Deb Shoquist Leads Operations at NVIDIA

Deb Shoquist leads operations at NVIDIA, overseeing the supply chain and manufacturing behind its AI platforms — and her network of partners builds technology that has never been built before. In a new interview, blogs.nvidia.com details how she turns blueprints into high-volume systems, from TSMC wafers to Foxconn and Wistron assembly, in as little as two weeks when supply and demand requires it.

The story, syndicated this week, landed as Shoquist paused her travels to accept the 2026 Dr. Lisa Su Woman of Innovation Award. Her résumé is the supply chain itself: TSMC has partnered with NVIDIA since 1998, and SPIL and King Yuan Electronics (KYEC) were assembling and testing its chips more than two decades ago.

What does Deb Shoquist actually do?

At the most basic level, our job is to convert designs into the physical world, Deb Shoquist explained, noting that the blueprint comes from the design team and her job is to transition it into a reliable, high-performance supply chain.

Her analogy is matchmaking. A supplier must be financially strong enough to invest in NVIDIA’s ramps, highly technical, and willing to do something that has never been done before. Three legs, every time. If a supplier only has two, you have to invest in the company to make them a partner — which is exactly what NVIDIA did with optics makers Coherent and Lumentum, whose technology and willingness outpaced their scale.

Why do human relationships decide chip volume?

Every three months, Shoquist sits down with her most important suppliers, and she also devotes six weeks annually to traveling to see partner operations across Taiwan, China, Korea, Vietnam and the U.S. Conversations move from meeting rooms to factory floors.

Over 20 years, she says she has learned it from the people running production lines — not just the CEOs. That trust is what lets floor personnel say, “I need better technical support on the floor. I need more quality engineers. We need a better design.”

Deb Shoquist Builds AI Supply Chain at NVIDIA: Blogs.nvidia.com
Photo: mingooland.com

Only then are you in a position to ask them for special favors, she noted, because you can do it surgically, and your ask is credible and reasonable. She recalls a supplier in Taiwan who calls her at 3 a.m. with fixes, and factory workers in Vietnam whose only ask is more tools.

How fast can the chain really move?

The answer is measurable — and the numbers are the sharpest part of the interview:

Supply chain metric Figure from the source
Cable cartridge defect target Below 1 DPPM (defect part per million)
Cable legs per cartridge 1,500

Every chip is tracked from the identification number on the die to the customer invoice. Suppliers who cannot support that traceability process cannot supply at all. Shoquist describes the chain as a circulatory system: when one branch needs acceleration, you have the power to do that. Weekly yield reviews determine which factory gets which chips — low-performing plants simply don’t get more chips that week.

What breaks when things go wrong?

The GB200 compute board ramp is the candid example. Foxconn and Wistron backed him without question, bringing in hundreds of operators to keep production moving.

forck.live
Photo: forck.live

What followed was the “NVIDIA swarm”: specialists from NVIDIA design engineering, operations industrial, and product and test engineering, together with Foxconn automation engineering, streamed onto the line to rework and restart it ahead of the full-scale ramp. Root causes spanned test tray assembly, tray design, automation hardware, control software and product yield.

Deb Shoquist said that they say “the mission is the boss,” but they also live “the mission is the boss.”

Why co-packaged optics rewrite the supply chain?

The transition from chips to systems changed everything. As primarily a chip manufacturer, NVIDIA could pick the best technology matches within an industry at scale. As a data center systems company building co-packaged optics (CPO), it must create entirely new supply chains with scale partners that are new to the optics world.

The CPO example matters beyond NVIDIA’s walls. Optical suppliers possess the expertise and the readiness, yet they still lack the financial resources and manufacturing footprint needed for fabricating new fabs on the necessary scale — a shift coverage of the interview framed as the frontier where supply bases are still maturing.

For enterprise buyers, the practical read is this: NVIDIA’s throughput advantage is not just a chip advantage. It is decades of accumulated manufacturing relationships — including TSMC, Foxconn, Wistron, SPIL, KYEC, Unimicron, Lumentum, Coherent and Amkor — that a competitor cannot replicate on a roadmap slide.

The Amphenol cartridge, in context

Why does one cable assembly deserve attention? Because 1,500 small cable legs per unit, each requiring a very precise assembly to yield a quality level of less than 1 DPPM, meets the strictest volume-sourcing bar NVIDIA sets — and Amphenol built over a hundred automated lines to produce this product at the required speed and scale.

Deb Shoquist said that that takes know-how and technology, but it also takes heart, willingness and dedication. The partners NVIDIA maintains, she notes, are the ones who supported it when it was small, and yet are willing to scale with it today.

Operations at NVIDIA stays a constant motion problem — supply chain changes designed and implemented constantly to meet the strong demand of new products. The challenges are high, but the team is always given the resources needed to be successful.

Deb Shoquist described closing the interview by sharing that the only thing between their goals and their success is their own creativity, adding that it is exhilarating.

Photo of author

Technology Desk

Technology Desk at Archyde.

Andrej Stojaković Scores 46 Points as Illinois Tops Duke