Nvidia's AI Chip Dominance: $10 Billion+ Spent by Each of Three Mystery Customers

Nvidia's Record-Breaking Q3 Revenue

Nvidia, the world's leading supplier of graphics processing units (GPUs) for data centers, reported a staggering $35.1 billion in revenue for its fiscal 2025 third quarter, marking a 94% increase from the previous year 1

1

. This unprecedented growth is primarily driven by the surging demand for AI chips, with the data center segment accounting for $30.8 billion of the total revenue 1

1

.

Three Mystery Customers Dominate Nvidia's Sales

In a surprising revelation, Nvidia's 10-Q filing disclosed that 36% of its Q3 revenue, approximately $12.6 billion, came from just three unidentified customers 1

1

2

2

. Each of these customers, referred to as Customers A, B, and C, contributed 12% to the quarterly revenue 3

3

. Over the first nine months of the fiscal year, two of these customers (B and C) have each spent around $10 billion on Nvidia's technology 3

3

.

Potential Identity of Top Customers

While Nvidia keeps the identity of its top customers confidential, industry experts speculate that they likely include tech giants such as Microsoft, Meta, Amazon, Alphabet, and possibly Super Micro 2

2

. These companies are investing heavily in AI infrastructure, with Morgan Stanley estimating that Microsoft, Amazon, Alphabet, and Meta alone will spend a combined $300 billion on AI infrastructure in 2025 1

1

.

Nvidia's Blackwell Architecture and Pricing

Nvidia recently launched its new Blackwell GPU architecture, offering a significant performance leap over the previous Hopper architecture. The GB200 NVL72 system, based on Blackwell, can perform AI inference 30 times faster than its H100 predecessor 1

1

. However, this cutting-edge technology comes at a premium, with a single GB200 GPU costing over $80,000 1

1

.

Market Dominance and Future Outlook

Nvidia currently holds a dominant position in the AI chip market, with an estimated 90% share in the data center training segment 2

2

. Mandeep Singh, global head of technology research at Bloomberg Intelligence, predicts that the data center training market could reach $1 trillion without any significant pullback 2

2

.

Potential Risks and Challenges

Despite Nvidia's impressive growth, the concentration of revenue from a few key customers presents potential risks:

1. Dependency on large tech companies: Any reduction in AI spending by these major customers could significantly impact Nvidia's revenue 1

   1

   .

2. Supply constraints: Nvidia outsources chip fabrication to Taiwan's TSMC, limiting its ability to rapidly increase production 2

   2

   .

3. Emerging competition: Some hyperscalers, like Alphabet, are developing their own training chips, which could dilute Nvidia's market share in the long term 2

   2

   .

4. Shift to inference: As the AI market matures, there may be a greater focus on inference chips, where Nvidia faces more competition from companies like AMD and custom silicon providers 2

   2

   .

Industry-Wide AI Infrastructure Investments

The demand for AI chips is driving massive investments in data center infrastructure. Oracle, for example, has announced plans to build a zettascale data center with over 131,000 Nvidia Blackwell AI training chips 2

2

. This trend underscores the scale of AI adoption and the critical role Nvidia plays in powering the AI revolution.