The chipmaker is already having to notify its investors of a delay of at least six months in the transition to an advanced manufacturing process of 7 nm - and of a new business approach that will not outright disqualify the use of external chip production as part of some products
Intel has had to publicly and officially admit, on several occasions, that its 10-nanometer manufacturing technology has actually failed and will never meet the targets set for it in performance and production outputs - causing us to stare at the next 7-nanometer rollout, hoping this process will take time Preparing to correct all the hardships and shortcomings of his predecessor.
Unfortunately, the production of tiny and fast chips on a huge scale is not a simple business at all, so now it seems that the transition to the next generation at the blue chip maker is not entirely smooth - the company was forced to declare a defect in one of the many production stages, and its repair will delay commercial availability. First 7 nm chips in at least half a year, to a time frame of late 2022 or early 2023.
Those who followed the course of events in the transition of Intel At 10 nanometers, they probably noticed that actual announcements of postponements and delays usually came shortly before the deadlines set by the company for itself - so the current announcement that arrives relatively long before the target naturally raises concerns and suspicions that the practical rate of production at 7 nanometers It is much slower than what has been revealed and reflected for us. However, in the meantime the company clarifies that the fault in question has been found and addressed, and now no further problems are expected that could slow down the pace of progress drastically - we would rather be careful and skeptical, but also keep our fingers crossed that the new technology will reach its destination in the updated time window.
Intel still has a commitment to bring to market its first 7-nanometer products during 2021 or 2022 at the latest, with the Aurora supercomputer designed to consist of state-of-the-art Ponte Vecchio accelerators. This puts the company in a very challenging position - and may explain why alongside the announcement of rejection we also received a very surprising statement that the company will start considering using third-party chip manufacturing services for some products, as a practical alternative to personal in-house production that has been the default in almost all Situation so far.
The new age of chips, too AMD And at Intel, based on hybrid arrays of small chips in different manufacturing processes that are connected together in an effective unified package with improved output and competitive prices - we received first evidence of this also in the Zen 2 generation and the familiar Chiplets technique of AMD And also as part of Kaby Lake-G processors and Lakefield processors Intel.
Different small chips for different components within the central processing unit will give the greatest flexibility to manufacturers with relatively small compromises on latency in communication between components or their relative operating efficiency, so it is understandable why Intel is willing to consider the idea of using chips not made in the company. Would make technological and economic sense. Especially when it suddenly appears that the jump to 7 nm can also be much faster and smoother than we estimated.