Regarding inflation, many operate from the premise "too much money chasing too few goods (and services)".

This doesn't convey the true essence of what is taking place since price increases could happen due to a variety of factors. For example, supply disruptions cause price increases even if the money supply is contracting. At this point, it is the pace of each that creates the results.

Nevertheless, if we want to operate from this premise, we are about to see how the situation in the quote, across the board, is impossible.

AI Makes Monetary Inflation Impossible.

Sam Altman made headlines with his assertion that $7 trillion dollars is required to be invested in the global semiconductor industry to meet the demand presented by AI. Thus stunned a lot of people yet does show the amount of money that is required to feed this beast.

What is amazing is this is simply one piece of the puzzle.

We are racing ahead at a rapid pace regarding technology. There are two components that are essential. At present, we are nowhere near sustained in either compute or energy. Altman was focusing upon the compute aspect of the situation but he didn't even touch energy.

The reality is both go together. There simply is no way to massively increase the compute without a similar move in energy.

How much does this require? Altman didn't offer a guess nor has anyone else that I have seen. What we can conclude is the price take will be large.

For all the benefits of AI, for its realization, compute and energy advancements are necessary on a scale we never saw.

Trillions of Dollars By 2030

If Altman is correct, to achieve AGI by 2030, trillions is required. We are seeing massive amounts of money being invested. Tesla is dropping over a billion this year in chips.

According to Altman, AGI is achieved when it has a noticeable difference on society. In a conversation with Lex Fridman, he said it has to change the world, economics, and accelerate the rate of scientific progress.

When we are talking about an AI world, the most valuable things are energy, compute, and AI data centers. We are quickly running out of data and a shortage of compute is quickly appearing. Both of these will be addressed. Of course, to handle all this, massive energy gains are required.

All of this makes the predictions and forecasts about how much energy is going to be required by 2050 simply laughable. Those are going to be vastly shy of what is the world needs. It makes sense since most did not anticipate the explosion in demand created by these technologies.

How will this all be solved?

At this point, I am not sure anyone knows the answer. What we can state is it will cost a fortune. We are looking at trillions of dollars in investment to have a chance to meet what Altman has stated as the path we are on. Of course, he might be wrong and AGI could be much further out. That is not going to stop the investment that is being made.

When money is sucked in by one area of the economy, it cannot go elsewhere. Thus, in aggregate, money is not going to force prices up. In fact, we could see it pulled out of other areas.

That does not mean prices will drop. We are facing the potential of World War 3. War always disrupts supply chains, often causing increases. There also are localized issues as we see with real estate in the United States. Some areas are facing declines in demand as migration occurs. This is offset in other areas where prices are going up as more people flow in.

These topics are never as simple as they make them out to be. What we can conclude is that AI is likely at the center of a lot of discussions over the Next 5-10 years. It is not something that is going to slow down, the investments amounts are showing us that.

Here is a chart of the H100 sales by NVIDIA. Keep in mind, these chips go for $50K apiece and there is a one year wait time.

Source

We can bet that Nvidia is going to sell a lot more of them along with their latest release, the B200 which are reportedly going for $30K each.

Before long, we are going to be talking about some real money.

---

What Is Hive

Posted Using InLeo Alpha