[REVISION HISTORY]

Amazon AI supply chain and acquisition efforts

Updated 4 times since CLSTR started tracking revisions of this situation.

What changed

2026-10-12 18:18 UTC → 2026-10-12 22:14 UTC · added ~~removed~~

Amazon has expanded its involvement in the artificial intelligence supply chain through strategic investments and potential acquisitions. Initially, Amazon secured a direct equity stake in Taiwan-based printed circuit board manufacturer Gold Circuit Electronics Inc. via a private placement totaling approximately NT$1.589 billion (USD 54.45 million). This investment, involving roughly 1.86 million shares, transitioned the company from a corporate client to a formal shareholder to meet the rising demand for specialized circuit boards used in AI servers. Amazon is currently in advanced negotiations to acquire Decart, an Israeli-founded startup now based in San Francisco, in a deal that could be valued at approximately $7 billion. This potential acquisition represents a significant increase from Decart’s May 2026 valuation of nearly $4 billion. The move follows a previous, abandoned attempt by Anthropic to purchase the company for roughly $6 billion in August, a pursuit the company reportedly ended in September after conducting due diligence. Decart, founded in 2023 by engineers Dean Leitersdorf, Orian Leitersdorf, and Moshe Shalev, specializes in the Decart Optimization Stack (DOS). This software allows developers to migrate AI workloads between different hardware architectures, such as Nvidia GPUs, Google TPUs, and Amazon’s Trainium chips, reducing the time needed to optimize neural networks from months to weeks. Decart also develops generative models like ‘Lucy’ for real-time video editing and ‘Oasis’ for generating simulated 3D environments used in training robotics and autonomous driving systems. The startup has raised over $400 million in total funding from investors including Nvidia, Sequoia Capital, and Benchmark. Amazon, a major investor in Anthropic, could utilize Decart’s technology to strengthen its AWS ecosystem and reduce reliance on Nvidia hardware by making its custom Trainium chips more accessible to developers.

Versions

Only revisions since CLSTR began indexing content versions appear here. Select a version to see what changed compared to the one before it.