Amazon has announced that it will spend $200 billion on capital expenditure this year, a spending programme which exceeds current investments by even Google and Microsoft. The CEO, Andy Jassy, reveals most of this spending will be AI-based, supporting Amazon’s cloud computing subsidiary - Amazon Web Services (AWS).

This comes as internal fears arise over AWS’s competitiveness in securing corporate AI contracts. Microsoft and Google managed to secure early investment opportunities with large AI players like OpenAI and Anthropic, which Amazon was only able to replicate later on and with smaller contracts.

This is important given AWS’s current dominant position as the world’s largest cloud provider, generating 60% of Amazon’s overall profits. AWS’s dominant position is clearly at threat from the AI boom, which Amazon failed to capitalise on in AI’s critical early phase - analysts already forecast that Microsoft’s more AI-based cloud business will overtake AWS in revenue by 2029.

As such, the spending programme represents Amazon’s attempt at a catch-up. This programme uses a number of parallel strategies: firstly, massive data centre expansion to provide the computing capacity which AI needs; secondly, development of custom chips which reduces reliance on chip providers like Nvidia; and thirdly, development of in-house AI models that can contend with offerings proposed by OpenAI and Anthropic.

What does this mean for the technology sector?

• Cloud providers are essentially engaging in an infrastructure arms race, creating large amounts of demand for data centre construction and energy capacity

  
  

• Strategic partnerships between cloud providers and AI developers are acting as a core distinguisher amongst tech competitors

  
  

• Attempts at creating in-house chips with less reliance on big players like Nvidia demonstrates wider efforts to shift the entire production process in-house such that there is minimal reliance on third parties