Quick Read
-
Amazon (AMZN) has quietly built a $20 billion chip business growing over 100% annually, with Jassy projecting a $50 billion run rate.
-
Jassy has signaled potential external Trainium sales, positioning Amazon as a direct rival to Nvidia inside a $150 billion annualized AWS cloud empire.
-
The analyst who called NVIDIA in 2010 just named his top 10 stocks and Amazon wasn’t one of them. Get them here FREE.
Artificial intelligence has turned semiconductor stocks into the market’s biggest winners. Nvidia (NASDAQ:NVDA), Broadcom (NASDAQ:AVGO), Advanced Micro Devices (NASDAQ:AMD), and other chipmakers have captured investors’ attention as demand for AI infrastructure continues to climb. But what if one of the fastest-growing chip businesses isn’t actually a semiconductor company at all?
That’s the question investors should be asking about Amazon (NASDAQ:AMZN). Most shareholders still view Amazon as an e-commerce and cloud computing giant. Yet beneath the surface, the company has quietly built a semiconductor operation that is becoming one of the most important pieces of its long-term AI strategy — and the numbers are getting too large to ignore.
Amazon’s Hidden $50 Billion Chip Business
Amazon’s custom silicon division is built around a simple idea: instead of buying every processor from outside suppliers, design chips optimized specifically for Amazon Web Services (AWS).
That strategy has produced a growing portfolio of custom processors, including Graviton CPUs and Trainium AI accelerators. Amazon’s chip operation is already generating roughly $20 billion in annual revenue and expanding at more than 100% year over year. If Amazon sold chips directly, CEO Andy Jassy says the run rate would approach $50 billion.
To put that into perspective, here’s how Amazon’s chip business compares with some well-known semiconductor segments:
|
Company/Segment
|
2025 Revenue
|
|
Amazon Internal Chip Business
|
~$20 billion
|
|
Broadcom AI Revenue
|
~$20 billion
|
|
Intel (NASDAQ:INTC) Data Center & AI
|
~$16.9 billion
|
|
AMD Data Center Segment
|
~$16.6 billion
|
The comparison is surprising because Amazon isn’t manufacturing chips for retail sale. Instead, AWS creates enormous internal demand by deploying its processors across one of the world’s largest cloud infrastructures.
The analyst who called NVIDIA in 2010 just named his top 10 stocks and Amazon wasn’t one of them. Get them here FREE.
When AWS installs millions of dollars worth of Trainium or Graviton chips into its own data centers, that economic value still flows through Amazon’s semiconductor operation.
24/7 Wall St.
Beyond the Prime boxes lies a semiconductor empire growing at 100% year-over-year. Amazon isn’t just buying chips anymore—it’s building the hardware that threatens to topple industry giants. © 24/7 Wall St.
Vertical Integration Creates an AI Advantage
Let’s look at why Amazon’s approach matters.
Nvidia’s dominance comes from selling GPUs to cloud providers. Amazon is taking a different path. It is becoming both the customer and the supplier. By designing its own processors, Amazon can reduce dependence on third-party vendors, lower operating costs, and optimize performance for specific AI workloads. The strategy appears to be working.
Graviton-based instances reportedly deliver better price-performance than comparable x86 alternatives for many cloud applications. Meanwhile, Trainium is being positioned as a lower-cost option for training large AI models.
AWS generated $150 billion in annualized revenue during the first quarter of 2026 and remains Amazon’s primary profit engine. Even small improvements in computing efficiency can translate into billions of dollars in savings and additional capacity.
That creates a built-in customer that few semiconductor companies can match.
Could Amazon Eventually Challenge Nvidia?
Granted, Nvidia remains the undisputed leader in AI hardware. The company generated more than $193 billion in data center revenue during its most recent fiscal year and controls much of the AI accelerator market.
But Amazon may not need to beat Nvidia to create enormous shareholder value. The bigger opportunity is that Amazon could eventually sell more of its custom AI chips beyond its internal infrastructure. Jassy recently acknowledged the possibility of expanding external sales of Trainium processors, potentially placing Amazon into more direct competition with Nvidia and AMD.
If AWS customers increasingly adopt Trainium-powered services, Amazon gains another growth engine layered on top of cloud computing.
In effect, investors get exposure to e-commerce, cloud computing, AI infrastructure, and semiconductor design through a single stock.
Key Takeaway
In short, Amazon is no longer just a cloud company that happens to design chips. It is becoming a semiconductor powerhouse that happens to run the world’s largest cloud platform.
Nvidia still sets the pace in AI hardware. That said, Amazon’s combination of AWS scale, custom silicon, and growing AI demand creates a competitive advantage few companies can replicate. With an estimated $50 billion chip business already growing at triple-digit rates, smart investors may want to stop viewing Amazon solely as an e-commerce stock.
Ultimately, AI chips aren’t just fueling semiconductor stocks anymore. They are becoming a major reason Amazon’s next decade of growth could look very different from its last.
The analyst who called NVIDIA in 2010 just named his top 10 AI stocks
This analyst’s 2025 picks are up 106% on average. He just named his top 10 stocks to buy in 2026. Get them here FREE.
