Tesla tightens Samsung’s GDDR6 and tightens the graphics market

Samsung Electronics has started expanding production of 8 Gb GDDR6 this month for Tesla, and the monthly volume requested by Elon Musk’s company is already about four times higher than in the first quarter, according to a report by eDaily also picked up by other industry outlets. Samsung has not officially confirmed this information, so for now, it should be read as a report from Korean press rather than a formal announcement from the manufacturer.

This data matters because it comes at a time of intense tension in graphics memory markets. TrendForce reported the average spot price of 8 Gb GDDR6 at $12.335 on April 13, 2026, a very high figure for this segment, even as Samsung continues to market GDDR6 in densities of 8 Gb and 16 Gb with speeds up to 24 Gbps. In other words, demand remains strong and manufacturers’ flexibility seems limited.

Tesla isn’t seeking gaming memory, but it could impact that market

According to the Korean report, Tesla has been requesting more GDDR6 supply from Samsung since the beginning of the year for its infotainment and autonomous driving systems, with requests in January exceeding five times the initial forecast. The same article states that Samsung does not plan to significantly increase its overall GDDR6 8 Gb production, and has even reduced some of its GDDR6 range, making any shifts in priorities among clients more delicate.

This preferential treatment doesn’t happen in a vacuum. Samsung and Tesla already have a significant industrial relationship in automotive AI chips. Reuters reported in July 2025 a $16.5 billion agreement for Samsung to manufacture Tesla’s future AI6 chips at its Texas facility, and Korean specialized media have also noted that the recent tape-out of AI5 relied on Samsung and TSMC. While this doesn’t necessarily mean Tesla will absorb all available memory, it helps explain why Samsung might prioritize its needs over other clients.

The potential problem for AMD, Intel, and NVIDIA is credible, though not confirmed

It’s important to differentiate confirmed facts from inferences. There’s no public evidence that Samsung has already reduced deliveries to AMD, Intel, or NVIDIA, nor that these companies depend exclusively on Samsung for GDDR6. For example, Micron continues selling GDDR6 for graphics applications, indicating the market isn’t supplied by a single vendor.

However, a reasonable conclusion is that if Tesla consumes much more 8 Gb GDDR6 and Samsung doesn’t increase total production in parallel, market pressure could shift to graphics cards still relying on this memory. Many current and recent models, especially outside the high-end segment, could be affected.

Some official examples of graphics cards still using GDDR6 include the Radeon RX 9070 XT and RX 9070, both with 16 GB and a 256-bit bus; the Intel Arc B580 with 12 GB and a 192-bit bus; the Arc B570 with 10 GB and a 160-bit bus; and NVIDIA models like the RTX 4060 Ti, RTX 4060, RTX 3050, and RTX 3060.

The practical implication is simple: if supply tightens, the first things to be affected are usually new architecture launches, availability, lead times, and final costs of models that still rely on GDDR6. This is especially impactful in mainstream and entry-level segments, where memory costs weigh heavily on the overall price compared to premium GPUs.

The background is an increasingly strained memory industry driven by AI

Tesla’s move also aligns with a broader context. The memory market has been under pressure for months due to AI trends and the prioritization by major manufacturers of higher-margin products, especially HBM and advanced memory for accelerators. The Verge recently summarized that the global RAM shortage could last for years and that the industry might not meet all short-term demand, as much of the new capacity isn’t expected to come online until 2027 or later.

Meanwhile, SK hynix announced this week an investment of around $13 billion in a new plant in South Korea to meet the demand for AI memory, confirming that capacity is still expanding but mainly where strategic profitability is highest. None of this is specifically targeted at short-term relief for the gaming GDDR6 market.

Thus, the news about Tesla and Samsung has deeper implications. It’s not just about a car manufacturer requesting more memory. Advanced automotive applications, autonomous driving, and embedded AI are starting to compete more directly for components traditionally associated with the graphics market. In a supply-limited environment, this often manifests first in prices and stock availability, rather than official statements.

Frequently Asked Questions

Has Samsung officially confirmed quadrupling GDDR6 for Tesla?
No. The information comes from an eDaily report, replicated by other outlets. Samsung has not issued an official confirmation with figures yet.

Does Tesla use Samsung GDDR6 for graphics cards?
No. According to Korean press, Tesla uses it for infotainment and autonomous driving systems, not for consumer graphics cards. The impact on gaming would be indirect, due to supply competition.

What current graphics cards still use GDDR6?
Official models still employing GDDR6 include the AMD RX 9070 XT and RX 9070, the Intel Arc B580 and B570, and several NVIDIA cards such as the RTX 4060 Ti, RTX 4060, RTX 3050, and RTX 3060.

Will this cause graphic card prices to rise?
It’s too early to say definitively. However, increased GDDR6 consumption by Tesla, combined with high spot prices and tight supplies, does raise the risk of stock shortages and higher costs for models relying on this memory.

via: edaily.co.kr