Working Smaller, So You Can Think Bigger
But the more nodes shrink, the more complicated their process becomes.
The road from 5nm to 4nm brought new challenges, as new innovations always do. Yield rates for 4nm mass production presented a hurdle, but Samsung Foundry has overcome that hurdle and is now quickly approaching the same yield rate for 4nm as our successful 5nm node – with an even faster learning rate.
Still to come are the lower power of 4nm products, which will provide an even higher yield that will go past the possibilities of 5nm.
And as the only foundry in the world providing 3nm GAA, Samsung Foundry is already seeing performance improvements of over 50 percent compared to 4nm.
As we move to smaller nodes and higher yields, we’re also dramatically reducing our Fault rate, leading to higher customer satisfaction throughout the entire semiconductor arena including automotive, IoT.
Finding New Growth in Auto and IoT
The growth of smart technologies is leading to a renaissance in semiconductor proliferation. From IoT solutions that keep getting smarter, to vehicles that are more responsive and even drive themselves, demand for semiconductors and specialty technologies is growing – and Samsung Foundry’s capacity is growing with them.
Samsung Foundry is right now expanding its capacity to produce the specialty products that power devices as small as wearables, to ones as large as transport trucks. We’re designing products for image sensors that capture sharper images, sensors for wearables that give you more accurate and useful information on health, driver IC, and PMIC.
Our growth as a foundry has been nothing short of remarkable. By 2023, our capacity will be 1.5 times what it was in 2019.
Amidst that growth is a renewed focus on the 8nm node – a solution that’s been a mainstay for our clients across the specialty technology spectrum. Known as a cost-effective node with superb scalability, it delivers FinFET backed high performance without using EUV.
Data Synergy: Stable Quality at Any Node
A major limitation of most Fabs is that they only produce a small number of products. They lack diversity in their line-up, and that leads to blind spots in their data. To put it simply, they don’t have enough data on hand for truly meaningful data engineering.
Samsung Foundry is overcoming the limitations of low-data volumes through Data Synergy – using large scale data gathered from multiple foundries and products. We’re using that massive pool of data to predict production and yield rates with greater accuracy than ever, an advantage owed to the vast and diverse line of products we produce.
Samsung has the largest DRAM fab in the world, providing a depth of data that can’t be recreated elsewhere. From there, IT experts at our DX Center can tap the expanse of Samsung Memory’s data to create predictive models, illustrating them for clients faster and more accurately than our competitors.
Those improvements lead to scheduling optimization, productivity improvements, and increased speeds that translate into a faster TAT.
That’s all the more important as TAT for advanced products grows longer with the use of EUV and other technologies. Data Synergy can help fast track it, while making the whole process more predictable.
New FABs, for New Challenges
US
