Most of us tend to be plodders, there are those who embrace the latest technology and know the companies that are leading edge, but most people tend to need a focal point to understand what could be. Sometimes what could be means what is, will change or could easily change dramatically and that has consequences, but it also has many benefits. No one knows until a few years down the line. When Apple came out with the first smartphone, people began to use it differently and the applications on the phone has changed the world to be a better place. If you considered half of developing world now has access to credit, before they only had access to money lenders at high interest rates. There are always benefits to some and many times to many and consequences for industries. If you think about cable companies and the move to streaming on smartphones. This year the defining focal point will be ChatGPT and the uses that are coming. To use ChatGPT, means an infrastructure will have to be in place and the company that is leading that process is Nvidia.

In an article by Don Clark of the New York Times News Service, the infrastructure has taken at least 10 years and it happens Nvidia made most of the correct decisions. Think about other larger chip companies who are not where Nvidia is, and the questions is why Nvidia?

Over the past 10 years, Nvidia has built a nearly impregnable lead in producing chips that can perform complex AI tasks such as image, facial and speech recognition as well as generating texts for chatbots such as ChatGPT. The company achieved dominance by recognizing the AI trend early, tailoring its chips to those tasks and then developing key pieces of software that aid AI development.

Jensen Huang, Nvidia’s CEO has offered customers access to specialized computers, computing services and other tools of their emerging trade. That has turned Nvidia into a one stop shop for AI developers.

According to the research firm Omdia, Nvidia accounts for 70% of AI chips sales and holds an even bigger position in training generative AI models.

Daniel Newman, an analyst at Futurum Group said customers are willing to wait up to 18 months to buy a Nvidia system rather than buy an available off the shelf chip from a startup or another competitor.

Mr. Huang, said we understood that the reinvention of how computing is done and we built everything from the ground up, from the processor all the way to the end.

In 2006, the company announced software technology called CUDA that helped program the GPUs for new tasks, turning them from single purpose chips to more general purpose chips.

In 2012, researchers used GPUs to achieve humanlike accuracy in tasks such as recognizing a cat in an image. Nvidia responded by turning every aspect of the company to advance this field. This effort has cost $30 billion in the past decade.

Besides collaborating with leading scientist and startups, Nvidia built a team that directly participates in AI activities such as creating and training language models. This lead Nvidia to develop many layers of software beyond CUDA which saves labor for programmers.

Nvidia is focused on the large customers and data centers and its chips are relatively expensive the H100 costs between $15,000 and $40,000 which is 2 or 3 times higher than the A100. However, Mr, Huang says if you can reduce the time of training to half on a $5 billion data center, the savings is more than the costs of the chips.

For smaller companies, companies such as Inflection AI and CoreWeave which allows computers to rent time on their computers rather than buying the chips.

Linking to dividend paying stocks, at the moment one of the reasons you want to own Nvidia is because it has the largest market share in chip manufactures for data centers. It is important to remember IBM and Intel had the largest market share in the past. Cycles happen and Nvidia made deliberate choices to concentrate on the promising AI market. That market was seen in the select government institutions but was not in the general public. Eventually the general public will have access to Nvidia chips and new competitors will rise, for now enjoy the ride and ensure Nvidia is continuing to be the choice for AI developers. If not look for alternatives.

There are more questions than answers, till the next time – to raising questions.