Nvidia’s recent earnings reveal a surprising fact: two anonymous customers accounted for a staggering 39 percent of its revenue in Q2, leaving industry watchers both impressed and uneasy. According to the SEC filing, “Customer A” contributed 23 percent, while “Customer B” made up 16 percent – up sharply from last year’s 14 and 11 percent combined share, as reported by CNBC.
That level of concentration is a double-edged sword. On the bright side, it underscores just how central Nvidia has become in the AI infrastructure space. The revenue surge is largely driven by growing demand for the company’s data center chips, especially for cutting-edge AI tasks. But depending almost two-fifths of revenue on just two clients carries serious risk – if either customer changes course, the impact would be immediate and severe.
Who might they be? While the filing labels them simply as “direct” customers, industry watchers speculate they could be major cloud partners or system integrators. Analysts point to large cloud providers like Microsoft, Amazon, Google, or Oracle as possible players – though these firms often qualify as “indirect” clients. Others suggest that major ODMs and OEMs such as Foxconn or Quanta, which assemble systems containing Nvidia chips, might fit the bill. Nvidia’s CFO Colette Kress noted that large cloud service providers make up roughly half of its data center revenue, which itself accounts for 88 percent of total company revenue.
In context, Nvidia posted $46.7 billion in Q2 revenue, a 56 percent year-on-year increase, with data center demand driving much of that growth. Still, gross margins have slipped from 75.1 to 72.4 percent, indicating growing cost pressures and competition.
This scenario isn’t unheard of in tech – many companies rely heavily on a handful of key clients. But the pace of Nvidia’s growth, driven by skyrocketing AI demand, amplifies the importance of diversification. While current clients are investing heavily in AI infrastructure, Nvidia must balance its reliance on them with strategies that spread risk – like expanding into new sectors, broadening its hardware lineup, or securing long-term contracts.