The artificial intelligence boom has created an almost bottomless appetite for computing power — and with it, an equally bottomless appetite for the engineers who design the chips that make it possible. Across the semiconductor industry, companies are confronting the same uncomfortable reality: the hardware demands of the AI era are scaling faster than the workforce capable of meeting them.

Trevor Bauer, corporate vice president at AMD and site leader for its Longmont, Colorado campus, has a clear-eyed view of that gap. He sits at the intersection of the company’s engineering ambitions and the talent pipeline that has to sustain them. What he sees is an industry that needs more engineers — but also different ones.

“We’re seeing increasing demand for engineers who understand not just chip design, but also AI frameworks, system optimization, advanced packaging, and how hardware and software co-design can produce world-class products,” Bauer says. “On top of a strong engineering foundation, it’s as critical as ever for engineers to also develop their skills in innovation, adaptivity, collaboration, and communication.”

That profile — technically deep, broadly fluent, and adaptable — reflects how profoundly the work of designing a semiconductor has changed.

What the Work Actually Is

AMD is a fabless company, meaning it designs chips but contracts manufacturing to outside partners. The model is common among leading semiconductor firms, but it has a significant implication for workforce strategy: when you don’t own a fab, your competitive advantage lives entirely in your engineers.

“This enables us to be laser focused on what we do best,” Bauer says. “Designing semiconductors.”

That focus demands engineering work that spans a disorienting range of scales. At the lowest level, engineers are managing atomic interactions inside individual transistors. At the highest, they’re building software toolkits that let customers rapidly deploy AMD-powered systems for robotics, autonomous vehicles, and high-performance computing. In between lies circuit design, firmware development, functional verification, performance validation, and full system demonstration — each a discipline unto itself.

Hardware, Bauer is quick to note, is only half the equation. “In parallel to hardware design, it’s critical to design a robust software ecosystem that enables the seamless use of each product,” he says. “Everything must be meticulously validated before, and again after, manufacturing.”

A Taller Stack

The craft of chip design has also changed substantially over the past decade — and not in the way outsiders might expect. The work hasn’t gotten simpler. It’s gotten more layered. Engineers who once drew transistor-level schematics by hand now work with hardware description languages, high-level synthesis tools, and increasingly, AI-driven frameworks that accept natural language and block diagrams as inputs. Each new layer of abstraction enables larger systems to be built by more engineers — without requiring every one of them to have deep expertise at the levels below.

Bauer describes this as semiconductor engineering evolving to “higher levels of abstraction.” The foundational knowledge still matters — it always will — but it now supports a much taller stack. And that stack keeps growing.

AI is accelerating that evolution while simultaneously raising the performance bar. As AI workloads push computing systems to new extremes, semiconductor teams are navigating challenges that barely existed a few years ago: high-speed optical connectivity, security vulnerabilities amplified by AI and quantum advances, cost pressures, and software tooling demands that grow more complex with every product generation.

Building the Pipeline

AMD isn’t waiting for the right engineers to appear. The company has built an active network of partnerships aimed at developing talent alongside advancing technology.

On the academic side, AMD works with universities to foster research into advanced technologies and cultivate a talent pipeline. Through consortia like IMEC and TIE, and participation in standards bodies, the company works to shape the long-range trajectory of the field. Partnerships with government organizations and national laboratories take on higher-risk, longer-horizon research — the kind of work that may not pay off for years but that underpins the next generation of breakthroughs.

“In each case, AMD benefits tremendously by being part of a broad ecosystem,” Bauer says. “We could not do what we do without our valuable partnerships.”

The Myth in the Machine

Underlying all of it is a misconception Bauer wants to put to rest. For decades, the technology industry has operated on an assumption baked into Moore’s Law: that semiconductors will get faster, cheaper, and more powerful on a predictable schedule, more or less automatically.

They won’t. Not without the right people making it happen.

“There is a misconception that Moore’s Law is a fundamental law, and that it is guaranteed to ‘just happen,’” Bauer says. “The truth is, finding such improvements is more difficult than that. They are achieved only through creative breakthroughs up and down the engineering stack. Behind every successful innovation, there is a tremendous amount of exploration, experimentation, and ideas that never pan out.”

That reality — that progress is earned, not automatic — is precisely what makes the talent question so urgent. And, Bauer would argue, so interesting.

“As an AMD engineer, I feel like I am participating in the most exciting game I’ve ever played,” he says. “There’s strategy required with long-term product planning, competitive prediction, industry evolution, and process development. There are new engineering ‘super-powers’ coming on-line with generative AI and advanced semiconductor integration technologies. The world is changing rapidly, and it’s very exciting to be both a witness to, and a part of, that change.”