amazon. To date, it hasn't received much credit for its AI chips, however, after recently providing investors with some insights into the scale and growth of this area of its business, that appears to be changing.
In Amazon’s 2025 annual letter3 to investors, posted in the first half of April, CEO Andy Jassy said that its chips business is “on fire”. He also said that it will be “much larger than most think”. More specifically, Jassy said that the company’s Trainium2 chips – which have about 30% better price-performance than comparable GPUs – are largely sold out. Meanwhile, Trainium3, which just started shipping at the start of 2026 and is 30-40% more price-performant than Trainium2, is nearly fully subscribed.
In terms of revenue, Jassy said that the annual revenue run rate for its chips business (inclusive of Trainium, Graviton, and Nitro) is now over $20 billion, and growing at a triple-digit percentage rate year-over-year. However, he noted that if the chips business was a stand-alone business, and sold products to other companies, its annual run rate would be close to $50 billion.
It’s worth noting that the fact that Amazon produces its own chips gives it a major advantage in terms of efficiency. In the annual letter, Jassy said that he expects that Trainium will save the company “tens of billions” of capex dollars per year and provide several hundred basis points of operating margin advantage versus relying on others’ chips for inference.
Broadcom: Major Chip Deals with Google and Meta
Another chip designer that is currently having success in the AI space is Broadcom. It makes custom chips (which it calls “XPUs”) for companies such as Alphabet and Meta Platforms, and it is seeing high demand as hyperscalers move to reduce their reliance on Nvidia.
On April 6, Broadcom announced4 that it had signed a long-term agreement with Google to develop and supply future generations of custom AI chips and other components for the company's next-generation AI racks through 2031. On the same day, it also announced that it had expanded its deal with Anthropic, giving the AI company access to around 3.5 gigawatts worth of computing capacity drawing on Google’s AI processors.
Shortly after this, on April 145, Broadcom announced that it will work with Meta to produce several generations of custom AI processors as the social media giant builds out the computing capacity needed to power AI features across its apps. This tie-up will last until 2029 and includes an initial commitment of over one gigawatt of computing capacity.
Looking at these deals, it’s clear that this chip company – which reported 106%6 year-over-year AI product growth in Q1 – is becoming an integral part of the AI revolution. As hyperscalers pivot from an "everything-Nvidia" model toward a more diversified, multi-layered infrastructure, they are increasingly relying on Broadcom.
Taiwan Semiconductor: The King of AI Chip Production
If there’s one thing that Nvidia, Broadcom, and Amazon have in common, it’s that they all rely on Taiwan Semiconductor Manufacturing Company for production. It’s the most dominant chip manufacturer globally and this is reflected in its growth.
On April 16, Taiwan Semi posted its earnings7 for Q1 2026 and they were very strong. For the quarter, revenue amounted to NT $1,134.10 billion (approx. USD $35.9 billion), a 35.1% increase year-over-year. Meanwhile net income was NT $572.48 billion, up 58.3% year-over-year. Note that advanced technologies (7nm and more advanced) accounted for 74% of total wafer revenue while 3nm shipments accounted for 25% of wafer revenue so it is clear that performance is now inextricably tied to the global buildout of high-performance AI infrastructure.
Looking ahead, TSMC raised its forecast for 2026 revenue growth. It is now projecting8 an increase of more than 30% in US dollar terms compared to 2025. In terms of capex, the company is guiding its 2026 capital budget toward the high end of its $52 billion to $56 billion range. This capital will be focused on advanced process technologies, advanced packaging, and global expansion.
Investing in AI Chip Stocks
Investors should note that while all of these companies are having success today, there is no guarantee that they will continue to do so. Therefore, there is risk involved in investing in these businesses directly.
One way to potentially reduce this risk is to opt for an artificial intelligence ETF such as the Themes Generative AI ETF (WISE). With this kind of product, an investor may be able to gain exposure to several AI chip stocks all at once, potentially reducing single-stock risk while still being positioned to capture the growth of the industry.
Disclaimer: The author has positions in Nvidia and Amazon
Footnotes:
1LSEG, as of April 22, 2025
2Investing.com, NVIDIA at GTC 2026: AI Expansion and Strategic Partnerships, as of March 16, 2026
3Amazon News, CEO Andy Jassy’s 2025 Letter to Shareholders, as of April 9, 2026
4Reuters, Broadcom signs long-term deal to develop Google’s custom AI chips, as of April 7, 2026
5Broadcom, Broadcom Announces Extended Partnership with Meta to Deploy Technology to Support Multi-Gigawatts of Meta’s Custom Silicon, MTIA, as of April 14, 2026
6Broadcom, Broadcom Inc. Announces First Quarter Fiscal Year 2026 Financial Results and Quarterly Dividend, as of March 4, 2026
7TSMC Reports First Quarter Earnings, as of April 16, 2026
8TSMC 2026 First Quarter Earnings Conference, as of April 16, 2026
Author is a contractor of Leverage Shares LLC, a U.S. affiliate of Themes Management Company LLC. Leverage Shares LLC provides certain services to Themes under an intercompany services agreement.
