Nvidia CEO Jensen Huang assured investors that Nvidia will continue to dominate the market for AI training chips despite increasing competition in the industry. Nvidia’s upcoming Blackwell technology is expected to solidify the company’s position as a leader in the field, with the H100 chips being regarded as one of the most sought-after products in the tech industry.
While Nvidia briefly surpassed Microsoft and Apple to become the world’s most valuable company, its market value has since fluctuated due to profits being locked in by investors. However, the company’s strong position in AI training chips is expected to remain stable in the long term.
Nvidia’s advantage lies in its CUDA ecosystem, which has become a standard for developers due to its widespread availability across major cloud providers and PC manufacturers. This has created a loyal customer base that is unlikely to switch to competing platforms.
On the other hand, Micron’s recent forecast for fourth-quarter revenue failed to meet optimistic expectations, causing a decline in the company’s stock price. The demand for high-bandwidth memory chips needed for AI training has been on the rise, but Micron’s forecast was only in line with market expectations, disappointing momentum investors.
High-bandwidth memory is crucial for solving the scalability issue in AI training, as it ensures that costs do not increase linearly with the complexity of the models. Micron’s HBM business is projected to generate significant revenue in the coming years, reflecting the growing importance of memory bandwidth in the AI industry.
The memory wall, caused by a bottleneck in data transfer speeds, has been a major challenge in scaling AI training models. As a result, companies like Micron and SK Hynix have seen strong demand for their high-bandwidth memory chips, with Micron’s HBM business already fully booked through the end of the next calendar year.
Overall, Nvidia’s continued dominance in AI training chips, coupled with the increasing importance of high-bandwidth memory in the industry, suggests that the company is well-positioned for long-term growth in the AI market.
Article Source
https://fortune.com/2024/06/28/nvidia-jensen-huang-micron-ai-training-semiconductor-chips-high-bandwidth-memory/