AI: The Coming Revolution

In this post I share key insights from Coatue's AI: The Coming Revolution Post

AI funding mostly sits in the model layer for now, BUT what's next?

I think highly specialized tools will emerge for each core process of Generative AI infrastructure. End-to-end platforms will not succeed due to the product nuance and user experience requirements needed to create value, change behavior in the enterprise, and create product moats.

Less than 10% of enterprise have adopted Generative AI.

Generative AI is by far the greatest value accelerator out of each major tech advancement so far.

The Generative AI Infra stack is fast emerging. Some MLOps vendors are repurposing their platforms to address the workflow-specific needs of Generative AI. It remains to be seen if non-generative AI MLOps platforms can transition to meeting the needs of Generative AI builders. Personally, I am quite skeptical because of how many nuances exist, and many of these MLOps platforms had overly broad offerings even before Gen AI went mainstream.

More data is coming online, and more models. This means curating and understanding data is more vital than in the past, and tools like Looker and Tableau are not suited to meet the needs of Generative AI builders or the data market as a whole due to not being built with ML and LLMs native in the platforms. Nomic is leading the charge on this front creating the new category of ML powered unstructured data curation.

Will VectorDBs be a true database market with public IPOS or just another fad…. I am on the fence about this.

Data is the moat, not the model.

I agree with this 100%.

Those with first party proprietary data can create and fuel models that create massive business value, and create monopolies in entire categories. 

Interested in investing in the best AI deals at the seed stage? Join Banyan’s syndicate for free just click this link.