AWS Slashes EC2 GPU Prices 40% to Counter Google's TPU Momentum
Amazon just cut GPU instance prices across the board. If you're running AI workloads, recalculate your cloud spend — the numbers just changed significantly.
🔍 What Happened
AWS announced pricing cuts of up to 40% on P5 (H100), P6 (Blackwell), and G6 (L4) GPU instances, effective immediately across all regions. Savings Plans commitments see additional 15% discounts. The move follows Google Cloud's aggressive TPU pricing, which underpriced AWS GPUs by 30-50% on comparable workloads.
💡 Why It Matters
AI workloads are the largest new growth vector for all hyperscalers. Losing GPU customers to Google Cloud would hurt AWS's AI narrative badly. This price cut — AWS's largest in years — signals that competitive pressure is finally cracking through the big three's pricing discipline.
🏢 Impact on Business & Users
Enterprises with committed AWS spend should renegotiate now. Mid-size AI companies save $100K-500K annually. Meanwhile, Google's response is already in motion — expect TPU v6 pricing details within weeks. Microsoft Azure hasn't responded yet but will need to.
👀 What to Watch Next
Watch for Azure's response and whether it cuts reserved pricing or introduces competing instance types. Also monitor CoreWeave, Lambda, and other GPU-specialty clouds — they may lose the pricing edge that's grown their market share.
Frequently Asked Questions
Enjoyed this article?
Get stories like this delivered to your inbox.
