Artificial intelligence is taking over more and more areas of life, and many are hard to imagine without it. By entrusting tasks and processes to AI assistants, ordinary people and international corporations forget about risks. One of them is the instability of stock prices associated with AI companies, from which, like the collapse of the “dot-com bubble” a quarter of a century ago, millions of people around the world can suffer. Tengrinews.kz talks about the ups and downs of artificial intelligence in the eyes of investors. How and why AI companies have risen sharply in price At the end of August, shares of processor and memory card developer Nvidia soared, causing the company to immediately rise in price by nine percent, or $440 billion. The reason was the publication of financial statements for the second quarter, which also contained a forecast for the next year. Nvidia's revenue for the three months amounted to $96.2 billion versus the projected $92.2 billion, and earnings per share were $2.2 versus analysts' forecasts of $2.1. The current surge in quotations reversed the trend of the last year, when Nvidia's positive quarterly reports were followed by a stock drawdown - despite good results,
wives in documents. The company's management is optimistic about the near future and expects revenue growth by 70 percent in 2027. Nvidia founder and CEO Jenson Huang noted that demand for products is even higher than expected, but production capacity is limited and cannot service all potential orders. In the future, the situation may only get worse, since Nvidia’s main contract manufacturer is Taiwan’s TSMC, and the authorities of the unrecognized island are afraid of China’s military operation. If the forecasts come true, the global chip industry and the industries that depend on it will face a massive crisis. Nvidia has other potential problems as well. Analysts point out that the recently announced Jalapeno chip from OpenAI (creator of ChatGPT), developed using new technology for more precise training of artificial intelligence, and its analogs could shake Nvidia's position in the AI chip market, which until now was considered a near monopoly. OpenAI itself claims that Jalapeno is better than its competitors at adapting to the needs of a particular user and processing more of the information he needs, while consuming less energy. More importantly, when creating it, use
called its own developments, which the company carefully protects, which means it will be difficult to repeat it. The dynamics of Nvidia's quotes have been considered for quite some time in the context of what is happening in the AI industry. The company, which initially specialized exclusively in chips, graphics processors and video cards, has been actively developing artificial intelligence in recent years. At the end of the last decade, Nvidia introduced a mini-computer with support for AI technologies and began to use parallel computing (breaking a task into several parts with parallel calculations on different processors) to solve artificial intelligence problems. In 2021, Nvidia launched the most powerful supercomputer at that time, capable of performing AI operations. The company is also actively investing in the AI sector, buying large stakes in cloud providers, laboratories and infrastructure developers. The portfolio already includes shares of OpenAI, which can now become Nvidia’s main rival, and Nebius (a European company that received part of Yandex’s assets). Over the past year alone, the total amount of investments in shares of AI companies has grown 10 times and reached $99 billion, which
Nvidia is one of the main investors in the industry. Who else did this affect? The explosive growth of Nvidia shares pulled with it the securities of other companies whose activities are directly related to both chips and AI, primarily the American software developer and semiconductor manufacturer Broadcom, the famous Intel and the same Nebius. But volatility appeared again in mid-September. On September 14, shares of AI-related companies fell sharply after the heads of several leading developers called for slowing down the development of the most powerful systems due to risks associated with the safety of artificial intelligence. Nvidia shares lost about 3.4 percent, and the index of major semiconductor manufacturers fell nearly six percent. AI is experiencing not only ups, but also downs. Two months earlier, the opposite picture was observed on the global stock market. Then other companies were in the spotlight, although almost all of them were also related to the production of chips and artificial intelligence: The first to come under attack was the Seoul Stock Exchange, whose main index, the Kospi, sank by almost 10 percent in one trading session on June 23, which is why trading even had to be suspended
pour. The leaders in decline were Samsung and SK Hynix. Next, American technology companies fell in price, especially Micron, which ranks third in the global memory card market, and Elon Musk’s space operator SpaceX, which had recently gone public. Because of the collapse, Musk ceased to be a dollar trillionaire. In mid-September, his fortune was estimated at about $900 billion. Experts then connected everything that was happening with the disappointment of investors in the artificial intelligence industry and doubts about the ability of its participants to recoup multibillion-dollar investments in development. High volatility, that is, ups and downs, has become a feature of the current stage of development of the AI market, says Sergey Kudryashov, partner in the Digital Transformation practice at Strategy Partners, whom we ask to comment on what is happening. “The market is extremely concentrated around the securities of a dozen companies, and within this group there is so-called circular financing: participants along the chain are connected by transactions and largely finance each other’s demand,” the speaker shares his vision. “Add here the gigantic gap between investments in hardware and real revenue
- and it becomes clear why individual securities like Palantir move by five to eight percent per day in the absence of fundamental news." Moreover, even high-profile technological announcements of the level of “announced the achievement of AGI,” notes Sergei Kudryashov, have ceased to move the securities of key players in the sector. In its pure form, the market is guided by expectations, he adds. Whether they will be justified will largely depend on how widely the technology can go beyond the already mastered scenarios – this is what another expert whom we ask thinks so give an assessment,” Vladimir Sveshnikov, founder of startups Stafory and Parodist AI. He believes that there is an intrigue here and it lies in three things: First, will there be in the coming years the same technological leap in robotics and so-called physical AI that occurred several years ago with language models? Huge investments are already being made in self-driving cars, industrial and home robots - from Tesla and Figure AI to Chinese developers and new startups. which remains intriguing is the question of whether new industries will emerge in which artificial intelligence will be used as massively as now
alone in programming? There are already some successful projects, for example, in the legal field, but software development remains the most obvious example of the large-scale use of AI. According to Sveshnikov, if there are several such industries, this will greatly change the market’s attitude towards current assessments of AI companies. The third and most uncertain thing, according to the expert, is the question of whether there will be an ASI (artificial super intelligence), which will be able not only to write code, solve complex problems or prove mathematical theorems, but also make important and useful discoveries in practice: for example, new drugs, new ways of obtaining energy, discoveries in physics, chemistry, biology. If this happens, the economic impact could be on a completely different scale. Vladimir Sveshnikov considers the first two scenarios quite probable, but the third is not clear. “In AI now there is a constant cycle from wow to disappointment. A new model comes out, it shows something very strong, and everyone thinks that this is the next big breakthrough. Then people start using it on ordinary tasks and quickly find a bunch of errors and limitations,” says the expert. And it is this gap
between expectations and the real capabilities of the technology, according to Sveshnikov, largely explains the sharp movements in quotes. In addition, it is worth considering that in recent months, the construction of data centers, the purchase of equipment, and the provision of new sources of critical energy for this technology have been added to the traditional expense items of AI companies. To achieve this, many developers are agreeing to build mini-nuclear power plants. “Power is becoming an increasingly severe limiter, and it’s not just about hardware: without electricity, new data processing centers are simply impossible to build,” explains Sergei Kudryashov. “The counterbalance is the bet of Chinese companies on open models: they are becoming more and more capable and can be deployed on their own hardware. If the trend continues, consumers will focus on open models hosted in the cloud or on their own infrastructure, and this will limit the growth of income of advanced American laboratories, and This means it can accelerate the decline of the market." All this forces businesses, including large ones, to reconsider their plans and abandon the use of artificial intelligence at least to the same extent. In addition,
New flaws are discovered in AI, making it less attractive compared to live performers. As a result, some companies have begun to bring back employees who were previously hastened to be laid off in favor of artificial intelligence. Geopolitical tensions around Iran, which affects global oil prices, and uncertainty about the policy of the Federal Reserve System (FRS) were also cited as reasons for panic in the markets in June. Constant exchanges of blows between Washington and Tehran (the latter hitting US allies in the region) are delaying the prospect of establishing lasting peace and the full opening of the Strait of Hormuz, the most important transport artery for fuel in the Middle East. By mid-September, geopolitical tensions around Iran continued to weigh on markets and energy prices. Ceasefire agreements have collapsed, and traffic through the Strait of Hormuz remains restricted. Against this background, the price of Brent oil rose above $105 per barrel. As for the increase in the base rate that was expected from the Fed and its new head Kevin Warsh, these expectations were justified: following the meeting on September 15-16, the regulator increased it by 0.25
percentage point - up to 3.75-4 percent. The decision was the first rate increase since 2023. The Fed explained it by continued high inflation and the need to quickly return it to the target level of two percent. And more expensive money could have an impact on the AI sector, making it more expensive for companies to raise funding for development, and reducing investment opportunities for their potential clients. What does this threaten? The June decline and August rise are just examples of noticeable jumps in the market. Fluctuations occur daily and bring investors around the world either profits or losses. The largest investors are traditionally banks, insurers and funds, including pension funds, which rely on millions of ordinary people. In May, a large American investment company, Acadian Asset Management, warned about the danger of capital concentration in one sector, and even more so in a limited number of companies. By her calculations, 17 percent of the global stock market's gain that month came from just two issuers: Micron and SK Hynix. This situation violates the main rule of investment - mandatory diversification, that is, risk sharing (according to the principle “don’t put all your eggs in
one basket"). Moreover, AI affects many sectors of the economy, which until recently could hardly be called related. Among them are energy (powerful data centers require energy), real estate and construction (warehouses for data centers), transport. AI-related companies purchase goods and services from many contractors, and their employees actively spend their growing income. The recent IPO (initial public offering) of SpaceX made 4.4 thousand of the company’s employees millionaires. Condition 400 of them already exceed $100 million. Investors are also closely watching the plans of the largest private AI developers to go public, although the timing of the placement has already been postponed several times. OpenAI, on the contrary, has postponed a possible placement until at least 2027 and is discussing a new round of private financing due to the fact that in the pursuit of profit from the AI boom, investors have been selling off shares of other companies over the past year. logistics, real estate and asset management companies. All of them are united by rumors that preceded the collapse about the ability of AI to replace some of their functions. Experts warn that massive abbr.
Concerns due to AI, which so frighten representatives of more and more new professions, can backfire on the business itself, which hopes to reduce costs in this way. A sharp increase in unemployment will lead to a decline in consumer activity: laid-off workers will simply have nothing to spend on, and budget funds that could go to soft loans and other subsidies to companies will have to be spent on benefits. Could the AI market become a bubble? Sergey Kudryashov, partner in the Digital Transformation practice at Strategy Partners, is confident that no one has any questions that this is a formed “bubble”. The only question that remains open is how and when it will burst. “The benchmark here is 2027-2029: the built infrastructure will begin to become obsolete, new large investments in modernization will be required, and revenue may not have time to reach the required volumes,” the expert predicts. “At the same time, the fork in the road is not whether the bubble will burst or not, but which layer will suffer: companies in the periphery operate mainly on borrowed funds and, if sentiment turns, they will be the first candidates for a fall, then as a core it is noticeably more stable." To this the speaker adds, h
then skepticism now outweighs enthusiasm - this moment has already passed. But at the same time, Kudryashov makes a reservation: there are no obvious signs that pessimism has been converted into real positions. According to market participants, hedges of about a percent of the portfolio are often against the sector, rather than short positions. Skepticism remains an insurance policy, not a bet on collapse, he adds. “The market is held back by the performance of leading laboratories: revenue volume is still insufficient, but the growth rate is very high - Anthropic’s annual revenue rate increased from nine billion dollars at the end of last year to forty-seven by May,” says Kudryashov. That’s why he believes that everything will become clearer by 2027-2029 - then it will become clear how the situation is with aging equipment, energy shortages and maintaining revenue growth rates: “For now, this is a “bubble” without a calendar: everyone is discussing the collapse, but no one will name its date - that’s why the market is trading in expectations, not facts.” The founder of startups Stafory and Parodist AI, Vladimir Sveshnikov, does not think that AI will turn out to be a bubble in the classical sense. “I think the industry will find enough ways to make money and eventually justify
a significant part of the money that is now being invested there. There may well be a correction in ratings, and a strong one at that. But even if a “bubble” bursts somewhere, I don’t think it will lead to a disaster,” he believes. History knows examples. A quarter of a century ago, the world already experienced a similar shock, called the “dot-com bubble.” Dot-coms were any companies whose activities were related to the Internet, which was then gaining popularity. Over the course of several months, they rose in price and attracted more and more new investors, confident that soon no ordinary process would be possible without the Internet. Feature Internet startups of that time was that they did not actually make a profit and did not even invest in development, spending most of the money they raised on marketing. In March 2000, the “bubble” burst: at one point there were no buyers on the market willing to pay even more for shares of Internet companies, and quotes sharply went down after the NASDAQ Composite index, which unites shares of the world's leading technology companies, fell by 70 percent. for Kazakhstanis to play on the stock exchange and not go broke. Guide for beginners.
investors How can Kazakhstanis invest with the help of professionals and not make a mistake? Guide for beginners What determine gasoline prices in Kazakhstan and whether the cost of oil affects them
