Amazon CEO Sees AI Infrastructure as Key to Future Growth
Amazon CEO Andy Jassy believes artificial intelligence infrastructure is still too costly for widespread enterprise adoption. However, Amazon Web Services (AWS) is working to change this by developing custom chips and reducing inference costs.
During an interview on CNBC’s Squawk Box,Jassy addressed concerns that recent AI model efficiency improvements might lessen the need for infrastructure expansion. He stated, “We have very high demand. I don’t see us slowing down our building centers anytime soon.”
Despite economic uncertainties and potential tariffs, AWS is not experiencing a decline in AI infrastructure demand. Jassy explained that while efficient models are beneficial, the challenges of AI infrastructure are more complex and enduring.
Lowering AI costs can unlock greater customer spending, similar to the early days of AWS. “It allows them to save money but doesn’t reduce their spending. Instead, it enables more innovation,” Jassy saeid.
Two main factors can reduce AI costs: custom chips and inference costs. AWS’s custom AI chips offer 30% to 40% better price performance than current GPU-based instances. Reducing inference costs involves hardware and software advancements.
Jassy emphasized AWS’s mission: “Our team feels responsible for making AI costs considerably lower than today.” This push could benefit the crypto sector, making blockchain-native AI applications more feasible.
