The market headlines this week were made by the S&P 500 hitting a new all-time high. As you know, I had expected that new high for some time to give a more complete look to the US uptrend. What went almost completely unreported was the new all-time high for the FTSE 250, the index focussed on the UK domestic economy – the smaller company index was within a whisker of doing the same.
Looking back over the last month the facts on the ground for these UK indices look even better, up just short of 6%, which was double the gain for the much-hyped US S&P 500 index. Our latest “What’s Hot? What’s Not?” analysis confirms this trend…
Of the Hot funds, you will see that five of the top ten were UK funds with a Value style. It is a similar story with the Hot sectors, with a rare hat-trick for the UK stock market sectors which took all three top slots.
Bizarrely, on Thursday the Telegraph chose to ignore all of this and run with the headline “Burnham tax fears spark stock sell-off”, which was self-evidently wrong – more people must have been buying than selling “stocks”, or shares, or these indices and sectors would not have gone up. The basis for their story was that:
“It was the fifth-worst month for equity fund withdrawals in the last 11 years”
This was a mix of retail investors and wealth managers selling funds, just as the UK stock market is gathering momentum and outperforming most global markets. This could be very astute timing, but that is very rare when the investment crowd move at once in a single direction. With hindsight we will know if it was a winning move, but it will have resulted from sheer luck, a random win, rather than any great insight. More often over the last 40 years it would have been correct to see this as a contra-cyclical indicator, and you should do the opposite.
Needless to say it was tech-centric funds which were smashed over the month, with losses in the region of 20%, particularly Asia-based funds.
In the wake of 50% falls in key Korean AI stocks, and a similar collapse in the SpaceX share price from its peak, the AI investment mania continued to unravel this week.
A leftover piece of a SpaceX rocket — launched way back in early 2025 and floating in space ever since — crashed into the Moon. It wasn't a failure as such, just old junk finally coming down, but the timing was unfortunate. China's space program is closing in fast. Its State rocket company just recovered a reusable booster for the first time, something only SpaceX had really pulled off before.
Same story in AI. Alibaba just released a huge new model, Qwen3.8-Max, that stacks up against the best from Anthropic and OpenAI. That's on the heels of Moonshot AI's Kimi K3, which matched top US models on coding for way less money, and DeepSeek's newest model, which came in almost as good, and 40% cheaper, than OpenAI's cheapest option.
The pattern is clear. Chinese AI companies are catching up fast, spending less to do it, and focused more on building things people can actually use, boosting economy-wide productivity - rather than chasing the giant, expensive dream of building smarter-than-human AI, on which governments are likely to clamp down hard.
AI technology, and its application, is still at an early stage. But it is already improving productivity in engaged businesses, and is fast improving. My personal experience is telling…
Something very noticeable happened in the last 4 weeks. Imagine I have a 400-page document, made up of 50 different small sections. I asked our AI tool to look at one section, change some words (too technical) to plain English, and do a bit of text formatting. In the way it responded, two things changed from a month ago.
Firstly, I no longer have to be pinpoint accurate in my instructions (it interprets in a common-sense fashion), and the outcome is also higher quality. The second change is the really interesting bit...
It observed similar things which could be changed in that section, and made the alterations without referring back to me, though told me what it had done and why - with the option to reverse them of course. It gets better…
It also had a look at the rest of the document for similar adjustments which might be merited - in this case it didn’t act independently, but listed all the possible changes and asked for my instruction. It understood that my original instruction was limited to it making alterations only within a certain section of a much larger document, which is good, but also understood that it would be daft not to check the rest of the document - it understands that it must justify its existence by making productivity and quality improvements.
Now let’s turn to the investment angle. We pay less than £20 per month for this valuable tool – which is rapidly improving. That’s two pints of Guinness a month at London prices. That isn’t a business model which justifies billions of expenditure, and obviously the big companies know this and have a variety of strings to their bows. The problem is that, although use of AI will be huge in every business in years to come, this does not mean it will be hugely profitable for the AI companies. AI will be a utility, utilities do not have fat margins, and competition will be considerable (and already is). Many will fail.
This is an extraordinary technology, but it doesn’t follow that it is an extraordinary investment opportunity.
On which note, the Dynamic Portfolio’s continue to perform rather well – not get rich quick, nor get rich slowly – just get rich steadily. UK, Japan, and World ex-UK are updated this month. Dynamic UK Blended outperformed the FTSE 100 over the 6 months, and since inception in February 2000 has generated +1,393.67% growth, more than 4 times the gain in the FTSE (income reinvested). No secret formula. Do this with my compliments.