#

Key takeaways

• Amazon and Nvidia expanded their partnership.

• The plan adds 2 million Nvidia GPUs to Amazon data centers.

• The GPUs are scheduled for 2027 and 2028.

• Nvidia technology will be integrated across AWS.

• Amazon continues to develop its own chips.

What TechCrunch reported

TechCrunch reported on August 26, 2026 that Amazon and Nvidia expanded their partnership. The report says the companies plan to add 2 million Nvidia GPUs to Amazon data centers in 2027 and 2028. It also says the GPUs include Blackwell Ultra, Rubin, and Rubin Ultra.

The report adds that Nvidia technology will be integrated across AWS. At the same time, Amazon continues to develop its own chips. That combination suggests a mixed compute strategy, with both partner technology and in-house development.

What the partnership covers

The source describes a large-scale expansion of AI compute capacity. It does not provide technical deployment details beyond the GPU families named in the report. It also does not explain how the GPUs will be distributed across AWS services or regions.

The key point is the scale. Two million GPUs is a major commitment in the context of cloud infrastructure. The report frames this as an expanded partnership, not a one-off purchase.

Why this matters for AI buyers

For cloud buyers and AI teams, the story is a signal to watch global compute capacity. It may also affect how buyers think about vendor choices and infrastructure pricing. That is an inference from the reported expansion, not a stated outcome.

The report does not say what this means for current AWS customers. It also does not say whether access, pricing, or product availability will change. Readers should treat any operational impact as an assumption unless AWS or Nvidia states it directly.

Technology mix and vendor strategy

The report highlights two parallel paths. Amazon is adding Nvidia GPUs to AWS, while also developing its own chips. That suggests the company is not relying on a single hardware approach.

This matters because AI infrastructure often depends on both supply and strategy. A cloud provider can expand capacity through partnerships while still investing in internal silicon. The source does not say which approach will dominate over time.

Operational considerations

The report gives a clear timeline, but not a rollout plan. It names 2027 and 2028 as the years for the added GPUs. It does not specify milestones, customer access rules, or service-level changes.

That leaves several questions open. Buyers may want to monitor how the expansion affects capacity planning and procurement decisions. They should avoid assuming immediate effects before more details are published.

Morocco relevance

The source reports no Morocco-specific facts. For readers in Morocco, the global lesson is to watch major cloud compute expansions closely. Large infrastructure moves can shape expectations around capacity and vendor options, but local availability should not be assumed.

Bottom line

This report points to a significant expansion in AWS AI infrastructure. Amazon and Nvidia are deepening their partnership, and the plan includes 2 million GPUs across 2027 and 2028. The source also shows that Amazon is still pursuing its own chip strategy.

For readers tracking AI infrastructure, the main takeaway is simple. Global compute supply is still evolving, and major cloud partnerships can change the market conversation quickly. The report does not provide enough detail to predict direct customer impact, so any such conclusion should remain cautious.