The U.S. has struggled with a worker shortage in semiconductor chip making, and even educating people about the field’s existence has proved challenging. In response, Apple and other companies have dedicated considerable money and time to addressing the skills gap and broken pipeline.
Apple began the New Silicon Initiative, a series of grants to tech-focused universities nationwide, to develop more skilled workers in designing and manufacturing chips. The initiative funds education and training in microelectronic circuits and hardware design. Eight universities participate, chosen for their engineering savvy and commitment to scaling up courses in creating integrated circuits.
One participant is Georgia Tech’s School of Electrical and Computer Engineering. ECE School Chair Arijit Raychowdhury spoke to TechRepublic about how Apple’s support has changed the school’s offerings and students’ potential places in the changing field of computer chip engineering and fabrication.
In October, Georgia Tech celebrated the beginning of its NSI involvement, representing an expanded collaboration based on a successful chip tape-out course already offered at the university.
“We’re thrilled to bring the New Silicon Initiative to Georgia Tech, expanding our relationship with its School of Electrical and Computer Engineering,” said Jared Zerbe, director of hardware technologies at Apple, in a press release. “Integrated circuits power countless products and services in every aspect of our world today, and we can’t wait to see how Georgia Tech students will help enable and invent the future.”
The full partnership will kick off in January 2025. Apple engineers will present guest lectures, review projects in several IC design courses, give feedback to students, and participate in mentorships and networking events. Apple also funds teaching assistants. Those mentors can answer students’ questions about what jobs will be available to them once they acquire chip design skills.
A highlight of the program is that the tape-out course offers students the opportunity to not only design their own chip but also have it fabricated and tested for bugs. This allows them to gain experience in revising and troubleshooting in conditions similar to those found in the real world. Graduates of the computer architecture, circuit design, and hardware technology courses at ECE can go on to be integrated circuit design engineers, chip design engineers, and analog designers.
SEE: Apple’s M4 chip powers AI features in upcoming devices.
“There was a huge interest among the students,” said Raychowdhury. “In the first semester, they designed a RISC-V microprocessor with some accelerators — and realize that these are seniors. These are not grad students. These are senior undergraduate students.”
Those designs were manufactured on TSMC’s 65-nanometer process node and shipped back to the students. Then, the students could write test modules for their own chips.
“Apple ended up hiring a bunch of the students from this first inaugural class,” Raychowdhury added.
The success of the initial tape-out class led to Apple getting even more involved in coordinating with the school to meet its workforce needs. Raychowdhury said the school has had similar arrangements with companies like Texas Instruments, GlobalFoundries, and Absolics.
Otherwise, “it’s very hard to find students who have that kind of expertise” in chip design, he said.
When companies have a hand in the curriculum, some of what would normally be on-the-job training could be done in the classroom. “That reduces the ramp-up time of the students when they join any of these companies,” Raychowdhury added.
Meanwhile, students will see that they are getting skills that lead directly to in-demand jobs.
They have the space to “figure out whether this is something that they’re really passionate about,” said Raychowdhury. “Even in this huge area of semiconductor jobs, what exactly are they interested in? Whether it’s a design, whether it’s working in the fab, whether it’s packaging, and so on.”
One of the components students build in the tape-out class is a RISC-V microprocessor with an accelerator. Designed to solve linear algebra problems faster, this accelerator could be students’ first step into the hot field of designing the hardware behind generative AI. Georgia Tech and Apple’s efforts don’t focus on generative AI unless they pursue it as a more advanced research project.
“There are some advanced research topics — they are not in a classroom setting yet — where students are actually pursuing ways to use AI, particularly language models, to design chips, including writing RTL,” Raychowdhury said. “That is one area which is growing in popularity.”
Georgia Tech’s Professor Sung-Kyu Lim is working on using AI to accelerate backend processes for chip design, such as layout generating and routing, to reduce the time to market. Some graduate students have the opportunity to work collaboratively on that project.
At Georgia Tech, up-and-coming engineers can work with technologies similar to the advanced manufacturing and processing tools they would use in everyday life as a chip designer. Georgia Tech’s AI maker space, launched in collaboration with NVIDIA, gives students access to H100 and H200 GPUs. That, in turn, gives them more processing power to figure out difficult chip design problems.
Ultimately, the plan is to produce enough skilled workers to cross the skills gap. McKinsey found in 2024 that the number of people working in the semiconductor manufacturing workforce in the U.S. has dropped 43% from its peak in 2000. The country may need 88,000 semiconductor engineers by 2029, but only about 1,000 new technicians join the workforce yearly.
As Raychowdhury explained: “We need a lot more engineers who can work in the fab, who can work in design, who can work in testing.”
