AI chip market "hundreds of battles"! How long can Nvidia's "Iron Throne" sit?

**Source: **Financial Association

Edit Xiaoxiang

Image source: Generated by Unbounded AI tool

Nvidia has been an early leader in the AI revolution. Now, rivals big and small are trying to close the gap.

Established chip industry heavyweights such as AMD and Intel are spending billions to enhance their artificial intelligence offerings, while a slew of start-ups is attracting investors eager to bet on the next chip giant. Meanwhile, cloud computing companies such as Amazon and Google are developing chips of their own design in an attempt to become bigger players in the field.

As everyone wants to become the second "Nvidia", a picture of "a thousand sails racing, a hundred boats competing for the stream" is being staged in the AI chip market...

The current AI boom began late last year when OpenAI's ChatGPT was born, performing a series of "miracles" in just a few months while firmly capturing the public's imagination. This unprecedented frenzy has led to a surge in investment in chips capable of creating and deploying ChatGPT and other so-called generative AI language model systems.

Nvidia is a pioneer in making such AI chips, thanks in part to its long history of making graphics cards used in video game graphics, which are also increasingly important in AI computing. The latest wave of market frenzy has boosted Nvidia's sales and propelled its valuation soaring above $1 trillion over the past few months.

Huang Renxun, the founder of Nvidia, has said bluntly that artificial intelligence is a revolutionary technology that is as important as personal computers and smartphones. "We are at the starting line of AI, and every industry will be revolutionized and reborn."

The Santa Clara, Calif.-based global semiconductor juggernaut now controls more than 80% of the increasingly lucrative GPU market, a business that is fueling a wave of artificial intelligence, analysts estimate. Of course, as more and more companies are trying to follow behind them to seize the "blue ocean" of the AI chip market, it is not easy for Nvidia to keep this big cake. **

AMD and Intel: Eyeing Pretenders

**AMD is currently seen as Nvidia's closest competitor. For a long time, AMD has been Nvidia's "old rival" in the field of game graphics cards. Today, AMD also has its own AI chip series and has a deep relationship with large data center operators who are eager for computing power. **

Forrest Norrod, AMD's head of data center hardware, said the generative AI boom led by ChatGPT initially took the company by surprise, but he said the industry ultimately wants Nvidia to have a competitor. "There are a lot of people who are desperate for alternatives," he noted.

AMD has been planning to produce a new generation of AI chips since last year, and released the AI chip MI300 series in June this year. Compared with the previous generation MI250X, the computing power (TFLOPS) of the MI300 on AI is expected to increase by 8 times. Sophisticated algorithms used to power AI models like ChatGPT.

Northland Capital Markets analyst Gus Richard recently estimated that AMD could reach 20% market share in artificial intelligence chips over time, both because of the strengths of its products and because the world needs a second chip besides Nvidia. supplier.

**In addition to AMD, Intel is also expected to be a spoiler. **The company recently refuted claims that only Nvidia chips can run generative AI. Intel touted its recent AI chip development in a blog post late last month, saying its chips could be a "compelling option for customers looking to break away from closed ecosystems."

** Both AMD and Intel have made huge acquisitions in recent years to bolster the strength of their AI product lines. ** In 2022, AMD spends $35 billion to acquire Xilinx. Xilinx is a company that designs chips that can be reprogrammed after production and are adept at AI computing. Intel, which acquired Israeli artificial intelligence startup Habana Labs for about $2 billion in 2019, is now producing its chips.

In fact, compared to Nvidia, Intel and AMD also have some advantages of their own. They can provide open source alternatives at the software level, which may attract some customers who need it. AMD CEO Su Zifeng said in an interview last month that AMD will provide a "menu" that includes all the components needed to build AI large models. Customers are free to choose the components they need and connect them through industry standards. "We're sure a lot of customers prefer to have options, and they want to be able to customize the components to their data center requirements."

Chip start-ups: newborn calves are not afraid of tigers

**In addition to these traditional chip industry giants, venture capital investors have also placed billion-dollar bets on the chip industry this year, hoping to tap into this emerging market. **AI chip startups Mythic and Tachyum raised new mega-funds this year after several years of funding.

According to Crunchbase, chip startups will attract $8.3 billion in 2021 and $7.9 billion in 2022. Although these figures are not small, they lag behind other popular industries, such as the encryption industry. But this year, AI is almost the only thing in the eyes of many venture capitalists in the technology field.

Nvidia's lead hasn't discouraged newcomers, either.

Graphcore, a British company that has raised $750 million from investors, is developing a more powerful artificial intelligence processor to compete with Nvidia, its chief executive Nigel Toon said.

Toon noted that Graphcore has been particularly successful in Asia. Chinese companies have been looking for alternatives to U.S. suppliers as geopolitical tensions mount and U.S. authorities restrict Nvidia’s sales of its most advanced chips to China.

** "Nvidia needs to share some of their market cap," Toon quipped, referring to the "trillion-dollar club" where Nvidia's market capitalization debuted last month. **

Cloud computing giants: cross-border also has its own advantages

**Amazon's cloud computing division, AWS, has also recently been aggressively marketing its own chips as a cheaper alternative to Nvidia's chips, and has attracted big internet companies like Airbnb, ByteDance and Snap. company. **The AI specific chips currently developed by AWS include Inferentia and Trainium, which are designed to provide convenience for developers to run large language models in the cloud.

AWS executive Nafea Bshara said that by using Amazon's chips, ByteDance has saved up to 60% of the cost of deploying artificial intelligence models.

He noted, “Amazon understood the potential of AI years ago. We realized then that we needed to help customers control their computing costs in order to make AI accessible to customers of all sizes and in all industries.”

Although Amazon seems to be lagging behind Microsoft and Google in investment and development in the field of artificial intelligence, Amazon CEO Andy Jassy (Andy Jassy) recently stated that through Bedrock, custom chips and other services such as CodeWhisperer, Amazon will be in the artificial intelligence market. The Admiral has a real advantage.

In terms of Google, it also confirmed in April this year that it would transfer its team responsible for AI chip development to the Google cloud department to enhance the competitiveness of the cloud computing team and catch up with the powerful enemy AWS. Google also claimed in early April that by connecting 4,000 fourth-generation TPUs in series, the supercomputer built would be 1.7 times faster and 1.9 times more efficient than an equivalent machine using the Nvidia A100 GPU.

The most critical "decisive battlefield" in the future of the chip industry

**It is worth noting that the emerging AI boom comes at a time when the semiconductor industry as a whole is suffering from a post-pandemic slowdown in sales of technology products such as smartphones and personal computers. **

According to International Business Strategies (IBS), a chip industry consulting firm, global semiconductor revenue will drop by 9% this year to about $511 billion.

**However, it is widely expected that the demand for artificial intelligence chips will further expand the scale of the industry in the next few years. **IBS had earlier predicted that chip industry revenue would double to about $1.1 trillion by the end of the decade, driven by advances in 5G networks, self-driving cars and other technologies. Now with the blessing of AI, IBS has increased its revenue forecast by another $150 billion, reaching about $1.25 trillion by 2030.

Handel Jones, CEO of IBS, said, “Generative AI is one of the biggest events the semiconductor industry has ever experienced.

As a whole, the AI chip market as a whole is expected to generate around $43 billion in annual sales this year, accounting for about 8% of the chip industry's total sales, according to analysts at Morgan Stanley. And within four years, AI chips are expected to roughly double their share of the industry, reaching $125 billion in revenue.