- Amazon recently closed its AGI Lab, cut AI roles, and retired products, signaling a strategic shift to focus on AI infrastructure and customer-facing engineering rather than developing frontier models.
- Amazon poured additional funding into Forward Deployed Engineering and raised its 2026 capital spending toward $220 billion.
- Amazon retained its own models, likely to force price pressure rather than chase model market share.
- Last week's arrival of the new DeepSeek model reinforces the need for buyer businesses to keep a deliberate portfolio of model resources.
- This move positions Amazon as a provider of the 'plumbing' for various AI models, including competitors'.
- For businesses buying AI services, this means prioritizing multi-model flexibility, but accepting potential infrastructure lock-in.
- Amazon bets that the model layer is now a commodity; the buyer's takeaway is that the model is an ingredient, not the dish.
- Companies must cultivate internal AI capabilities and judgment to select appropriate models for specific tasks, rather than outsourcing core strategy.
Amazon's recent strategic shift involves closing its AGI Lab and reallocating resources towards AI infrastructure and customer-facing engineering. This decision reflects a broader trend where the model layer is viewed as a commodity, allowing Amazon to focus on providing essential services for various AI models.
The company aims to be the "plumbing" for AI, offering infrastructure that supports models from competitors like OpenAI and Google. Amazon's investment in metering and hardware-adjacent layers ensures revenue generation regardless of the model being utilized. This shift encourages businesses to prioritize multi-model flexibility while acknowledging potential infrastructure lock-in.
Experts suggest that companies must develop internal AI capabilities to effectively select and manage models for specific tasks. George Westerman, a senior lecturer at MIT, emphasizes that organizations must adapt their operations to leverage AI effectively. He notes, “Technology changes quickly, but organizations change much more slowly.” Companies that succeed in AI transformation do so by integrating AI into their core strategies rather than merely adopting new technologies.
As Amazon continues to provide its own models, the focus remains on creating competitive advantages through “Corporate Taste,” which leverages unique domain knowledge to apply AI effectively. This strategic pivot positions Amazon as a key player in the evolving AI landscape, emphasizing the importance of infrastructure and customer engagement over model development.
“Amazon is likely keeping its own models to force price pressure rather than chase market share. Buyers should build in-house judgment, since “Corporate Taste” — leveraging unique domain knowledge — is the only layer that plays offense, while MIT Sloan's George Westerman cautions that “organizations change much more slowly” than technology.”
Forbes
Yahoo Finance UK
