The debate whether AI is in a bubble or not continues.
In a number of articles, I covered some points of what we are not dealing with a bubble. There is a lot of base layer stuff taking place, such as infrastructure. That could only lead to a bubble if utilization does not occur.
Of course, markets can far outpace human adoption curves. On this, some might have a legitimate point as a case for it being a bubble. I will, however, debunk that in a bit.
Amazon is going deeper into the AI game. This, in my view, harkens against the notion of a bubble.
Amazon To Invest $50B In AI Infrastructure: AI Bubble?
Before getting to the bubble discussion, let us look at what Amazon is doing.
Today, Amazon announced an investment of up to $50 billion to expand AI and supercomputing capabilities for Amazon Web Services (AWS) U.S. government customers. This investment, set to break ground in 2026, will add nearly 1.3 gigawatts of AI and supercomputing capacity across AWS Top Secret, AWS Secret, and AWS GovCloud (US) Regions by building data centers with advanced compute and networking technologies. Federal agencies will gain access to AWS's comprehensive AI services, including Amazon SageMaker AI for model training and customization, Amazon Bedrock for model and agent deployment, Amazon Nova, Anthropic Claude, and leading open-weights foundation models, and AWS Trainium AI chips, as well as NVIDIA AI infrastructure, equipping agencies to develop custom AI solutions, optimize massive datasets, and enhance workforce productivity. These new capabilities will be available to existing and future U.S. government customers across AWS Top Secret, AWS Secret, and GovCloud (US) Regions, strengthening America's AI leadership and giving federal agencies the secure, scalable infrastructure they need for the next era of innovation.
The company is looking to expand the AI infrastructure capabilities of its government cloud services. These are ongoing operations that the US government conducts on Amazon's servers. As noted, federal agencies will gain access to the AI services, logically increasing the capabilities for government employees.
Like most AI services, the goal is to increase efficiency in areas such as decision making. Obviously, governments are not known for this in general. That said, there are certain areas where the focus, especially on the back end, nets results.
With the development of AI touching areas such as healthcare, the government has a keen interest on the services being created. This extends to energy, cybersecurity, and overall network systems improvements.
This is only going to add to the ongoing data center expansion that Amazon is undertaking. The company already surpassed 900 data centers globally.
Why No Bubble
Much of the AI discussion keeps bringing up the same names:
- Amazon
- Meta
- Nvidia
- OpenAI
- Elon Musk Companies
- Alibaba
- Tencent
- Microsoft
In other words, the nexus of this is that only a few companies in the market are behind the major push. There are certainly a lot of start ups but we are not dealing with dozens of pets.com like in the Internet era. There are AI starts ups that are receiving funding. However, when we look at this list, we see some of the biggest names in tech.
Is there some "pull" to the overall market? Probably. But the rest of the companies are lagging. They have a bit of a push but nothing that one would consider crazy with regards to traditional metrics.
The other fact is that, outside of OpenAI (and perhaps Musk's xAI), nothing on the above list is depending upon new financing. These companies are paying for this entirely outside of operations.
I noted in another article that Nvidia is likely to register $100 billion in profits this year. That means a $4 billion investment equates to 2 weeks.
Ponder that for a second.
Looking back that the list, of the 7 American companies (separating Tesla and xAI) only OpenAI could be in jeopardy. xAI is similar in the need to raise capital but, if Musk wanted to, he could fund the next $50 billion.
The point here is we are not going to see these companies collapse. There will be no implosion. Hence, a bubble is difficult.
Could we see a pullback? Certainly. But the bubble activity is not widespread (if it exists) and there is valid development of a brand new technology that is going to alter much of the digital world (at a minimum).