Nico Caprez, the son-in-law of NVIDIA CEO Jensen Huang, is becoming a more closely watched figure inside the AI chip company, according to The Information. Caprez recently married Huang’s daughter, Madison Huang, in Hawaii. The report says the bigger story inside NVIDIA is that, less than three years after joining the company, he has already risen to vice president of AI Infrastructure Growth, with responsibility spanning AI infrastructure, neocloud relationships, and financing partnerships.
From selling GPUs to building out a neocloud ecosystem
One of Caprez’s main responsibilities is managing NVIDIA’s ties with neocloud operators. In the report, neocloud refers to a newer class of cloud computing providers built for AI training, inference, and GPU-heavy workloads. Compared with traditional hyperscalers such as Amazon AWS, Microsoft Azure, and Google Cloud, these companies are smaller, but they often depend more heavily on NVIDIA GPUs and tend to move faster in deploying the latest AI accelerators.
That has turned neocloud providers into an important customer group for NVIDIA. The Information says Caprez’s role goes well beyond selling GPUs to them. NVIDIA also invests in some neocloud companies and helps them line up financing, including credit support that can make it easier for them to borrow money to buy NVIDIA hardware.
The relationship is far removed from the traditional role of a chip supplier. NVIDIA’s older model was straightforward: data centers needed GPUs, and NVIDIA made and sold the hardware. As AI data center capital spending has climbed into the tens of billions, and in some cases hundreds of billions of dollars, the main constraint on GPU sales is no longer only chip supply. It is also whether customers have enough capital to build data centers and purchase the hardware.
That has pushed NVIDIA deeper into capital markets. Alongside his work with the neocloud ecosystem, Caprez is involved in other NVIDIA equity investments and in the company’s AI infrastructure projects with major Wall Street financial institutions. The report says he also took part in the related Wall Street partnership project NVIDIA announced in August, valued at $500 billion. Not every deal is led by him alone, but he has become one of the central figures in transactions that cut across chips, finance, and data center infrastructure.
The Firmus deal as a signature example
One of the clearest examples of Caprez’s role is NVIDIA’s deal with Australian neocloud company Firmus, reached in June. According to the report, Firmus planned to finance purchases of more than 170,000 current and next-generation NVIDIA GPUs and CPUs.
For a young neocloud company, borrowing enough to buy 170,000 AI chips is difficult on its own. Banks and creditors would need to assess the company’s credit profile, cash flow, and whether it could actually find enough customers to rent that GPU capacity in the future.
NVIDIA is set to provide what the report calls credit support, a form of backing or guarantee meant to help convince the market that Firmus will be able to absorb and monetize that future compute capacity. That structure could make it easier for Firmus to raise debt in capital markets and potentially lower its financing costs.
NVIDIA stands to collect revenue on two levels in the arrangement. The first is the traditional sale of GPU and CPU hardware. The second is recurring income tied to cloud services. The Information says the Firmus structure could also give NVIDIA a portion of cloud revenue generated when Firmus sells AI compute capacity in the future.
That would extend NVIDIA’s role beyond one-time hardware sales and into long-term cash flow from AI cloud services. Several people interviewed for the report described the Firmus transaction as a representative example of Caprez’s dealmaking ability because it required coordination among NVIDIA, the neocloud operator, creditors, investors, and financial and infrastructure partners across different regions.
Why NVIDIA is willing to take on more risk
For NVIDIA, building a neocloud ecosystem serves another strategic purpose: reducing long-term dependence on the largest cloud companies. Amazon, Google, and Microsoft remain among NVIDIA’s biggest customer groups, but they are also among its biggest potential competitors.
Google has TPU, Amazon has Trainium and Inferentia, and Microsoft continues to invest in in-house AI chips. Those companies still buy large volumes of NVIDIA GPUs today, but their long-term strategy is to reduce reliance on any single chip supplier.
NVIDIA wants to cultivate another group of cloud providers that are deeply tied to its ecosystem, which could soften the effect if hyperscalers gradually shift more workloads to their own chips. Neocloud companies fit that need. They generally do not have the ability to design AI chips themselves, so they rely heavily on NVIDIA GPUs. If NVIDIA also helps them expand through equity investment, financing, and credit support, it is effectively creating more long-term buyers for its hardware.
The financing model is drawing scrutiny
The model is not without controversy. One of the biggest concerns is whether NVIDIA could create an overly dependent or circular capital relationship if it invests in customers, provides credit support, and then enables those same customers to borrow money to buy NVIDIA GPUs.
The report says some larger neocloud companies do not want that kind of arrangement. Nebius, for example, previously declined a similar support structure because it had stronger financing options of its own and did not want to become too financially dependent on NVIDIA.
NVIDIA is also aware of the risk. For smaller neocloud companies, the company is said to worry that some may become too reliant on this kind of support and struggle to operate without NVIDIA-backed financing.
As a result, Caprez and NVIDIA’s Structured Finance and Corporate Development teams are rethinking how these programs and future deal structures should be designed. The next test for Caprez is not only whether he can close more deals. It is whether he can help build a model that can keep scaling without exposing NVIDIA to excessive financial risk.
From Swiss ski racer to corporate development executive
Caprez’s background is unusual for a rising executive in AI infrastructure. He is Swiss and was once a professional skier, competing in high-speed events including Downhill and Super-G, and he represented Switzerland in competition.
He later moved into business and finance. Caprez and Madison Huang both studied at London Business School and graduated in 2021. Madison later joined NVIDIA and now works on product marketing inside the company.
Caprez stayed in Zurich for a period before joining NVIDIA around three years ago. He started in the company’s Corporate Development team, which typically handles investments, mergers and acquisitions, strategic partnerships, capital allocation, and large transactions. That work became the foundation for his later move into AI Infrastructure Growth.
Rapid rise inside NVIDIA
Caprez has not been at NVIDIA long, but his influence has grown quickly. The Information interviewed several people who have worked with him. Some described him as sharp and deeply committed to the job, saying he often works more than 15 hours a day.
Those interviewed also said he quickly grasps the financial details inside complex transactions. That matters in NVIDIA’s current AI infrastructure push because large AI data center deals are no longer simple product sales.
A major AI infrastructure deal can include GPU procurement, data center leases, debt financing, power agreements, equity investment, and revenue sharing at the same time. The capital structures behind these projects are starting to look more like large energy or infrastructure project finance than traditional enterprise sales at a technology company.
People cited in the report also said Caprez remains relatively easy to work with in high-pressure negotiations and sometimes uses humor to ease tension in the deal room. That may be one reason he gained Jensen Huang’s trust so quickly.
Succession speculation has started, but no plan is in evidence
As Caprez and Madison Huang take on more visible roles at NVIDIA, a more sensitive question has started to circulate internally: whether Jensen Huang is gradually preparing the next generation of family members. The Information says some NVIDIA employees have begun to speculate about whether Madison and Caprez could eventually move into higher C-level management roles.
At the same time, the report says there is no evidence that NVIDIA has put any family succession plan in place. It also stresses that NVIDIA is a public company with a board of directors and formal governance procedures, meaning major executive appointments are not something Huang can simply decide on his own. The report adds that succession talk is still early. Huang is 63 and has said publicly that he wants to keep working for a long time, with an attitude described as wanting to stay until the end.
When asked whether Caprez could be seen as a symbol of NVIDIA’s willingness to commit more financing and take on more infrastructure risk, people interviewed for the report said he is indeed skilled at designing creative financing structures and bringing together different partners to get deals done. Even so, Jensen Huang remains the one in charge.
Because Huang has access to a wide range of industry information and speaks directly with key figures across nearly the entire AI supply chain, his view of AI infrastructure demand and future market size still drives NVIDIA’s capital allocation decisions.