The Bank of England and a few other places are warning that the Ai bubble is in danger of bursting which would sound scary if only I knew what it meant, so i asked someone who should know and came away still not certain if i'm being honest.
From what i could glean, a bubble grows when new investors drive the share price of something up to unsustainable heights and then when they realise that the thing isn't as good or as profitable as they thought, they start selling their shares which causes prices to fall and more investors sell their shares and the whole thing comes crashing down, or in economic terms, the bubble bursts.
OpenAI's boss Sam Altman did say this week that: 'There are many parts of AI that I think are kind of bubbly right now' which doesn't sound great if you have invested in an AI company.
A burst bubble results in massive financial losses for investors who bought into the hype, during the dot-com bust in the early 2000's, £5 trillion was lost and apparently the AI bubble has inflated even bigger, 17 times bigger according to some financial analysts which will make one hell of a pop if it does go with a bang.
More people, such as the people at the Bank of England and the International Monetary Fund are now saying that these AI companies are overvalued and mentioned circular financing with Nvidia investing $100 billion in OpenAI who buy their chips from Nvidia so they are therefore giving OpenAI the money to buy their chips with and the infrastructure needed from another Nvidea company, Coreweave, so the company invests in its own customers so they can continue making purchases which is a headache, especially as OpenAI has never turned a profit but has cost the company a projected loss of $27 billion.
If investors don't see a return for their outlay soon then they could try and get some of their money back and things could go poof pretty quickly and the people who seem to know are saying we have the first signs of a bubble and the pin which could very well pop it.
Thursday, 30 October 2025
Ai Looking A Bit Bubbly
Subscribe to:
Post Comments (Atom)
1 comment:
"bubbles bursting" is an emotion generating headline created by journalists and talking heads that need to sell their opinions (which are really just echos of their readers/listeners/watchers and voters) - very manipulative and self-serving...
it seems very likely that some of the companies are overvalued (because enough investors follow trends instead of looking at the financials, capacity, and consumer adoption). this doesn't mean all of the LLM companies are overvalued. Also, many of the LLMs are owned by trillion-dollar companies...
regarding the dotcom bust circa 2000, if one held their shares, they recovered in 2 years. ditto the banking bust circa 2010. these "bursting bubbles" are better seen as corrections in stock value.
if you bought shares early, and they have doubled or tripled, a 30% correction means you gained much value on the investment. the whiners are the ones that bought at $400/share instead of at $30/share...
the "circular" bit is also bullshit. if you need hardware, you buy it or lease it from someone. if that someone is buying your software services (which they also need) it is more like an interest free loan for both entities. further, it is simply more direct than party a (LLM), buying from party b (servers), that buys from party c (components), that buys from party d (LLM), that buys from party b (servers), that buys from party a (LLM), etc.
in re profits, consider amazon profits. it went at least 10 years before it had a profitable quarter - much less profitable year. as a society becomes more complex, the innovations to solve that society's problems become more complex - and takes longer for the innovation to be developed, made production ready, then be operationally leveraged by consumers (in this case other company - this isn't a consumer tech yet).
i'm told by people that worked where i retired from that my former employer has over 200 efforts in place to automate work (using agenic ai) that could not be automated using conventional IT methods.
points:
- a correction is likely (including some competitors being acquired)
- if your shares are at least 2 years old you probably okay
- if your shares are newly acquired, then get out now or hold them when the correction happens
- it takes at least 5 years for new techs to become profitable
- look at the potential market (every company in US, uk, eu, canada, japan, australia, etc.) and judge the potential value (think trillions of dollars) to determine the potential value of the providers
- look at current consumption (the big corporations are already using the LLMs, many small companies run by younger owners are using LLMs, and many small startups are built on LLM workflows - agenic ai).
- many IT professionals are learning AI/ML to make a transition to AI/ML because they see what their employers are doing (i'm retired and i just completed and Andrew Ng course on agenic ai, to go along with my MS Data Analytics) - this is a good weak signal of future conditions - automating with LLMs instead of pure code...
- i could be wrong and the whole house of cards will collapse...
Post a Comment