Yes, that is correct.
The investment community is pouring in all the big bucks into the Gen AI Applications space but it is unclear if the profitability will come from there at all.
It is possible that the big money is actually at the bottom – infrastructure layers, which is Cloud service providers and chipset vendors read @AWS, @nvidia, @gcp, where the dollars could eventually gather over time.
It is an interesting fact that the infrastructure players need the bucks too but they are already managing it with their size and clout (Amazon Web Services, Microsoft Azure, Oracle, Google Cloud) – apparently, the top 3 cloud providers are spending upwards of $100 billion a year updating their cloud assets and capabilities.
The chipset vendors for this sector are making hay too (Nvidia just crossed $1 Trillion market cap recently) a spectacular milestone (while competing Intel which was a big player in the data center hardware space still seems to be finding its way).
The infrastructure layer is where the majority of the costs associated with generative AI are incurred. This is because generative AI models require a lot of computing power and data storage. As a result, companies that provide infrastructure for generative AI, such as cloud providers and hardware manufacturers, are likely to be the most profitable in the long run.
Generative AI models require a lot of computing power. Generative AI models are trained on massive datasets, which can require a lot of computing power. This is why cloud providers, such as Amazon Web Services (AWS) and Microsoft Azure, are so important for generative AI. They provide access to powerful computing resources that can be used to train and run generative AI models.
Generative AI models require a lot of data storage. Generative AI models also require a lot of data storage. This is because they need to store the training data, as well as the output of the model. This is why hardware manufacturers, such as NVIDIA and AMD, are so important for generative AI. They provide high-performance GPUs that can be used to accelerate the training and inference of generative AI models.
As the demand for generative AI continues to grow, the companies that provide infrastructure for generative AI are likely to become more profitable. This is because they will be able to charge a premium for their services.
In addition to the infrastructure layer, there are other areas of the generative AI space that are also likely to be profitable in the long run. But there is bound to be crowding out before the sector settles down with sanity.
source