At NVIDIA, Kshitiz Khatri Is Learning the Business Behind AI

Every day, people type questions into generative AI tools and get answers in seconds. What most users do not see is the enormous amount of computing power those few seconds take and the business decisions around performance, energy and cost that lead to the answers appearing.

At NVIDIA this summer, Kshitiz Khatri (Class of 2027) is getting a close-up look behind the answer box in between his first and second year at The University of Virginia Darden School of Business.

Before coming to Darden, Khatri spent five years working in semiconductors. This summer, he is back in that world, but with a new perspective.

As a product management intern on the NVIDIA’s accelerated computing GPU team, Khatri is working on the business questions behind the technology: how hardware performs, how much power it uses and what it costs for data centers to run the systems supporting today’s AI tools.

Before the internship, Khatri said he assumed electricity would be the biggest cost involved in running a data center. This summer, he has learned that energy is only one part of a much larger picture. The bigger cost structure includes GPUs, networking, cooling, storage, memory and other components that make large-scale AI computing possible.

That broader view has helped Khatri see AI not only as a technology story, but also as a business story shaped by cost, performance and infrastructure decisions.

 The Business Side of AI

While Khatri brought technical experience into the internship, he said his first year at Darden has helped him understand the business side of the work in a much deeper way.

The clearest examples have been accounting and finance. In his role, Khatri works with complex cost models for data centers, including spreadsheets with hundreds of line items, discounted cash flow analysis, depreciation assumptions and financial reports.

Before Darden, Khatri had an engineering background but little formal exposure to accounting. He had not spent much time reading income statements, balance sheets or 10-K reports. This summer, those concepts have become directly relevant to his work.

“Without the finance one and two core courses, I would have not been able to capture the meaning of those cost models,” he said.

The case method has helped in a different way. Khatri said Darden has trained him to take on unfamiliar problems, form a point of view and contribute to a discussion. That habit has been useful in meetings and brainstorming sessions at NVIDIA, where he is often learning new concepts quickly and then applying them in real time.

One example has been tokenomics, or the economics of tokens in AI systems. Khatri was asked to study the topic, review industry reports and share what he learned with teams at NVIDIA. He said the pace and structure of Darden’s case method helped prepare him for that kind of work.

A Clearer Path Forward

The internship has also helped Khatri clarify what he wants to do after Darden.

During his first year, he explored other career paths, including consulting. But returning to semiconductors this summer reminded him how much he enjoys the field and the kind of work he wants to keep doing.

“Coming back to semiconductors, I realized 100% that that’s what I love,” he said.

If given the opportunity, Khatri hopes to continue at NVIDIA. If not, he plans to pursue product management or product marketing roles elsewhere in the semiconductor industry.

Advice for Students Interested in Tech

For other Darden students interested in tech, Khatri’s advice is to start early.

He said students often prioritize networking for consulting and investment banking, but tech recruiting also requires building relationships well before applications open. That is especially true for students interested in large technology companies, where internship recruiting can begin early in the academic year.

His advice: make connections early, keep them warm and learn as much as possible about teams and roles before the process formally begins.

For Khatri, the summer has brought together the technical foundation he had before Darden and the business skills he has built during his first year. At NVIDIA, he is seeing up close how AI infrastructure gets built, evaluated and improved, and how much business thinking guides the technology.