If an investor had been told 20 years ago that microprocessors would be the hottest sector in 2024, he would have bet on Intel. A mistake. The dominator of yesteryear has been left behind in the race of artificial intelligence, the new goose that lays the golden eggs. Intel wants to change the chip, but it will cost time and money to recover from the delay. It has redoubled its commitment to manufacturing with multimillion-dollar investments, while developing new artificial intelligence processors. The current CEO, Pat Gelsinger, has been asking for patience since taking office in 2021, but the market is tired of waiting. In 2004, Intel was worth $148 billion on the stock market, more than 30 times what Nvidia, which was around 4,000 million. In fact, Intel's annual profit then almost doubled Nvidia's capitalization. Since then, its neighbor in Santa Clara (California) has multiplied its capitalization by more than 500 and is worth 2.2 trillion dollars, while Intel is around 135 billion, 10% less than two decades ago. Yes, it's true: Nvidia is the great winner of the era of artificial intelligence and any comparison with it is somewhat unfair. However, in that same period Advanced Micro Devices (AMD) has multiplied its value by 30, TSMC has multiplied its value by 15, and Avago-Broadcom, by more than 20 in 10 years (it was not even listed in 2004). ASML, a manufacturer of lithography machines to produce chips, has multiplied its value by 40. Microprocessors occupy a strategic place in the value chain of leading sectors of almost any technology. However, the undisputed king of the sector is now the ugly duckling. The company had a turnover of 79,024 million dollars in 2021; 63,054 million in 2022 and 54,228 million in 2023, according to their annual reports. This drop in income of 31% in two years has been accompanied by a drop in profit of 91%, from 19,868 to 1,689 million dollars, a lower result than 30 years earlier, when it was savoring the success of its ubiquitous microprocessors. Pentium.Intel rose to the top with an integrated business model in which it designed microprocessors and produced them in its own factories. However, it has been surpassed by specialized firms that are more innovative in the design of integrated circuits (Nvidia, Qualcomm, Arm) or more efficient and advanced in manufacturing (TMSC, main supplier to Apple and Nvidia, and Samsung). Its main production experiences for third parties have been disappointing. While demand for cloud computing or small microprocessors for smartphones grew, the personal computer market, where its microprocessors were ubiquitous, has been limping for years. In addition, Intel has given up quota to AMD and Qualcomm, while losing Apple as a star client when the company led by Tim Cook bets on its own microprocessors. It has been left with a smaller piece of a smaller pie. To turn this situation around, Gelsinger has undertaken a restructuring with which, beyond the short-term adjustment, it wants to recover the time lost both in design and in manufacturing. On the design side, he wants to put a spotlight on artificial intelligence, with the accelerators that the market demands. Last month it assured that its Gaudi 3 artificial intelligence processor will be more powerful than Nvidia's H100, the benchmark to beat. In addition, it has presented new chips for PCs of the era of artificial intelligence, the Core Ultra. His thesis is that as the technology spreads, personal computers, mobile phones and networking equipment will need chips capable of directly executing AI tasks beyond access to the remote data centers of companies like Microsoft and Google where Now it is concentrated. As for manufacturing, the so-called foundry, Intel is undertaking multimillion-dollar investments in the United States, Europe and Israel, in the heat of public incentives, especially those of the Joe Biden Government, to reduce dependence on Asian supplies. In the United States alone, Intel plans investments of more than $100 billion in five years to expand the country's advanced chip manufacturing capacity, especially to meet the demand generated by artificial intelligence. The factor plays in Intel's favor. geostrategic and customers' desire for alternatives to Nvidia. Against, in addition to the accumulated delay, the new competition from large technology companies, such as Amazon, Microsoft and Google, which develop their own microprocessors, and the resistance of designers to contract the foundry to a competitor instead of a pure manufacturer, such as TMSC. .“We are executing our strategy to significantly improve profitability over time. Obviously, we have not achieved it yet, given the large initial investment we have needed to develop this business. But we always said it was going to be a multi-year plan, and we are on the right track,” Gelsinger said in April in the presentation of the results of the first quarter, which Intel closed with losses of 381 million dollars. “Semiconductors are the currency that will boost the global economy in the coming decades. “We are one of two, maybe three companies in the world that can continue to enable the next generation of chip technologies and the only one that has capacity and R&D in the West, and we will participate in the entire AI market,” Gelsinger added. . He assured that the company has hit rock bottom, but at the same time gave disappointing forecasts for the second quarter. The shares reacted downwards and accumulate a fall of more than 30% in 2024. “We would like to believe that the bottom has been hit, but we have lost count of the times we have heard it,” said analysts at an investment firm. American in a report for clients in which they described Intel's short-term situation as “extremely difficult” and expressed doubts about the medium term. “While we believe they are doing everything they can to try to put things right, it is clear that the company is deeply broken, and it will take years to see the fruits of their (currently exhaustive) labor, with success in their efforts far from assured. «in the midst of execution difficulties and structural headwinds,» they added. Gordon Moore, one of the founders of Intel, predicted in 1965 that the number of transistors on an integrated circuit would double each year for the next 10 years with minimal increase. of the cost. Ten years later, he revised what is known as Moore's Law to set a doubling rate every two years. This exponential evolution has marked the semiconductor sector for decades, challenging unthinkable limits with new materials. Over the past two decades, Intel has done little by its co-founder's law. Now he wants to change the chip. Follow all the information from Economy and Business on Facebook and xor in our weekly newsletter
Subscribe to continue reading
Read without limits_