“Nvidia’s business model could collapse if AMD, Arm and other Nvidia competitors sell higher-quality chips at lower prices,” Chey Tae-won, chairman of the Korea Chamber of Commerce and Industry (KCCI) and SK Group, warned attendees at the recent 47th KCCI Jeju Forum. In his speech, he likened the current AI boom to the California Gold Rush of the mid-1800s.
Choi expects AI chip maker Nvidia to remain the top company by market capitalization for the next three years, just as pickaxes and jeans makers thrived during the height of the gold rush. But “when the money ran out, the people selling pickaxes stopped selling,” Choi was quoted as saying by The Korea Times at the forum. “If we can’t make money, the AI boom could die, just like the gold rush died,” he continued.
Nvidia became the world’s most valuable company last month after forecasting record data center GPU sales for 2023. While the company’s stock has already experienced a market correction, it remains the world’s third most valuable company. With no signs of AI development slowing down, we can expect Nvidia to maintain its lead for the foreseeable future.
But the cost of training the next generation of AI law graduates is rising exponentially: Anthropic CEO Dario Amodei says that AI models being trained today cost $1 billion, with $100 billion models expected to emerge by 2025. The rising cost of AI development has concerned Goldman Sachs, with some of its staff saying AI is over-hyped and questioning whether the huge investment in it will pay off.
We believe that AI is now part of our reality, but the AI race could turn into a bubble that could burst unless companies harness this tool in a profitable way, at which point investment in specialized hardware designed to run these processes could plummet, and Nvidia could be caught in the crossfire.
Still, we don’t think this scenario alone will make the green team give in. After all, AI didn’t exist when Jensen Huang conceived Nvidia. The company can always count on the gaming industry as its main customer base, and its cards will always be put to other uses.
The real threat to Nvidia will be its competitors: AMD makes decent GPUs but is behind on supersampling technology, with some even exploring FSR 3.0.
It’s still a long way from DLSS 3.5, but it’s catching up fast, and Intel is also making inroads in the GPU market with its Intel Arc GPUs and XeSS supersampling technology.
Individuals and institutions will continue to buy Nvidia’s products as long as they offer unparalleled performance, but competitors are not slacking, and with AI companies like Microsoft, Amazon, Google and OpenAI investing in their own hardware research, there may come a time when Nvidia is left behind in AI acceleration.
