Emerj, an Artificial Intelligence research and advisory firm, recently published a multi-part article addressing the advantages and disadvantages of various Artificial Intelligence business models. In Part 1 they define the 5 different business models, 3 product models and 2 service models. In Part 2 they compare these models. Emerj talks about the SaaS AI model as having the advantages of both scalability and near term customer value. Although these types of firms are the least effective in helping their customers actually develop their own AI competencies, this is exactly why they can more easily scale and why the business model is a good one, especially in the shorter term.

Emerj has a specific definition of AI SaaS, AI products and AI platform businesses with the main difference between AI SaaS and AI products being that AI products make use of customer data in the AI models used. Per this definition AI products are still likely to be SaaS businesses, they are of the type that myself and others have written about as being a more challenging model to execute. Emerj defines this challenge in a different way. They define the main difference as a tradeoff between stickiness (low churn) and near-term value/scalability. That says to me that if you have a strong product and figure out some ways to scale you'll have a very defensible business model but it will take some time to get there.

IBM announced on October 8th, 2020 that they would be spinning off their Managed Infrastructure Services business as a seperate public company. A good description of the reasoning and strategy from the IBM perspective are available in their announcement and in their presentation deck on the changes. In general, the analyst comments have been mostly positive both from an investment and a product perspective.

It's clear from this announcement that IBM believes that the hybrid cloud market has significant growth opportunities along with AI software. Increased focus on these areas should certainly help them innovate in these areas. I also think this change in focus reflects the market reality of the areas where they have not been very successful such as with their public cloud services.

I found this Computer Weekly article a good balanced perspective talking about the various areas of the cloud where IBM has and has not been successful.