As of November 20, 2026, Level lists 104 live roles that match "chip design". Only 15 of them sit at Builds AI, the top level. This week OpenAI published results from using its models on chip design with AMD. Most open chip design jobs still describe classic hardware work.
What happened
On November 18, 2026, OpenAI published a post on using its models for chip design. The post was also discussed on Hacker News.
OpenAI says it has used its models on its own inference chip, called Jalapeño, since early 2025. It reports an estimated standard-cell area reduction of about 13 percent on the chip's vector unit. On its matrix unit, the reported figure is about 10 percent.
The post also describes work with AMD. According to OpenAI, model rewrites of SystemVerilog code improved power, performance and area on AMD's flagship GPU design by up to 14 percent. OpenAI adds that AMD included some of those changes in a design sent to fabrication.
OpenAI also announced a partnership with Synopsys, a maker of chip design software. The stated goal is to train a model on Synopsys tools.
One caveat matters. These are results reported by OpenAI about its own models. The post is not an independent study, and it says nothing about hiring.
How many chip design jobs are open right now?
As of November 20, 2026, 104 live roles on Level match the search "chip design". They come from 13 companies.
The list is concentrated. NVIDIA has 64 of the 104 roles, which is 62%. Amazon follows with 18.
The other 22 roles are spread across 11 employers. They include Hewlett Packard Enterprise, Axon, OpenAI, Intel, Etched and Cisco. You can compare employers on the NVIDIA, Amazon and OpenAI company pages.
Verification is the most common specialty in the titles. Of the 104 roles, 32 mention verification in the job title. Fresh postings are a minority, since 16 of the roles were published in the last 30 days.
Which chip design roles involve AI?
Most of them involve little AI. Of the 104 roles, 84 are tagged Little AI, which is 81%. Level explains each label on the levels page.
The other 20 roles sit at Uses AI or higher. Of those, 15 are tagged Builds AI. Another 3 are Works on AI, and 2 are Uses AI.
Those 20 roles come from seven employers. NVIDIA has 8 of them and Amazon has 4. OpenAI has 3, Etched has 2, and three other companies have one each.
Read the label with care. A Builds AI tag means the job is centered on AI. It does not always mean the engineer uses AI models to design the chip. Some of these roles design hardware that runs AI.
A few titles do name AI as the design method. OpenAI lists a Software Engineer, AI for Chip Design and a Research Engineer, AI for Chip Design. Anthropic lists a Research Engineer, Chip Design RL.
Others build chips for AI workloads. Cerebras lists a Senior ASIC RTL Design Engineer, AI Hardware. Etched has two matching roles on its company page.
Where are these jobs, and what do they pay?
Location is the surprise. Of the 104 roles, 69 are in Israel, which is 66%. NVIDIA alone lists 53 of those Israel roles. You can browse AI jobs in Israel for the wider picture.
The United States accounts for 23 roles and India for 6. The remaining 6 are in China, Taiwan, Malaysia and Germany.
Remote work is rare in this field. Only 7 of the 104 roles are remote, which is 7%. For other options, see remote AI jobs.
Seniority leans senior. Of the 104 roles, 67 are senior, which is 64%. Another 15 are mid level and 11 are lead. Entry points exist but are few, with 5 internships and 2 junior roles.
Pay data is thin. Only 19 of the 104 roles publish a yearly salary range in US dollars. The median low end of those ranges is $185,000. The median high end is $250,000.
Those medians describe posted ranges, not what people are paid. They cover 18% of the roles, so treat them as a rough guide.
What to take from this
First, read the OpenAI post as a vendor report, not a hiring forecast. It measures chip area and performance. It does not measure jobs.
Second, the open market is still classic hardware work. On Level, 84 of the 104 matching roles are tagged Little AI. Skills such as RTL design, verification and physical design dominate the listings.
Third, the AI side is small and specific. If you want to apply AI to chip design, start with the 20 roles at Uses AI or higher. Check whether each role uses AI as a design tool or builds hardware for AI.
For related reading, see Level's posts on NVIDIA jobs and CUDA jobs.