Is NVIDIA losing its crown: A new leader emerges in the AI market

NVIDIA, which has become the absolute leader of the tech world in recent years, is under unexpected pressure. According to Bloomberg analysts, the company's shares have fallen 15% from their peak in May. Although revenues continue to grow, investors have begun to take a cautious approach to NVIDIA's future. This situation indicates that the balance in the AI market is shifting. This is reported by Techcrunch.com reports .

Currently, investors' main focus is shifting from GPUs to memory chip manufacturers. Specifically, the value of Micron shares has nearly tripled in a short period. This trend shows that the main bottleneck for data centers is no longer computing power, but data transfer speed and memory capacity. NVIDIA may be becoming a victim of the high-tech market it created.

Memory chips — the new source of deficit

The CUDA platform and highly complex GPU devices, which were the foundation of NVIDIA's success, were revolutionary steps at the time. However, the GPU shortage observed last year is gradually easing. Instead, the market is demanding High Bandwidth Memory (HBM) chips. These components, produced by companies like Micron, allow data to move between processors at maximum speed.

Interestingly, there were no technological leaps in the memory chip market like those seen with NVIDIA; demand simply outpaced supply by several times. As a result, the price of DRAM chips on the open market has increased tenfold compared to last year. Conversely, the hourly rental price of NVIDIA H100 GPUs has shown a steady decline after peaking at $3.20 in May.

Increasing competition and market prospects

Another reason for the decline in NVIDIA's value is the push for independence by major tech giants. Google, Amazon, Microsoft, and even OpenAI have begun developing their own custom processors. Although these chips are not as advanced as NVIDIA's products, they serve to lower the overall cost of computing power in the market.

Wayne Nelms, co-founder of Ornn, notes that while market participants can create their own processors, no one can manufacture their own proprietary DRAM memory chips. This could make leaders in memory technology the primary winners of the AI race in the near future. NVIDIA will now have to adapt to a new reality, not just technologically, but economically.

In conclusion, the flow of investment into AI infrastructure has not stopped; its direction has simply changed. The ecosystem created by NVIDIA now serves as a source of revenue for other participants, which will take competition in the tech world to a new level.