Intel unveiled new performance-boosting transistor technology for its 10-nanometer chips at its annual Architecture Day.

The new 10nm SuperFin technology — the name is a play on fin field-effect (FinFET) transistor with up to five-time more metal-insulator-metal (MIM) capacitance, or “SuperMIM” — “enables the largest single, intranode enhancement in Intel’s history, delivering performance improvements comparable to a full-node transition,” wrote Raja Koduri, Intel’s chief architect, in a blog post.

In an interview with Reuters, he said the new transistors technology could boost the performance of Intel’s 10nm processors up to 20%.

While Intel says its Tiger Lake mobile processors, which will be available in the fall, will use SuperFin technology, its long-delayed 10nm data center chips won’t be available until next year. Currently, all of Intel’s high-end Xeon Scalable data center chips use 14nm technology.

Meanwhile, rivals AMD, Marvell, and Ampere all have 7nm chips available today, and 5nm chips are in the pipeline. And it’s worth noting that the world’s largest contract chip manufacturer, Taiwan Semiconductor Manufacturing Co. (TSCM), which designs chips for companies like AMD and Apple, plans to use super-high-density MIM capacitance in its 5nm chips.

Unlike most chipmakers — which contract out their designs to fabs like TSMC — Intel is one the few still designing and manufacturing its own chips. This has allowed Intel’s competition to bring 7nm chips to market as early as 2018, and 5nm chips are expected to arrive this fall, followed by 3nm chips in late 2022.

During Intel’s second-quarter earnings call, CEO Bob Swan acknowledged these shortcomings and assured investors that there was no “fundamental roadblock” to resolving the issues with the company’s 7nm chips. Despite delays, he said the first would ship sometime in late 2022 or early 2023.