Jensen Huang – CEO Nvidia
Photo: David Paul Morris | Bloomberg | Getty Images
According to a report from Senior Writer Tae Kim from Barron’s, Nvidia is currently making up to a staggering 1000% profit on the sale of every H100 Hopper graphics card accelerator.
These profit margins, when converted to dollars, are up to $30,000 for the sale of each High-Performance Computing accelerator. Despite the selling price, it only costs around $3,320 for Nvidia, including the cost per chip and other components used on the PCB.
Even though Nvidia has incredible profit margins, it is not just the physical costs that are included in the making of a product. These include costs like product development and R&D that take a massive chunk of the budget before a product like H100 is manufactured.
Furthermore, Glassdoor estimates the average salary for an electronics hardware engineer at Nvidia is around $202,000 annually. Considering the number of engineers and the hours spent on the development of a graphics processing chip like the H100 Hopper, costs start to add up.
Nvidia is currently the sole manufacturer of the H100 Hopper graphics card accelerator and is expected to make an estimated $31 billion from the recent AI HPC boom. Nvidia has also sold all of the stocks of the H100 Hopper graphics card accelerator, and new stock is not expected before 2024.
With this new profitable sector, Nvidia might shift its attention from the desktop gaming graphics card space since the current RX 40-series graphics cards are not doing too well on the market.
A recent report also claimed Nvidia aims to focus on the mobile sector to make up for the lack of demand in the desktop graphics cards segment. However, no other category can be compared to the AI accelerator market, which is expected to be worth around $170 billion by 2027.
