This week, xAI raised $6 billion in funding, placing the value of the company at $24 billion. It is another leg in the AI race that we are seeing emerge before our eyes.
Whether xAI is able to catch OpenAi, something it has its eyes set on, remains to be seen. What we do know is the AI game is one where tremendous value is being generated.
Ironically, it was Sam Altman, the CEO of OpenAI who made the profound observation that the future currency might be compute. If this is the case, these companies that are spending billions on compute are actually, using his forecast, transforming their assets. In other words, it is really just moving from cash to something like bonds.
By buying the compute, it is now in a position to generate a return, similar to a bond. While there could be some decline in value, the utility of the processors will not happen overnight. That means each of these units will have a number of years of use.
All of this is helping to feed the explosion in LLM advancement. For the moment, the monetization is rather limited. That will be something these companies need to address in the future.
We all need to look at the potential impact.
image generated by Ideogram
There is a big difference between these LLMs and what the next layer is going to produce. In my view, there is no money in the models themselves since they will, ultimately, be commodities.
We can already see how there are a number of companies all generating different ones. The winners are going to be the applications these entities create that are tied to the models.
For a platform like LEO, things are a bit different. We are seeing the emergence of SLMs, which take the LLMs and train them on specific data held by the company or entity developing the SLM. This will be different from the LLM since it will have additional datasets that were added.
This is exactly what LeoAI seeks to do.
It will take a LLM (I presume OpenAI) and then train it on Hive data. We know long form posts along with Threads are fed through. There is also all the data from those activities such as upvotes and payouts.
Naturally, this is not data that will be in the LLM. It is something that will be unique to Leo. To access this, InLeo will be the front end, since it will be developed for this platform. It is also going to be a value add for Leo premium.
The question is what this does for Leo from a valuation point of view.
Here is where pointing out the value of the larger entities comes into play.
OpenAI is sitting somewhere around a $100 billion valuation. xAI, Anthropic, and OpenAI are likely to see higher levels in the future. This is not stopping anytime soon. Much like Leo, the value will eventually come down to what services are tied to the LLMs.
Leo is not likely to see billions, let alone hundreds. Could it see, however, millions based upon this addition?
That is something that I find to be rather plausible.
One of the benefits to this approach is the fact that we are looking at something that can scale. By this I am referring to the services offered.
Whatever LeoAI is capable of when released, that is but a beginning. From what I can gather, with each new training run, the features and services can expand. Not only is more data added, but additional options built in.
This is where the continuing evolution of these models drives higher valuation.
Of course, what would this be without mentioning the feedback loop.
As LeoAI used, more data is generated. This means that each prompt ends up as a feedback to improving LeoAI. That is what enables the development of another generation of features.
I am looking forward to the day when we have personal AI assistants. Could something like LeoAI ever evolve into that? I have no idea about the answer on that. What I do know is we could see Leo ride the wave of progress. Whatever the advancement of the underlying LLM, that is available to LeoAI.
Based upon the progress so far, what will that look like in a couple years? That is where the fun really starts.
LeoAI is another layer that we can see building taking place. The next year should focus upon not only bringing this to market but also add more services to it.
That will drive a great deal of value to the ecosystem.
