Amazon is doubling down on artificial intelligence infrastructure, aiming to make AI significantly cheaper through the development of custom chips. CEO Andy Jassy stated that AWS sees no decline in demand, despite macroeconomic pressures and emerging efficient models from competitors.

AI Still Too Expensive—For Now

Amazon wants to make AI accessible through cost-cutting innovation

Speaking on CNBC, Andy Jassy emphasized that AI infrastructure remains prohibitively expensive for many enterprises. AWS is addressing this by building its own AI chips to reduce the cost of inference—the process of generating predictions from trained models.

According to Jassy, “The lower that we can make the cost of AI, the more customers are going to use it.” His comments suggest that Amazon is focused not just on building infrastructure, but on making it more efficient and affordable.

Custom Chips Drive AWS Strategy

Performance gains and lower cost per inference

AWS has engineered custom chips that outperform traditional GPUs by 30% to 40% in price efficiency, according to Jassy. These chips are central to Amazon’s mission of reducing AI inference costs, which are expected to dominate spending as AI usage scales.

Hardware development is paired with improvements in software and architecture, allowing Amazon to stay ahead as inference becomes the primary cost burden in enterprise AI workloads.

Rising Demand Defies Market Fears

No slowdown in infrastructure expansion despite global pressures

Contrary to predictions that more efficient AI models might reduce demand for new infrastructure, AWS reports unabated growth. Jassy dismissed the idea that companies are scaling back on building centers, indicating that the hunger for AI compute remains strong.

Even with geopolitical and economic uncertainties—such as tariffs and inflation—companies are investing heavily in AI, treating it as a critical part of future competitiveness.

Echoes of the Early AWS Days

Lower costs lead to greater innovation, not lower spend

Jassy compared the current AI boom to the early days of cloud computing. Lowering unit costs didn’t reduce total spending—it unleashed innovation. The same principle applies today with AI: as costs fall, companies deploy more advanced solutions, expanding their usage.

“Lower cost allows them to save money, but they don’t spend less. It unleashes them to do more,” he said, underlining the expansive potential once barriers to entry are removed.

Impact on Web3 and Crypto

Cheaper AI could unlock blockchain-native innovation

Jassy’s focus on reducing AI costs could have a ripple effect in blockchain and crypto sectors, where high compute costs have long been a barrier. More affordable infrastructure may finally enable on-chain AI applications, such as autonomous agents and smart analytics.

With AWS pushing boundaries in AI hardware, developers across sectors could soon have access to powerful, low-cost compute—accelerating adoption in fields where cost has been a bottleneck.