Generative AI - a Costly Process?

Generative AI - a Costly Process?

I read two articles recently that are eye opening.  If you are interested in AI and how it is progressing, you should be aware of these.  You can read the full versions at these links:

https://www.ihlservices.com/news/analyst-corner/2024/06/cloud-pricing-in-ai-world/

https://www.sequoiacap.com/article/follow-the-gpus-perspective/

There are some industries that will not care too much about the cost of AI, but the potential processing costs of Generative AI could be mind blowing.  The machine learning segment of AI is being deployed now quite readily for things like demand forecasting and I believe will continue to roll out quite happily.  However, once customers start to see the real costs of rolling out their Gen AI pilot projects, I believe it will sober up end user businesses and deployments will take much longer to become reality.  There is a big risk here that many companies won’t pay the premium for Gen AI capabilities because the benefits will be eaten up by the processing costs.  You can get more detailed background in the two articles above.  Long term there may be improved solutions to address the costs, but this will surely slow progress until then.

You can read more about application areas that can be implemented affordably here


Posted by Brian Hume
14th June 2024

Back to The Martec Minute