After completing the last GTC, more post-event details and statements are now known to the CEO of the world’s most profitable chip company. NVIDIA would still depend on gamers for most of its revenue, especially the 47%, but the volume is reduced because the cards are not sold and companies start to buy more and more complementary products such as A100 GPUs, Quantum systems etc …
Supply, demand, shortage of chips and NVIDIA
Supply continues to exceed demand and although we have already seen that there is a date for the end of it, Huang takes it one step further and brings the date forward by almost a year, which makes sense. surprising:
“I think that over the next year the demand will far exceed the supply. We don’t have a silver bullet to solve supply chain problems. We also have the support of our suppliers. We are fortunate to have multiple sources and our supply chain is diverse because our business is quite large so we have the support of a great ecosystem around us ”
In other words, next year, manufacturers will have the opportunity to create more chips than assemblers and users demand, which will undoubtedly increase the runaway inflation that we are currently experiencing due to excessive investments, including the United States and China.
NVIDIA could open up to other designers
We have to go back to 2013 to remember Huang’s own comments on the issue of intellectual property. And it is said that many of its GPU’s IP addresses would be licensed to third parties, but there is little or nothing of that today.
Instead, it is the CEO himself who, after this GTC 2021 slipped that NVIDIA is open to the idea of creating processors that are not their own as long as they are functioning and are integrated into the software ecosystem for the development of CUDA. Unsurprisingly, Omniverse goes a bit like this as we saw in its presentation, but it has a quid pro quo that won’t happen.
There will be no open source at NVIDIA. And that makes sense, since their software is part of the appeal of their platforms and products, as Huang himself made clear:
Under CUDA is the NVIDIA hardware. There is really nothing from our open source. If anyone wanted to make an app for CUDA or make another chip for CUDA, we’re not fundamentally against it, but nobody ever asked for it. “
Why is NVIDIA claiming it? Because Amazon, Google or Facebook They need custom chips for their tasks, to be faster and more efficient and NVIDIA won’t say no if they knock on your door like they did with AMD or Intel for example, just look at the TPUs.
In the end, NVIDIA secretly opts for the strategy of Intel IDM 2.0 but on a smaller scale, since it’s not a wafer maker, but it’s open because Intel is now up to companies to design under their IP, so we could see some interesting products in no time with their label.
