Samsung has a great opportunity.

The Samsung HBM4 chip was showcased at the Korea Technology Exhibition in December 2025. Photo: Reuters .

Samsung is close to securing certification from Nvidia for its latest version of the HBM4 AI memory chip. This would be a significant step in its efforts to close the gap with rival SK Hynix.

According to Bloomberg, the Suwon-based company has entered the final stages of evaluation with Nvidia, after providing initial samples to the American chipmaker since September. Nvidia uses massive amounts of high-bandwidth memory (HBM) to power its AI accelerators.

The news agency reported that Samsung is preparing for mass production of the HBM4 in February. The company will be ready to ship the device soon, but no specific timeline has been set.

Samsung currently lags behind SK Hynix and Micron in the race to lead the AI memory market. However, shares of all three companies have surged in recent weeks, as the AI wave has created a global memory chip shortage in the electronics industry.

Since the beginning of September, the total market capitalization of this leading memory chip manufacturing group has increased by approximately $900 billion . Investors are increasingly expecting Samsung to join the group of component suppliers for Nvidia's upcoming flagship Rubin processor line.

HBM4 is a mandatory component for Nvidia's next-generation GPU architecture, codenamed Rubin, expected to launch in late 2025 or early 2026. This chip expands the interface to 2048-bit, double that of HBM3E, allowing for a dramatic increase in data transfer bandwidth without needing excessively high clock speeds, thus saving energy.

Samsung's mass production in February suggests they are racing to meet the deadline for sending samples for the early development phase of the Rubin. If deliveries are made at this time, they could be one step ahead of SK Hynix in providing technical samples for Nvidia's Rubin, creating an opportunity to turn the tables.

To date, Nvidia remains most heavily reliant on SK Hynix for the supply of the most sophisticated memory chips to be used in its high-end AI accelerators. This gives SK Hynix a pivotal position in the global AI supply chain, while also putting significant pressure on Samsung to demonstrate the stability and production capabilities of HBM4.

Previously, the Korea Economic Daily reported that Samsung was expected to begin shipping HBM4 chips to Nvidia and AMD as early as next month. Samsung and SK Hynix are expected to announce their earnings results on January 29th, and will likely provide updates on the development progress of their HBM4 chip lines.