NVDA的模型无关灵活性和新技术（Vera Rubin）推动需求；预计到2028财年CPU收入将翻倍。

Listen, I think the primary reason why everyone is talking about Nvidia stock today is because the company's share price is up almost 10% after they reported quarterly financial results. And that's a big deal for a company that already had a roughly $5 trillion market capitalization. Roughly $500 billion of shareholder wealth was added the day after the company reported these results.

And so, investors have been curious, why would these companies continue to purchase Nvidia chips at premium prices when they can invest more in their proprietary chips? Well, Nvidia's CFO, Colette Kress, did a great job explaining Nvidia's competitive advantages.

First of all, Nvidia's architecture runs every model. So, it's model agnostic. And their growing share as closed and open model adoptions grow. Nvidia is saying that their system operates well using open models and closed models. So, regardless of which model proliferates, which kind of model proliferates, Nvidia benefits. And Nvidia said that they think both types of models will thrive in the future. There's room for both types of models, open and closed.

They go on to say that they're great at small models and giant models, large or video, automated, progressive, or diffusion, in the cloud or on the edge. Nvidia is great at training, great at inference, great at agentic workloads. One platform, fungible for every model and every workload. And that flexibility is driving a lot of demand for Nvidia's product.

One of more One of the more difficult things lately has been financing for purchasing Nvidia's products. And the fact that Nvidia's products are so fungible makes it a lot easier to finance.

These chips from Nvidia are more fungible, and so it allows a less risky finance equation, which lowers the interest expense lenders would be willing to lend to. And that facilitates more transaction. It creates higher value for Nvidia's products.

The management team also saying that it's upcoming Vera Rubin technology, which is going to be its latest generation technology, is delivering 30x higher throughput per megawatt and 35x lower token cost relative to Grace Blackwall Ultra, which is the most recent technology before Vera Rubin.

Those figures look really high to me. I haven't been able to confirm this through other sources, but if this is true, this is a huge upgrade between the previous cycle. I mean, this is a significant increase in value that Nvidia is providing its customers. And that increase in value is very likely to drive increasing demand for these products.

Nvidia's saying that production shipments of Vera Rubin started earlier this month.

So, Nvidia is well-known for its GPUs, its accelerating GPUs, but their CPU is doing good business. Introduced in 2021, revenue on a trailing 12-month basis exceeded $5 billion. They're in full production of their next-generation Vera CPU. It completes agentic tasks 1.8 times faster on the benchmark and provides five times the bandwidth per watt than any other data center CPU on the market today.

but Nvidia's CPUs are now gaining in advantage and gaining in market share. The management team highlighted this saying that they continue to see demand for approximately $20 billion in total CPUs for the current fiscal year, and their preliminary expectation is for CPU revenue to more than double in fiscal 2028. So, $20 billion in fiscal 2027, their expectation is to more than double in fiscal 2028 to $40 billion.

That's a significant proliferation in CPU sales from Nvidia, which previously that market share was going primarily to AMD and Intel. So, let this be a cautionary note to those of you that are very bullish on AMD and Intel. By the way, the valuations of those two companies are significantly more expensive than the valuation of Nvidia when measuring on a forward price to earnings basis or when measuring using a discounted cash flow valuation model.