NEWS
NEWS

Mobiles, laptops, electric cars... the AI fever inflates the price of everything that runs on a chip

Updated

The price of phones will rise by 34% in the next two years, and that of PCs by a staggering 67%

Inside an electronics store MediaMarkt.
Inside an electronics store MediaMarkt.E.M

If you are an investor and want to bang your head against a wall for a missed opportunity, look at these two values on Wall Street: the American Micron and the South Korean SK Hynix. The first's shares have risen by 728.02% in a year. The second's, by 677.93%.

They are not penny stocks, meaning securities with minimal liquidity that can skyrocket or plummet with minimal speculative movement. Micron was valued at 972 billion euros at the opening of Wall Street on Tuesday. SK Hynix, 950 billion. Both companies almost simultaneously entered the trillion-dollar market capitalization club on May 26 and 27.

Unlike SpaceX, OpenAI, or Anthropic, Micron and SK Hynix are not losing money. Quite the opposite. In the first quarter of the year, Micron had a gross profit margin of 84.5%. In the just-ended quarter, it expects to have reached 86%. In other words: Micron sells its products at seven times what it costs to make them.

Those who buy their products the most are those AI companies that are not profitable. Micron, SK Hynix, and also the South Korean Samsung hold 89.7% of the global DRAM memory chip market, according to the Taiwanese company specialized in supply chain analysis, TrendForce. In turn, DRAM functions as the system's immediate memory: it does not store information permanently but keeps the data that the GPU is using available at high speed.

The GPU microchips for Artificial Intelligence (AI) are designed by the world-renowned Nvidia and manufactured by another giant, the Taiwanese TSMC. But, to function, these chips need DRAM memory chips (or their version for AI known as HBM). During the North American gold rushes of the 19th century, it was said that the only ones making money were not the precious metal prospectors but those selling them shovels to dig or sieves to filter water. Nvidia is the most important manufacturer and seller of the "shovel." But now it turns out that Micron and SK Hynix have joined the club.

Without these companies, there is no AI because Nvidia's microchips cannot process data at high speed.

Neither is there without NAND memory chips, which store data permanently, even when the device is turned off. In this product, the DRAM oligopoly has slightly less power: it only reaches 63.1% of the global market. But if two other companies are added - the American SanDisk and the Japanese Kioxia - they reach over 91%.

Without NAND, there is no AI either because it is impossible to store the data for the DRAM and Nvidia microchips to work with them.

But, without DRAM and NAND, there are no mobile phones, computers, tablets, televisions, consoles, servers, the cloud, 5G telephony, robots, electric vehicles, hybrids, or those with electronic driving aids, such as automatic braking, lane control, or parking sensors. These chips are the cornerstone of all "smart" devices, including washers and refrigerators with sensors and even the cranes at container ports.

And now, all these are being taken by AI companies. The prices of these components have increased by between 50% and 70% in just six months. And production will not start to increase until at least the end of 2027. The reason is that the memory microchip sector is characterized by booms and crashes. None of the major players wants to overinvest to then have a factory running at half capacity.

That is why Apple announced an unexpected and immediate 20% price increase on several models of its MacBook computers and iPad tablets on June 25. A week earlier, the company's CEO, Tim Cook, had compared the price increases of memory chips so far this year to "a once-in-a-century flood." For Cook, who steps down on September 1 after a decade and a half in which he multiplied the company's value by thirteen, "in forty years, I have never seen anything like this in any area." And let's not forget that Cook's specialty is the supply chain.

Apple has been the first major company to acknowledge the blow of the chip shortage. Not in vain, the world's second most valuable company has been the largest buyer of microchips worldwide for at least close to a decade. In 2025, it still managed to hold its ground. But in 2026, it is very likely to lose that position to the AI giant Nvidia.

The shortage of these microchips will cause the price of mobile phones to increase dramatically. Investment banking giant Morgan Stanley estimates that, to maintain its profit margins, Apple will have to charge $270 more for the mobile phones it releases this fall - including the foldable version of the iPhone 19 - compared to last year's models. It's not just that each iPhone contains memory chips worth $50, which will now increase to $200, according to market analysis firm Tech Insights. It is also the increase in phone memory. It only costs Apple $15 to increase a phone's memory from 256 GB to 512 GB, but it charges $200 more for it. In the past, Tim Cook's company could leverage its immense purchasing power to impose draconian conditions on memory chip manufacturers. Now, the AI giants have put an end to that glory period. Apple has to bid like anyone else for a scarce product.

And it not only competes against AI but also against the two largest American car manufacturers. General Motors and Ford have signed long-term chip purchase agreements with Micron. The American trade associations NCTA (cable and television infrastructure) and MDMA (medical equipment manufacturing) have sounded the alarm about the shortage of these technological components, while Cook may have directly addressed Donald Trump. The German digital industry association Bitkom has also echoed these alarms, and the UK's largest consumer electronics store chain, Currys, despite stocking up on mobile phones and computers, has already acknowledged that price increases will be inevitable in the last quarter of the year. According to some analyses in the US, the price of phones will rise by 34% in the next two years, and that of PCs by a staggering 67%.

It is a combination of exorbitant prices and bottlenecks. And a serious problem for governments facing price tensions due to Trump's tariffs and the US and Israel's war against Iran, not wanting to think about a third inflationary wave, this time due to the chip shortage. Another option that attracts companies but worries governments due to its obvious geopolitical risk is to relax the import licensing system for China's memory chips. But this would entail the risk of incorporating chips from the West's biggest strategic rival into their AI models and agents.

And then there is the regulatory component. It has been exactly four years since the US passed the CHIPS Act, a massive industrial policy plan to increase the country's manufacturing capacity in high technology, including, as the name suggests, chips. However, a massive Micron investment in the state of New York of $100 billion (¤86 billion) has been in limbo until this spring when legal hurdles to its construction were resolved due to the presence of an endangered bat colony in the forest to be cleared for the factory. For the same reason, the European Commission's Technological Sovereignty Plan approved last month, which foresaw the development of the microchip industry within the EU, will take years - at least - to have tangible consequences. And then, it is possible that the microchip sector has entered one of its characteristic downturn phases.

What is clear is that if you want to buy a phone or a personal computer this year, you better hurry. Either that, or buy a share of Micron or SK Hynix and hope for the appreciation to continue. Looking at things cynically, it turns out that AI is not only bringing us unemployment but also hyperinflation. This is called stagflation, and the last time the world suffered it was half a century ago during the oil crisis of the eighties. While we await the future benefits of this technology, its short-term costs seem increasingly burdensome.