Samsung is betting big on AMD for Exynos – here’s how it actually works

Samsung’s custom chip efforts haven’t exactly enjoyed smooth sailing. After years of production failures, design changes, and repeated outages, the company is finally aiming to get Exynos back on track. For global customers, this the comeback starts with the Galaxy S26where Samsung’s in-house silicon returns to its flagship lineup.

Exynos endured a turbulent decade. The Galaxy S23 completely skipped it due to inconsistent performance, overheating and poor efficiency, while last year’s Galaxy S series also edged it out – likely influenced by Qualcomm’s big jump with the Snapdragon 8 Elite and underperforming Samsung’s 3nm process. Although the Exynos 2500 was revealed Z Flip 7its limited use highlighted ongoing problems. In response, Samsung has restructured its chip design and manufacturing teams, including long-rumored ambitious plans to revive Exynos. a dedicated GPU for the upcoming Exynos 2800.

Graphics have been a definite part of this journey. For the past four generations, Samsung has moved away from Arm’s Mali and relied on AMD’s RDNA architecture for its Xclipse GPUs, starting with the Exynos 2200 in 2022. With Exynos back – and more ambitious graphics plans ahead – it’s a good time to evaluate the strategy surrounding the last three phone processors, Samsung’s proprietary processors. ultimately helped or hindered flagship devices.

For starters, let’s look at Exynos in isolation among recent Galaxy S flagships (we’ll leave out the Z Flip 7’s Exynos 2500 for a more like-for-like comparison). Between the Exynos 2200 and 2600, single-core CPU performance increased by 111% and multi-core by 211% in Geekbench 6. This is a huge leap and a clear sign of it Arm’s off-the-shelf CPU cores still scale welleven without the fully custom designs used by Apple and Qualcomm.

Graphics tell a similar story. During the same period, performance increased by 212% in 3DMark Wild Life Extreme and 253% in the Solar Bay ray tracing test. AMD’s Xclipse GPU has also matured nicely, overcoming the initial driver issues and now delivering the raw power expected of a modern flagship GPU.

All told, the Exynos nearly tripled in performance between the 2200 and 2600. That said, most of that jump came with the 2400, which nearly doubled performance. The 2600 is a more modest step up, offering a 40-60% gain in various tests. In isolation, it’s still a great trajectory – one that the PC market would be envious of. The problem is that Exynos is not a single upgrade.

Samsung’s move to AMD’s untested RDNA architecture was a bold, risky move. The Exynos 2200 was the first to hit the market with hardware-accelerated ray tracing in a mobile phone, giving Samsung a clear feature advantage and a strong marketing angle around gaming.

This leadership did not last long. Qualcomm added ray tracing with the Snapdragon 8 Gen 2, while Arm updated its lineup with the Immortalis GPU series later that year. Barring Qualcomm’s built-in Adreno, a more meaningful comparison is AMD vs. Arm – and here the results are mixed.

Check out MediaTek’s Dimensity 9200 (with Immortalis-G715). Size 9500Samsung’s early ray tracing quickly evaporated. The Exynos 2200 lags the Dimensity 9200 by around 33% in ray tracing performance in 3DMark’s Solar Bay test, despite its first release – although the time gap makes this comparison somewhat unfair.

This comparison points to the missing piece: the discontinued Exynos 2300 in 2023. Based on the trend line we drew, it was likely to be highly competitive, but performance issues kept it from ever shipping. As things stand, Arm’s Immortalis GPUs now consistently outperform Exynos in an area where AMD should dominate. For example, the Exynos 2600 lags the Dimensity 9500 by about 9% in this ray tracing test.

The biggest problem is traditional rasterization, which is still more important for most mobile games. Here, Exynos consistently lags behind. The 2200 was about 45% slower than the Dimensity 9200 in Wild Life Extreme, and even now the Exynos 2600 is about 19% behind rival chips from Qualcomm and MediaTek.

That gap is actually wider than previous dual-chip Galaxy S generations, which isn’t a great look for Samsung’s flagship chipsets. Perhaps the price, space, and power considerations still made the switch worthwhile from Samsung’s perspective, but from a gaming perspective, AMD’s Xclipse didn’t provide a clear advantage.

None of this will come as a surprise to longtime Samsung followers. A decade ago, Exynos and Snapdragon were true peers, but recently Snapdragon has moved on and stayed there.

While Exynos has appeared in newer Galaxy S models, it has become a secondary option. Samsung’s Ultra models They just switched to Snapdragon, reflecting the chip’s consistent leadership in CPU performance, gaming and, increasingly, AI across devices. This has created a familiar imbalance in global markets, with some customers getting the faster Snapdragon variant and others getting the Exynos.

It is not clear whether this strategy significantly reduces costs or simply offsets Samsung’s chip development costs. Given the company’s restructuring efforts and reported revenue challenges, any near-term benefits are likely limited.

To be clear, Exynos hasn’t stopped yet. Its generational gains are truly impressive, and the Exynos 2600 is by no means a slow chipset. But over the past five Galaxy generations, Samsung’s proprietary silicon has consistently outperformed its closest rivals (Google’s Tensor aside). The transition to AMD graphics did not change this dynamic.

Instead, Exynos remains a strategic investment. It gives Samsung control over its own silicon roadmap, reduces dependence on external suppliers, and enables deeper customization — whether features like Arm SME2 for AI or Heat Pass Block (HPB) technology For thermal management at 2600. This could be an optimistic sign of things to come, we’ll just have to see what’s to come when Samsung’s partnership with AMD finally comes to an end.