Summarize this content to 2000 words in 6 paragraphs in Arabic The WhatsApp messages began on Sunday. Fellow fund managers had spotted social media buzz on the success of China’s DeepSeek AI app. Was this the iceberg that would sink the apparently unsinkable Magnificent Seven US tech Titanic? Nvidia shares took a 16 per cent tumble in just two days, bounced and then stumbled again. The company lost about $600bn of market value. Microsoft held up, but later in the week reacted badly to slower growth in its core cloud business, while Meta reacted well to decent advertising income. Chips stocks so far have fallen a lot, while the software companies have held up.Is this a false alarm, or a chance for alert investors to jump into a lifeboat, while the rest of the passengers are still on the dance floor bouncing to the 80s classic “The only way is up”?It is very tempting to call the end of the technology bubble. History rings warnings. In January 2000, AOL (new media) bought Time Warner (old media) for $182bn. For some this signalled the peak of the technology bubble. The right investment decision then was to sell the lot — all of the newfangled technology, media and telecoms stocks — and reinvest in old-fashioned banks and mining stocks.The first thing investors ought to do whenever stories like this appear is to be sceptical — doubly so when private Chinese companies are involved And today? Almost a quarter of global equity indices are in the Magnificent Seven stocks, so calling the top might be calling the top for passive funds that track these indices too. With bond yields in the US and the UK delivering more than 4.5 per cent — higher than inflation — why take so much risk? I discussed the appeal of pocketing gains on these pages a couple of weeks ago. The attraction just got greater.For those who want a precis of the reams of coverage on the topic, here’s why the DeepSeek story is important in a sentence. The DeepSeek app is said to work as well as those of US rivals, but was developed at a fraction of the cost and uses much less processing power. There are claims and counter claims about how it was built and how much power it needs. What we know is that the DeepSeek R1 app works. Both this app and the expensively created ChatGPT app were asked a fiendishly complex question: name famous Scottish footballers. There was a score draw, not quite nil-nil. On programming, DeepSeek came out marginally ahead.I am exaggerating a bit, but in a world where Meta is said to be spending more than $60bn a year in capital expenditure costs, this is the tech equivalent of making a working rocket with toilet roll tubes and sticky-back plastic. Necessity is the mother of invention and US restrictions on access to chips seem to have made the Chinese tech wizards particularly inventive.It is interesting that the DeepSeek company’s founder is also co-founder of a hedge fund built on algorithmic trading technology. In the 2000 tech boom financial traders, known as “flash boys” led innovation in using fast data connections. Similar circumstances.The first thing investors ought to do whenever stories like this appear is to be sceptical — doubly so when private Chinese companies are involved. How much aid has the Chinese government given, for example? DeepSeek had a stockpile of older more powerful chips secured from before the sanctions were applied. How important were they?At the same time, Nvidia is booking orders for its next-generation Blackwell chips — OpenAI says these will allow it to stay a step ahead in functionality. It also has good prospects in the robotics arena a few years hence.But Nvidia is a tricky stock to value. Though some retail investors seemed to have seen this week’s fall as a buying opportunity, many other shareholders clearly deemed it a good time to pocket gains.Hard questions will now be asked of the “hyperscalers” — Microsoft, Meta, Alphabet (Google) and Amazon — which have been in a race to lead in AI. They spent around $240bn last year (marginally more than the NHS budget) on capital investment. Shareholders will be concerned about how this capital has been allocated/spaffed.OpenAI says DeepSeek may have taken a free ride on its work, but insists it is still ahead in the race. Microsoft suggests cheaper AI means lots more of it and so that will be good for them overall. However, both companies’ chief executives are up to their ears in past capital investment that they will need to justify. You might feel that “they would say this, wouldn’t they”. But running to cash and bonds is not your only option if you are reducing some of your exposure to AI risk. Many equity analysts have noted the relatively poor performance of shares in European and Asian companies, as well as US smaller companies. I see similarities with 2000, when old-fashioned stocks looked cheap. It might soon be catch-up time.A new source of inexpensive AI might accelerate the spread of this new technology, boosting lots more companies — new and old. As long as the platform remains open to developers, applications will emerge for a range of sectors and services. If cheaper AI leads to more AI this should be particularly good for AI consultants, such as Accenture and IBM (whose recent results were excellent). And also for the cyber security companies that defend your AI data. These smaller stocks performed well through this week’s turbulence. So investors are moving to the next wave and not selling everything. While the DeepSeek story was beginning to take root, President Trump was basking in the glory of his inauguration. The main technology announcement he made was the launch of Stargate — a good Hollywood name for a project. SoftBank, OpenAI, Oracle and others have pledged to back the $500bn project to build AI infrastructure in the US. Details were scant, but the participants no longer look like the visionaries they claimed to be. We knew that the second Trump presidency would throw unexpected challenges at investors. DeepSeek should be a reminder that the rest of the world can produce some surprises in return. But if those surprises are innovations that can make the world a better place I am not complaining. In 2000, investors did well to sell all tech stocks. There are similarities to today, but they should not be overstated. The turn of the millennium was a moment of excess greed among investors. That seems more evident in the bitcoin community — at least in equities we have valuations, however stretched. The shares of the market leaders of 2000 — Vodafone, Nokia, Time Warner — were on daft valuations, far beyond the Magnificent Seven’s today.I was always taught that stocks that go up like a rocket tend to fall like a stick. But I was also told to run my winners. Not helpful, I know! I will say this: if you think we are nearing the top for the Magnificent Seven you can start moving money elsewhere. Better to leave early than late. The good news is that there are plenty of other places you can put it.Simon Edelsten is chair of the investment committee at Goshawk Asset Management, which owns shares in Nvidia, Broadcom, Microsoft, Meta, Alphabet, Amazon, Accenture, Oracle and IBM
