Just the other day I was writing how much value is tied into a handful of companies in the S&P 500. This makes the S&P 500 index far riskier than people realize, because 38% of the total value comes from the top 10 companies, which is incredible when you consider how many well-known huge companies are in that entire list.
But, something interesting happened today.
That is about 500 billion wiped off of Nvidia alone.
The reason for this is due the quiet release of DeepSeek, a Chinese LLM that is making some waves in the news, because unlike all of the predecessors, the claims are the latest version took 55 days to train and cost a pittance of 5.6 million, compared to the billions that the others have cost that has warranted them to charge enormous amounts for their usage, and has driven Nvidia stock through the roof in the last couple years, because they make the best specialised processors for the job.
But, those chips aren't available to be sold to China, due to "US national security".
So, if this isn't a Milli Vanilli-esque smoke and mirrors situation, it would appear that a company founded about a year and a half ago, has developed and undercut the big boys by a long way. For instance, the cost of a monthly license for ChatGPT is $30 a month, but DeepSeek is looking at 50c a month for access. That is about 1.6% of the cost. If this is the case, and this is only the first stage of progression, the stranglehold that the US has on the artificial intelligence sector at the moment, isn't just loosened, it is blown apart.
Not because of the cost of access to the consumer, but because all the companies that are currently looking to build use cases and integrate AI into their daily routines, have a far cheaper alternative. And, the companies that have been utilising the current available models from OpenAI and the like, now have an alternative development platform that is far, far, far cheaper. This does a couple things, because lowering the cost gives more companies access to develop, and they can also offer what they develop at a far lower price point, and they don't have to wait long to get an ROI, because the initial costs are so low.
This also opens the floor for far wider experimentation, which will lead to an increase in the innovation, which of course leads to more competition, and increased fracturing of the marketplace. This undermines the established players, and makes room on the shelf for competitors to chip away at the market share, splintering it.
Microsoft was planning on spending something like 70 billion and Meta 50 billion on AI development in 2025 alone, but is that still going to happen, as much of that would have gone to Nvidia. Will they be able to use offers like DeepSeek for their development, or will it be banned by the US government? Or, will they just create their own offering using a similar process as DeepSeek?
The reason that there is so much being spent on AI is because there is so much to be made from it, including control over the masses - the biggest prize perhaps. In a world where information is the most valuable commodity, the ability to gather, collate and now make informed sense from it in moments, makes it a weapon. And as we know about how governments and corporations work, it is a weapons race to keep on ramping up to be both a deterrent, and a tool of aggression and control.
So much conversation is had around AI, but the fact is that the majority of the development isn't being done on things that actually help us as a society, they are spent on ways to control and extract value from us instead. The value of data itself isn't in the data, it is in the scarcity of access to the amounts that those few companies at the top. With their APIs sunk into the internet, they leech off our activity, recording, compiling and selling our digital profile to the highest bidder, and using it to target and manipulate us for whatever cause maximises their profit margins.
Perhaps, if all that information was available for all, and with the help of AI, niche developers who want to actually do something useful, can have the opportunity to develop and innovate also. The global conglomerates aren't interested in solving local problems, because they are looking to make their money by scraping through the billions - but locals are willing to build locally, because they are affected by the problems too.
all technologies will eventually be superseded and replaced, but what is interesting in the tech field is that the lifecycle might be getting increasingly short, because the tools are getting more efficient and effective. The costs come down, more people have access, new versions are created, some die, some thrive, and the evolution continues.
There is obviously a lot of fear mongering that is going to be generated by this, because it is a Chinese company, but there is an inevitability with it all anyway. Eventually, some team is going to find a creative solution to a costly problem, and will make an impact on the market. And then, they will be undercut again. What is interesting with all of this is that the cost of living should be getting cheaper with all this technology, not more expensive. But, that is part of the scheme, isn't it?
One way or another....
User pays.
Taraz
[ Gen1: Hive ]
