Samsung’s foundry dreams have slammed into the hard maths of yield. A fresh rumour says the Exynos 2800 will dodge sub-2nm bravado and stick with a tuned 2nm GAA process instead.
If the rumours are correct, the chip will not be mass-produced on the 1.4nm SF1.4 node, with Samsung preferring an upgraded 2nm option. The reasoning is simple enough; there is no point designing silicon that you cannot make in volume.
ZDNet Korea claims Samsung had pencilled in a 2027 move to 1.4nm for the Exynos 2800, codenamed Vanguard, then backed off. The firm is said to be prioritising yield stabilisation and practical optimisations rather than chasing the smallest number.
The report says the Exynos 2800 will lean on SF2P+, an improved flavour of Samsung’s 2nm GAA node, to refine performance and cost. That should let Samsung tweak the design without turning each chip into a pricey boutique item.
Exynos 2800 is tipped to be Samsung’s first SoC paired with an in-house GPU. If the manufacturing process tanks yield, that shiny “in-house” bit becomes a supply problem.
The 1.4nm node is expected to take time to settle, and pushing mass production to an immature process would drag yields down and drive up the cost per unit. That would clash with Samsung’s apparent goal of getting the Exynos 2800 into more Galaxy S28 phones.
Samsung is said to have finished the basic design of its second-generation 2nm GAA process, SF2P, with SF2P+ planned within about two years. The roadmap is meant to boost its foundry pitch, starting with the Exynos 2600.
Yields are still the sore point, with another report claiming Samsung’s 2nm GAA process is sitting at about 60 per cent. Before the Exynos 2800, Samsung is expected to roll out the Exynos 2700 this year, codenamed Ulysses, for higher-volume use in the Galaxy S27 line.
