Business 4 min read By Bethany Hadley
Amazon’s Custom Chip Business Hits $25 Billion Run Rate
Amazon’s custom chip business has reached a $25 billion annual run rate, signalling strong growth in its cloud and AI infrastructure operations.
Amazon’s custom semiconductor business has crossed a $25 billion annual run rate, a milestone that underscores the company’s deepening push into chip design and its growing reliance on in-house silicon to power its cloud and artificial intelligence operations. The figure, reported by Yahoo Finance, reflects the scale of Amazon Web Services’ (AWS) custom chip programme, which includes the Graviton processor family and the Trainium and Inferentia accelerators designed for AI workloads.
The run rate, which measures the annualised value of current revenue, suggests that Amazon’s chip business is now a significant contributor to the company’s overall financial performance. Analysts have noted that the growth of custom silicon is part of a broader industry trend, as major cloud providers seek to reduce their dependence on traditional chip suppliers such as Intel, AMD, and Nvidia. By designing its own processors, Amazon can optimise performance and cost for its specific workloads, offering customers more efficient computing options.
The news comes amid a wider push by Amazon to expand its AI capabilities. The company has invested heavily in machine learning infrastructure, and its custom chips are designed to handle the intense computational demands of training and running large language models. AWS customers, including startups and large enterprises, are increasingly using these chips to reduce costs compared with off-the-shelf alternatives.
Amazon’s chip business is part of a larger strategy to differentiate its cloud offering. The company has been developing custom silicon for several years, with the Graviton processors first launched in 2018. Since then, the line has expanded to include multiple generations, each promising better performance and energy efficiency. The Trainium and Inferentium chips, introduced more recently, target AI training and inference workloads respectively, competing directly with Nvidia’s GPUs.
The $25 billion run rate is a notable achievement for a business that was initially seen as a niche experiment. It also highlights the growing importance of semiconductors in the tech sector, as companies across the industry grapple with supply chain constraints and rising demand for computing power. For Amazon, the milestone could bolster investor confidence, as the company seeks to justify its heavy capital expenditure on data centres and chip development.
However, the news also raises questions about the competitive landscape. Nvidia remains the dominant player in AI chips, and its GPUs are widely used across the industry. Amazon’s custom chips are designed to offer a more cost-effective alternative, but they may not match Nvidia’s performance in all scenarios. Still, the rapid growth of Amazon’s chip business suggests that customers are willing to adopt alternatives, particularly as AI workloads become more diverse.
For investors, the $25 billion run rate is a positive signal, but it is not the only factor to consider. Amazon’s overall business remains heavily reliant on its retail and cloud segments, and the chip business is still a relatively small part of the company’s revenue. Nevertheless, the milestone reflects Amazon’s ability to innovate and execute in a highly technical field, which could support its long-term growth prospects.
