Amazon Says Its AI Chips Will Run a Popular Coding Language

Amazon has said its own AI chips will work with a popular coding language, a step that puts the company into more direct competition with Nvidia, The Economic Times reports.

The claim matters because software, not raw silicon, decides most AI chip purchases today. Developers build models with established tools, and for years those tools have been tied to Nvidia's hardware. If Amazon's chips accept code written in a widely used language, its hardware becomes a realistic alternative rather than a niche product.

Why does software support change the picture?

Nvidia's strength rests on two legs. The first is its GPUs, which train and run large AI models. The second is the software layer around them, which lets developers write code once and run it on Nvidia hardware without rework.

That second leg is the harder one for rivals to break. A competing chip can match on speed or price, but if it forces engineers to rewrite their code, most will stay put. By adding support for a popular coding language, Amazon attacks exactly this barrier. Developers could, in principle, move workloads to Amazon silicon with far less friction.

In plain terms: the announcement is about removing switching costs, not about announcing new transistors.

What is Amazon actually building?

Amazon designs its own AI accelerators through its Annapurna Labs unit and deploys them across its AWS cloud business. The chips serve Amazon's internal workloads and are offered to cloud customers as an alternative to Nvidia GPUs.

The reported move fits a pattern. Amazon, like other large cloud providers, wants to reduce how much it pays Nvidia for AI compute. Every workload that runs on in-house silicon cuts that bill and improves Amazon's margin on AI services. It also gives AWS customers a second option when Nvidia GPUs are scarce or expensive.

The commercial logic is simple: own more of the stack, keep more of the revenue.

How does this challenge Nvidia?

Nvidia has held the largest share of the AI chip market, and its position has looked secure partly because of its software ecosystem. Each rival that adds compatibility with mainstream developer tools narrows that advantage.

Amazon's claim, if it holds up in production use, would mean that code written for one ecosystem could run on Amazon hardware too. That shifts the competition back toward price, capacity and performance — areas where a cloud operator with its own fabs-scale purchasing power can press hard.

For customers, the practical effect is bargaining leverage. A credible alternative gives them room to negotiate on pricing for Nvidia-based instances as well.

What remains unconfirmed?

The report describes an announcement about compatibility, not benchmark results or shipment figures. Amazon has not, in this report, quantified how much performance developers should expect when they run code compiled this way, or which specific chip generations gain the support.

That gap matters. Compatibility claims are easy to make; performance after translation or compilation is where competing chips have historically fallen short. Developers will want independent numbers before they move production workloads.

The takeaway is that Amazon is now competing on developer convenience, not just on hardware specs — and the AI chip market will likely turn on whether that convenience survives contact with real workloads.