Intel is still having a hard time realizing the ambitions of its leader, Patrick Gelsinger.

Intel’s quarterly report will be released this week, but Reuters decided to take stock of Patrick Gelsinger’s 3.5-year tenure as CEO now. Just two months after taking office, he announced ambitious plans to build new enterprises, later setting a goal of regaining technological leadership by 2025, but so far Intel is struggling to achieve these ambitions.

Image Source: Intel

At least, this is the opinion of Reuters based on communications with several dozen former and current Intel employees at various levels. As the source explains, back in 2021, TSMC offered this company to become its privileged client, able to receive advanced chips at a discount of up to 40%. However, Gelsinger, with his characteristic energy, relied on building his own enterprises and competing with TSMC in the field of lithography. As a result, TSMC management somewhat cooled towards the potential client and canceled significant discounts, and therefore Intel, inevitably faced with the need to order chips from it, is now paying for them at the full rate. Gelsinger’s pride, as noted, ultimately cost Intel dearly. In public, both companies now speak respectfully about their relationship, without commenting on rumors of alleged difficulties in past negotiations.

According to Reuters, Gelsinger’s expectations regarding the pace of development of new technological processes and attracting large clients to them are also not being met. At least at this stage, Intel, although it declares the rapid advancement of its advanced 18A technology among third-party chip developers, has still not acquired large clients, and supplies smaller ones with the necessary tools to work with long delays. As a result, it will not see significant volumes of orders for the production of third-party chips using Intel 18A technology before 2026.

Intel was unable to attract Apple and Qualcomm to use this technical process. An attempt to cooperate with Broadcom also did not take place, as sources explain, since the yield level of suitable products turned out to be no higher than 20%, which is worse than TSMC’s performance at the stage of creating early prototypes. In general, potential Intel customers are not ready to give up the services of TSMC, which they are very happy with, for the sake of unpredictable cooperation with the company headed by Gelsinger. Publicly, Intel continues to deny that there are any problems with the development of the 18A process technology and insists that in 2025 it will allow it to regain technological leadership in the field of lithography.

Reuters also reports on the failed cooperation between Intel and Alphabet (Google), which was considering the possibility of ordering from this company the production of chips for its Waymo robotic taxis. The restructuring of its own assets prevented Intel from bringing the deal to fruition, and Alphabet eventually even received a certain amount of “compensation” from it.

Intel’s financial performance is depressing for investors even before the publication of its third-quarter report. The company’s revenue has fallen by about a third during Gelsinger’s tenure as CEO, to just under $54 billion last year. This year, it could post a net loss of $3.68 billion, the first time since 1986. Intel’s stock price dropped threefold during Gelsinger’s leadership.

At the same time, Intel finds support in the US government in its desire to revive advanced chip production in the country. The government is ready to encourage large American companies to order their chips for production from Intel, but it will not be limited to administrative measures in this area. Intel also managed to secure the largest financial support under the “Chip Act” in the amount of $11 billion, which should be used to build new enterprises on American soil, but it is forced to reduce its staff and abandon the construction of enterprises in Europe for an indefinite period.

Intel was unable to gain a foothold in the production of computational accelerators for artificial intelligence systems, although as of 2019 it had three promising families of accelerators: one of them it could develop on the basis of independent developments, and the other two were obtained as a result of the acquisition of third-party developers . Only Gaudi accelerators have more or less advanced to the market, but the company is forced to offer them at low prices, while ordering chips from the production of rival TSMC. Gelsinger himself, as Reuters notes, continually overestimated the revenue targets for the sale of such accelerators, while in reality they were not in proportional demand.