Guest Blogger Sanjay Gangal
Sanjay Gangal is a veteran of Electronics Design industry with over 25 years experience. He has previously worked at Mentor Graphics, Meta Software and Sun Microsystems. He has been contributing to EDACafe since 1999. EDACafe Industry Predictions for 2023 – AnsysJanuary 24th, 2023 by Sanjay Gangal
By Rich Goldman, Director Product Marketing, AnsysTaking a moment to look back at 2022 reveals a year full of surprises, a year difficult to predict. 2022 opened as the year of 3D-IC, Bespoke Silicon and AI-enabled IC design and ended as AI astounded us all in the ’person’ of ChatGPT. In between, Intel became a foundry, ARM sued one of their largest customers, and semiconductors went from extreme shortages to glut. (We used to call that boom to bust, or undercapacity to oversupply.) And perhaps most of all, the tragic war in Ukraine affected us all and the global economy in unforeseen ways. 2022 was an unpredictable year. 2023 is setting up to be just as unpredictable, but we can rather confidently project trends from 2022, and make some less likely predictions. Bespoke silicon will explode in 2023, as the 2022 pioneers press their advantage, and 3D-IC becomes available to more systems companies. The advantages of bespoke silicon as delivered in optimized end systems are becoming clear to everyone, driving the technique toward the mainstream. 3D-IC is the enabler, and as foundries make this technology more widely available, bespoke silicon will become more the norm. Classic semiconductor companies will team with systems companies to develop their own bespoke silicon.
2023 will indeed be a year of IC glut, returning us to the boom/bust cycle we left behind in 2008. This bust cycle will be extended by all the capital that governments around the world are plowing into advanced foundries in an effort to secure their semiconductor supply chain domestically. In 2023, governments will make the realization that funding foundries is not enough to accomplish that goal, as they import exotic materials for manufacture, then see their domestically produced semiconductors exported for test, and assembly before being returned in end systems. This will result in government investment as varied as rare earth mining and OSAT companies. The result is likely to be duplicate and suboptimal capabilities around the world, and a new era of semiconductor economics. It’s not at all clear that this era will be kind to semiconductor companies. 2023 will usher in the advent of IC level multiphysics. The shrinking of geometries combined with the stacking of chips on interposers in 3D-IC will require new physics analysis such as thermal and structural, and these physics must be analyzed together (multiphysics), not sequentially (multiple physics). Because of the extreme complexity and interactions, this is one phase of IC design that will dictate the use of only best-in-class analysis tools. Nothing short of this will do. No single company can provide all this, so open, extensible platforms will be key. Systems companies entering bespoke silicon design are already analyzing these physics at the board and system level and will continue to bring these design techniques with them into 3D-IC. 2023 will be the year that we learn to harness AI. AI in chip design will continue to improve the outcome of ICs, and AI applied to systems design will improve our end systems. At least one EDA company will release a new AI tool that will generate an optimal IC design requiring only a very high natural language description. It’ll take us another year to figure out whether this is absolutely revolutionary or just smoke and mirrors. Ultimately, we’ll figure out that combining artificial intelligence and human intelligence will yield the best result. Before that happens, we’ll see dozens of AI IC startups go out of business (who needs 50 of them?) and a small handful become fabulously successful. The visionary Silicon Catalyst will play a big role in the success of new semiconductor startups. Silicon photonics will continue its rise, as data centers transition to 800 MHz and 1 GHz data transfer speeds, speeds that silicon cannot deliver. Companies like Ayar Labs will lead the way with new enabling technology and chiplets/3D-IC will enable the integration of photonics and electronics. Strap in. 2023 is going to be quite a ride. That’s one prediction I have high confidence in making. About Author: As Director of Product Marketing at Ansys, Rich leads a team of product marketers in the areas of electronics, semiconductors, photonics and optical simulation. He is responsible for digital marketing, branding, marketing strategy, product marketing, and content marketing for Ansys’ leading simulation software products. Rich joined Ansys with the acquisition of Lumerical, where he was the head of marketing. Prior to that, he was a Partner at Silicon Catalyst where he coached semiconductor startup executives. Prior to that, he was Chancellor at Global Technology University in Silicon Valley, California. He was Vice President of Corporate Marketing and Strategic Alliances at Synopsys where he led the corporate marketing team and its University Program. Category: EDA Predictions |