Amazon is heavily restructuring Alexa+'s AI architecture to reduce dependence on Anthropic models
7x24h News
Amazon is heavily restructuring Alexa+'s AI architecture to reduce dependence on Anthropic models
According to Business Insider, Amazon is aggressively restructuring the AI architecture of Alexa+, routing more requests to in-house models and reducing calls to Anthropic Claude models through caching mechanisms, while optimizing GPU utilization to cut operating costs. Internal documents show that Alexa+'s AWS cloud costs are projected to reach approximately $1.7 billion in 2026, nearly triple the previous year, and exceeding target costs by about 60%. Even with approximately $450 million in potential savings identified, internal assessments still indicate that the business will struggle to meet financial targets. Amazon CEO Andy Jassy has publicly emphasized the urgency of reducing AI inference costs, and this architectural adjustment is the concrete implementation of this strategic direction.
TechFlow news, July 23, according to Business Insider reports, Amazon is significantly restructuring the AI architecture of Alexa+, routing more requests to self-developed models, and reducing calls to Anthropic Claude models through caching mechanisms, while optimizing GPU utilization to reduce operating expenses. Internal documents show that Alexa+'s AWS cloud costs are expected to reach approximately $1.7 billion in 2026, nearly triple compared to the previous year, and exceeding target costs by about 60%. Even though approximately $450 million in potential savings have been identified, internal assessments still believe the business will struggle to meet financial targets. Amazon CEO Andy Jassy has publicly emphasized the urgency of reducing AI inference costs, and this architectural adjustment is precisely the implementation of this strategic direction.



