news

after more than a year, the chatgpt craze sparked by openai seems to be cooling down rapidly, and the website's monthly traffic data seems to reflect this.

the report released on thursday emphasized thatgoldman sachspeter oppenheimer, head of macro research and chief global equity strategist at global investment research in europe, cited the latest data from the analysis agency similarweb, showing that the total number of monthly visits to chatgpt websites dropped sharply from spring to mid-summer. the report wrote:

“in terms of monthly users, the initial ‘excitement’ about chatgpt is fading. of course, this doesn’t mean that growth rates in related industries won’t be strong, but it does suggest that the next wave of beneficiaries may come from new products and services that can be created based on these underlying models.”

the plunge in monthly visits doesn’t mean the end of openai. customers may grow tired of gpt-4 or turn to other large language models (llms) backed by tech giants, such as grok from musk’s xai. some users may well find it unnecessary to integrate ai chatbots into their daily lives.

recent news shows that as openai's llm becomes more advanced, chatgpt will be improved, and openai will launch a new high-end chatbot, charging white-collar workers much higher than the $20 per month for gpt-4.

on the same day that oppenheimer's report was released, which was also thursday, technology media the information reported that openai executives discussed new pricing models, with upcoming high-end llms costing up to $2,000 per month, such as a new llm focused on reasoning called "strawberry" and a new flagship llm called orion.

the report cited people familiar with openai’s proposed subscription price as saying that openai may soon charge some users $2,000 per month for using the high-end llm, and that the price has not yet been finalized, suggesting that this makes people "strongly doubt that the final price will be that high."

while the final bill may not reach $2,000, the information points out:

“this is a notable detail because it suggests that even though the paid version of chatgpt, which is expected to generate $2 billion in annual revenue primarily from $20-per-month subscriptions, may not be growing fast enough to cover the huge costs of running the service, including the hundreds of millions of people using the free tier each month.”

what does the fading of chatgpt craze mean? this friday, in a recent report co-authored by oppenheimer, it was pointed out that the fundamentals of the technology industry are strong, but the concentration risk is high. it is recommended to seek diversified investment, which can not only reduce the concentration risk, but also enable investors to enjoy the growth of the technology industry while not missing out on the growth opportunities driven by ai technology in other industries.

the report data shows that although the valuations of top technology companies are not as high as those of top companies in other bubble periods, their market share is the highest in decades, accounting for 27% of the s&p total market value, reflecting an unprecedented high concentration.

the report compiled the average total returns that could be obtained by buying and holding the top ten stocks in 1-10 years since 1980, and found that although the absolute returns of dominant companies are still good, these strong returns will gradually disappear over time, and they tend to remain solid "compound companies." more importantly, if investors buy and hold dominant companies, and other faster-growing companies emerge and outperform the market, the returns of dominant companies will usually turn negative.

taking all the above factors into consideration, goldman sachs believes that:

• the dominant companies are unlikely to be the fastest growing companies over the next decade.

• the stock-specific risk in the s&p index is currently very high, and returns can be increased through diversification.