Share it with your senior IT friends and colleagues
Reading Time: 2minutes
While many are talking about better AI models, Amazon is building something even bigger, an entire AI ultra-cluster.
Under Project Rainier, Amazon is constructing a constellation of massive data centres, starting with 7 in Indiana (and potentially up to 30) in other locations.
All connected into one high-performance AI network.
What’s the mission?
To power AI workloads for partners like Anthropic (yes, the Claude folks).
Amazon has already invested $8B in them, and now Anthropic may use all of New Carlisle’s compute to run a single AI system.
Custom chips. Custom network. Custom scale.
Instead of relying on Nvidia or AMD, Amazon is going all-in on its own Trainium processors and its own Elastic Fabric Adapter for data centre interconnects.
Trainium 3 is expected to deliver 4x the performance with 60% less energy.
The big picture:
Amazon is joining the arms race for AI infrastructure:
– Microsoft: $80B investment in data centres
– Alphabet: ~$75B planned for infra
– OpenAI’s Texas mega-complex: ~$60B
But Amazon could have a dual advantage:
If Anthropic succeeds, → massive strategic edge in AI
If not → all that infrastructure boosts AWS, which is already a cash machine
What are your thoughts? Are we witnessing the quiet rise of Amazon as the AI infrastructure kingmaker?
The most up-to-date AI + Gen AI Coaching for senior IT professionals
In case you are looking to learn AI + Gen AI in an instructor-led live class environment, check out these courses
