Amazon is reportedly exploring a new, multi-billion-dollar investment in Anthropic, an AI firm known for its Claude models. This potential funding round would build on Amazon’s substantial $4 billion commitment in 2023, which cemented Amazon Web Services (AWS) as Anthropic’s primary cloud provider.
However, this time there’s a catch: Amazon’s backing is said to be contingent on Anthropic adopting Amazon’s custom AI chips, Trainium and Inferentia, instead of Nvidia’s GPUs.
Custom Chips vs. Nvidia’s Industry Hold
Amazon’s push for custom chip integration is not just a financial move; it’s a strategic step to minimize reliance on Nvidia’s dominant hardware, which has become the go-to for AI model training due to its high performance. Trainium and Inferentia, designed specifically for machine learning tasks, are said to offer up to 40-50% better price-to-performance ratios.
For Anthropic, which leans on Nvidia for its robust and proven capabilities, this presents a dilemma. Yet, with funding needs reportedly reaching $2.7 billion to scale its AI models in 2024, Amazon’s offer could be critical for sustaining operations and growth.
|
Related: |
Anthropic’s pursuit of new funding from Amazon could elevate its valuation to as much as $40 billion, doubling its worth from earlier this year. The company, founded in 2021, has already raised around $9.7 billion to date. Competition remains fierce in the AI sector, with OpenAI seeking $6.5 billion to boost its valuation to $150 billion.
Claude Models Enter U.S. Defense Sphere
Earlier this week, Anthropic announced a partnership with Palantir Technologies and AWS to bring its Claude 3 and 3.5 models into U.S. defense and intelligence work. The models will operate on AWS’s Impact Level 6 (IL6) cloud, which is certified for managing highly classified government data. This collaboration extends AI’s footprint in government applications, stirring ethical discussions about deploying advanced AI technologies in military settings.
The latest version, Claude 3.5 Sonnet, released in October 2024, introduced a “Computer Use” feature enabling the AI to perform actions like typing and handling desktop software. The new capability, which simulates keyboard and mouse inputs, is aimed at helping enterprises automate repetitive processes, although it remains experimental and may not yet be reliable for all use cases.
Regulatory Spotlight on Amazon’s Investments
Amazon’s increasing investment in AI firms has not gone unnoticed by regulators. In August, the UK’s Competition and Markets Authority (CMA) launched an inquiry into Amazon’s $4 billion investment in Anthropic to determine whether it could pose anti-competitive risks. This aligns with a broader pattern of scrutiny faced by tech giants for their influence in the AI sector. Microsoft and Alphabet have also seen their investments subjected to regulatory checks.
|
Related: |
Amazon’s Broader Energy Investments
In a related strategic move, Amazon has been investing in sustainable energy solutions to power its expanding cloud infrastructure. In October 2024, the company committed $500 million to develop small modular reactors (SMRs) with Dominion Energy and Energy Northwest. These reactors offer quicker deployment and lower emissions than conventional nuclear plants, aligning with Amazon’s goal of securing reliable, eco-friendly energy for its data centers.
Military AI and Ethical Concerns
The inclusion of AI in military projects continues to raise questions about ethical use. Anthropic’s entry into defense follows OpenAI’s policy shift in January 2024, which allowed for military applications of its models. This change was closely followed by the appointment of Paul M. Nakasone, former NSA director, to OpenAI’s board. The move recalls Google’s decision to step away from Project Maven in 2018 after employee backlash, highlighting the ongoing challenges of navigating military AI ethics.
