Nvidia (NASDAQ: NVDA) has done it again. The $4.7 trillion tech behemoth delivered another blockbuster earnings report on Wednesday, solidifying its near-monopoly on the computing infrastructure that powers the artificial intelligence (AI) revolution. With third-quarter revenue hitting a record US$57 billion, up 62% year-over-year, and earnings per share soaring 67% to US$1.30, the AI stock’s execution is nothing short of flawless.
Cloud GPUs sold out during the quarter, market demand for Nvidia’s Blackwell architecture remains insatiable, and management forecasts near US$10 billion sequential growth in next quarter’s revenue to US$65 billion. But for investors looking at 2026 and beyond, Nvidia’s perfection poses a paradox: the law of large numbers is finally catching up.
While Nvidia remains the undisputed king of AI, its size is a critical factor now. Adding US$10 billion in revenue, a massive feat for any other AI stock, barely moves the needle for a stock valued at nearly US$5 trillion. The easy money has been made. The AI-accelerator market share is Nvidia’s to lose, and competitors are circling.
For 2026, I am shifting my focus to three AI stocks that have more room to run. These growth stocks are smaller than Nvidia, meaning new AI revenue can propel their stock price growth much faster than it can for the incumbent giant. They have already outperformed Nvidia stock so far this year.
Source: Getty Images
Advanced Micro Devices (AMD)
The narrative for Advanced Micro Devices (NASDAQ: AMD) has shifted in 2025 from “if” it can compete to “when.” While Nvidia dominates the present, AMD, with its MI400 series next-gen GPUs hitting the market in 2026, is positioning itself to capture significant market share starting in the second half of next year.
Recent updates from AMD’s management suggest a massive inflection point is approaching. Beyond competing on raw silicon, AMD could win on value and availability over the next three years. As the AI market matures, hyperscalers (major cloud computing providers) are looking for a second source to avoid vendor lock-in and supply bottlenecks. AMD’s MI-series accelerators are gaining traction, offering a compelling price-to-performance ratio that is becoming impossible for data centre CFOs to ignore.
Unlike Nvidia, which is priced for perfection, AMD is priced for skepticism. This creates an asymmetric upside: if AMD captures even 10–15% of the AI accelerator market in 2026, a conservative estimate given its roadmap, the impact on its bottom line would be transformative in a way that another US$10 billion quarter for Nvidia simply isn’t.
Up 83% year-to-date, AMD stock has widely outperformed Nvidia’s comparable 41% return.
Broadcom
While Nvidia sells general-purpose GPUs, Broadcom (NASDAQ: AVGO) is mastering the art of the “custom silicon” or ASIC (Application Specific Integrated Circuit) market.
The biggest threat to Nvidia isn’t another GPU maker; it’s Nvidia’s own customers. Tech giants like Alphabet, Meta Platforms, and Amazon are increasingly designing their own in-house AI chips to run specific workloads more efficiently than general-purpose GPUs. Broadcom is the primary partner helping these giants design and build those custom chips.
Broadcom’s AI revenue growth is driven by this custom silicon boom, which is accelerating as companies look to lower the total cost of computing.
As AI model workloads move from “training” (where Nvidia dominates) to “inference” (running the models), the efficiency of Broadcom’s custom ASICs becomes a massive competitive advantage. Broadcom offers a way to invest in the AI infrastructure build-out without betting solely on the GPU monopoly holding forever.
Broadcom stock’s 57% year-to-date gain is better than Nvidia’s.
Celestica
Canadian investors who want an AI play directly listed on the TSX could pick Celestica (TSX: CLS) despite its market-beating 230% year-to-date gain. The $49 billion semiconductor manufacturing services company fits the description of “another smaller AI stock that could keep growing revenue in 2026 and beyond”.
Celestica is a “pick and shovel” play on AI. It builds the complex, high-performance connectivity and networking hardware that makes AI clusters work. As data centres upgrade to 800G switches and liquid-cooled server racks to handle the heat of Blackwell and AMD chips, Celestica is the manufacturing partner making modern data center upgrades and new builds happen.
As a manufacturing and supply chain partner, Celestica makes money regardless of whose AI chip wins the war in 2026 and beyond. Whether a data centre installs Nvidia, AMD, or custom Broadcom chips, it still needs the high-speed switching and advanced server racks that Celestica provides. Sustained revenue and earnings growth should help sustain and support CLS stock’s momentum.
