NVIDIA has invested in the startup Decart, which specializes in AI infrastructure software that enables developers to run models across various processors, including those from NVIDIA, Amazon, and Google. This investment comes as Decart raised $300 million at a valuation approaching $4 billion, with Radical Ventures leading the funding round. Decart’s flagship product, the Decart Optimization Stack, simplifies the process of moving AI models between different hardware systems without requiring significant rebuilding efforts. This aligns with a growing trend among AI labs and enterprises that prioritize hardware-agnostic solutions to enhance pricing flexibility and reduce vendor lock-in in an increasingly competitive landscape dominated by major cloud providers like Amazon and Google, who are advancing their own AI technologies to challenge NVIDIA’s position.
Amazon: Amazon is a global technology and e-commerce company whose AWS cloud division provides custom AI chips such as Trainium and Inferentia alongside NVIDIA GPUs. In this news, Amazon is highlighted as Decart’s largest customer, using Decart’s optimization stack to deploy AI across properties like Twitch, its online retail business, and content studios while enabling workloads to shift across different accelerators.
Decart: Decart is an AI infrastructure company that specializes in ultra-optimized systems for running large models and real-time world simulations across heterogeneous hardware. In this news, Decart is the startup raising a major funding round and attracting investment from NVIDIA to expand its platform that lets AI labs seamlessly switch workloads between chips from providers like NVIDIA, Amazon, and Google.
NVIDIA: NVIDIA is a leading semiconductor and computing company best known for its GPUs, which power many of today’s advanced AI training and inference workloads. In this context, NVIDIA is both a strategic partner and investor in Decart, backing the company’s software that makes it easier for AI developers to run and migrate models across NVIDIA hardware and competing accelerators.
Radical Ventures: Radical Ventures is a venture capital firm focused on backing companies in artificial intelligence and deep technology. In this story, Radical Ventures is the lead investor in Decart’s latest funding round, signaling strong institutional belief in Decart’s role as a key abstraction layer between AI models and underlying chips.
Decart Optimization Stack: The Decart Optimization Stack, or DOS, is Decart’s core software platform for orchestrating and optimizing AI models across diverse hardware, including GPUs and custom accelerators. It is central to this news as the product that lets AI labs move models between NVIDIA, Amazon, Google, and other chips without extensive re-engineering, making cross-chip flexibility commercially viable.
Vendor_strategy: Chip makers and cloud providers have started investing in orchestration and optimization startups as a way to keep their hardware attractive in multi-vendor environments rather than insisting on exclusive stacks.
Cloud_chip_competition: Cloud providers like Amazon, Google, and Microsoft are rapidly advancing their own AI accelerators to compete more directly with NVIDIA’s GPUs for training and inference workloads.
AI_infrastructure_trend: AI labs and enterprises are increasingly seeking hardware-agnostic software layers so they can negotiate better chip pricing and avoid being locked into a single vendor.