Microsoft, Meta, Google & Nvidia In AI Battle For Supremacy

There are many powerful players and enormous wins at stake if gotten right, making it thrilling to see various moves happen simultaneously. Not referring to the richest poker event on the planet, but the global battle to control the future of computing power is the real deal. To compete for billions more in this battle, billions of dollars must be invested. But perhaps billions are undervaluing the project. The CEO of OpenAI, an AI research and development company, Sam Altman, has been meeting with influential members of the capital market to discuss funding $7 trillion, yes, seven trillion, for an ambitious long-term plan to rebuild the semiconductor industry as a whole. But that’s a subject for a different blog post.

Several of the largest IT companies in the world, including Google, Apple, Microsoft, Amazon, and Meta, are engaged in a battle to control key assets, pocket share and, consequently, their destiny by producing their own AI chips. Companies like Intel are also making specialized AI processors. However, they are all contending with Nvidia for supremacy in the semiconductor market, since the company accounts for over 70% of sales of AI chips.

Companies can produce chips internally for less money than they can get them for in the highly competitive market. For instance, The New York Times reports that Nvidia chips are valued at approximately $15,000 per chip, whereas Google spends an average of $2,000 to $3,000 on each of its own.

With the AI chip market expected to do more than double to around $140 billion by 2027, according to Gartner, there is a lot of money at stake. With $52.7 billion allocated, including $39 billion in subsidies for U.S. chip manufacture, the CHIPS and Science Act offers $280 billion in additional financing to support domestic research and semiconductor manufacturing in the country. Additional $130 billion in government subsidies are provided in other regions to support Europe’s and other nations’ advanced technological sovereignty, including Japan, India, China, South Korea, and Taiwan.

It can be difficult to keep up with the news surrounding the chip race. These four reports listed below and many more have surfaced this month alone:

During Intel’s Vision event, the company unveiled two China-only models of its Gaudi 3 AI Accelerator, which will compete with Nvidia’s H100, and touted its Lunar Lake processor, which is expected to be released before the end of the year. With the expansion of capabilities in multiple states, Intel anticipates that its U.S. investments will surpass $100 billion over the next five years.

Samsung Electronics declared intentions to increase its entire semiconductor investment in Texas by over 100%, to approximately $44 billion.

By 2028, TSMC intends to build a third production facility at its proposed site in Phoenix. This is more than a $65 billion investment.

Google said developers may now access its Cloud TPU v5p, one of the few alternatives to Nvidia’s AI hardware. According to Google, the new tensor processing unit, or TPU, can train big language models—the basis for AI chatbots like ChatGPT—nearly three times faster than its TPU v4 predecessor.

Making a significant impact with the billions set out for U.S. chip production will take time; the CHIPS Act stipulates that this must happen by 2030. However, by producing their chips, companies like Google, Microsoft, and Amazon will be less reliant on Taiwanese and Chinese suppliers, who carry risks because of geopolitical unrest. This also holds for Nvidia, which produces the majority of its processors at TSMC facilities in Taiwan. Because of these connections, there is a risk of regional dependence on the supply chain network in the event of a significant interruption in the Taiwan region, which might drastically alter the course of events once more.

Geographically, the high-tech supply chain network is becoming more diverse, with nearshoring and regionalization strategies being the most popular supply network models. These strategies are being used to mitigate the numerous disruptions that have recently impacted the sector and to adapt to the growing number of laws that are impeding technology, particularly in the cutting-edge technological hubs in rival regions.

In a recent report, more than 80% of high-tech industry respondents stated that they are growing their supply chain networks and moving to a China Plus One strategy. Either that or they’re making a whole shift to areas other than China; important replacement sites include the United States, Thailand, Vietnam, Mexico, India, and Malaysia. On the other hand, China is intensifying its efforts to attain technological independence to safeguard its advanced manufacturing capabilities, encompassing GenAI, automotive, and other developing areas.

Over the next five years, the semiconductor industry, which is anticipating historically unparalleled growth, could see more total market opportunities opened up by these growing economies. However, there will be difficulties because expanding the production footprint to new sites would need time, money, and resources. The lack of port infrastructure is causing logistics issues for many nations, and in these rising areas, talent shortages are making manufacturing growth capabilities even more problematic.

In this high-stakes chip battle, who will win? In the upcoming years, will Nvidia be able to hold onto its leading position, or will one of the other tech giants manage to overtake it?

There is a saying that says how you finish a race matters more than how you start it. It doesn’t appear as though this race will ever truly finish. Furthermore, a lot of the rivals will probably make progress, despite how overwhelming the odds may appear. After all, in five or ten years, the world might appear very different. Diversification of the geographic network could be a decisive factor.