Amazon CEO Andy Jassy’s annual shareholder letter includes a diss song to Kendrick Lamar. What if rappers were the smooth-talking CEOs of corporations and not poetic Pulitzer Prize-winning musicians?
That means you need to know the history to understand all the competition Jussie aspires to, along with a cute personal story about his unfulfilled dreams of becoming a sportscaster and watching hockey games with his father.
Of course, Jassie doesn’t directly throw the gauntlet. He takes a more nuanced approach. For example, in his challenge to NVIDIA, he wrote, “We have a strong partnership with NVIDIA and will always have customers who choose to run NVIDIA and will always support these chips in the cloud.”
But he also notes that “nearly all AI to date has been run on NVIDIA chips, but a new transition has begun.” He said AWS customers “want better value for money,” which means Amazon’s own Trainium AI chips.
Demand for the chip is so high that the latest Trainium 3 capacity is nearly sold out, Jassy said. Surprisingly, Trainium4 capacity is also almost sold out, and it will still be 18 months before it becomes available, he said.
This means that Trainium’s annual revenue run rate has reached $20 billion. But if Amazon were a chipmaker selling its products to other companies, he hypothesized, the ARR would be $50 billion.
Indeed, Nvidia’s actual revenue last year was $215.9 billion. Nvidia may not be budging yet. Still, Jassy presents Trainium as a formidable up-and-comer.
tech crunch event
San Francisco, California
|
October 13-15, 2026
Jassy didn’t spare Intel either. He points out that AWS’s homegrown Graviton CPU, a competitor to the Intel x86 architecture, is “currently used extensively by 98% of the top 1,000 EC2 customers, or some of the world’s largest enterprises.” He writes that the two companies even asked to “purchase all of our Graviton instance capacity in 2026” (my emphasis). “We can’t agree to these requests given the needs of other customers, but we can understand the demand.”
He also promised that Amazon’s Starlink competitor Amazon Leo, scheduled to launch in mid-2026, is already a success. Contracts include Delta Air Lines, AT&T, Vodafone, Australia’s National Broadband Network and NASA.
Interestingly, he also said that Amazon might consider selling robots at some point. All the data from a million warehouse robots could be turned into “robotic solutions” for industrial and consumer use, he wrote. Will Amazonian humanoids exist in our future? Let’s take a look. He also talked about other Amazon businesses, including same-day delivery, groceries and drones.
But primarily, Jassy sought to make a case for the hundreds of billions of dollars in capital investment he had committed. In February, he announced plans to spend $200 billion in capital spending in 2026, primarily to build AWS data centers. That’s more than other large technology companies that also spend a lot of money on capital expenditures. Mr. Jassy’s pitch to shareholders makes sense, given that Amazon’s stock price has fallen below $200 a share and hasn’t recovered.
“We are not investing approximately $200 billion in capital investment in 2026 based on a hunch,” he wrote, citing as an example the model maker’s commitment to spend $100 billion on AWS in its contract with OpenAI. Of course, some doubt whether OpenAI will be able to deliver on all of its spending promises.
In a nod to this, Jassy claims that aside from OpenAI, there are “several other customer deals completed (unannounced) or currently in the works” to purchase AWS capacity.
We’ll have to wait and see. Those who create bubbles are never aware of (or admit to) the existence of bubbles. “I have observed public debate about whether this technology is overhyped and whether we are in a ‘bubble.'” But in the letter, he asserts that is not the case, at least for Amazon.
