Which AI chip companies have generated over a billion dollars in revenue

AI chips were not among headline generators until Nvidia dedicated its GTC 2016 to ‘Deep Learning’. The industry then realized potential and at least 70+ startups were born based on JP Data’s tracker. Some of them have gone to market and started generating dollars since, whereas others are still trying to get their products to the market. My report was the only report available that focused on both qualitative and quantitative aspect of AI chips in 2016. The first version of the report, that was published prior to the boom, predicted that the market will reach $12 billion by 2025 and the update, that was published after the boom, projected that AI chip market will reach $66 billion by 2025. The primary reason for the large upgrade was the thinking that intelligence will become part of every aspect of compute, and every chip in the future will sport some sort of AI accelerator.

AI chipset market predictions come true

Fast forward to 2022, the prediction seems to be holding. Intelligence has become an integral part of majority, if not all chips. On the data center side, top selling CPUs have been enhanced with AI accelerators. GPUs have become de-facto standards for training and FPGA products have been specifically created for AI acceleration. Same on the edge- ARM based SoCs have been enhanced to incorporate AI acceleration engine.

The question then is – which companies have made most money since then? Which ones have generated over billion dollars in revenue from AI chips? And how are they positioned for the future? Well, here are some companies in our tracking list that have made billions of dollars. The future positioning could be a long topic and I’ll cover that in another blog.

Companies that have monetized

There should be no question that Nvidia is the greatest success story of the latest AI chip boom. The company’s stock literally went from $7.3 on Jan1, 16 to $326 at its peak in November 2021. AI was the big driver for this 46X increase in value. The results are evident in the fact that the company’s data center business, that comprises of its AI products, has become the company’s largest business unit and generated over $10 billion in FY 2022.

Intel is yet another company that has benefitted immensely from the AI chip boom, The company announced that it generated over $1 billion in revenue in 2018 from its Xeon product line and over $3.5 billion in 2019. We are in 2022 now and Intel also has lot more AI chips than it had back then.

Qualcomm is the another prominent chip company that doesn’t get as many headlines for AI but has generated more dollars driven by AI. Company’s Snapdragon chip is used in top end cell phones and features DSP, GPU and NN accelerator for AI. If you attribute a dollar value to this silicon and multiply by number of smartphones shipped by the value, we are looking at well over $10 billion in revenue.

Other companies that have generated over $1 billion in revenue from AI includes Mediatek, AMD and Mobileye. Mediatek goes into second tier phones and sports similar AI acceleration as Qualcomm The company generated $17 billion in revenue in 2021. The company was somewhat late to the party but nevertheless has incorporated several AI related silicon blocks in its chip. AMD was also somewhat late to the party to incorporate AI related hardware but has since added silicon to support AI acceleration. Mobileye is yet another company that has done well in AI chip world. In its FY 2021, the company generated over $1 billion in revenue. The company’s chips are widely used in automotive market.

What about AI chip startups

That leaves out many companies from the list. There are a few companies that are close to $1 billion mark but haven’t quite made it yet. The question then comes, what about the AI chip start-ups that have been generating headlines with the funding and valuations? Well, let’s just say that $1 billion in chip sales is lot of chips. With ASP of say 2000 in Data Center world, thats 500 thousand chips. With ASP of say $50 (common in edge world), that is 200 million chips. So unless an enterprise chip company sells half million chips that go into servers or a whopping 200 million chips that go into smartphone and such, we won’t see a billion dollar AI chip contribution. That is not to say that these companies can not generate billion dollars in revenue. Most of the chip start-ups are moving up in the value chain by introducing higher value products such as cards, workstations, servers, racks and even services that go with the hardware. So we should not be surprised to read a headline from one of the start-ups about reaching $1 billion in revenue. But the fact is that, not all of that will come from chips.

AI chip market has long ways to go

We are only five years into the latest AI chip revolution. There are plenty of use cases and applications for AI and we are just scratching the surface. It would have been impossible to think in 2010 that AI would contribute billions of dollars of new revenue in less than five years. Given the way things are going, it seems that AI revolution is here to stay and so are AI chips.