Samsung’s overall chip business achieved a record-breaking operating profit in Q1 2026, and that’s only going to go up as the global chip buyers are turning to Samsung Electronics for manufacturing future chips, as reported by Nikkei Asia. The list includes big tech giants such as AMD, Google, NVIDIA-backed Groq, Tesla, and BYD, all of which are looking to diversify their chip production to Samsung as demand continues to soar beyond TSMC’s available capacity.

AMD is in talks with Samsung about manufacturing future CPUs starting in 2028. Google is also reportedly expanding ties with the South Korean giant to produce its next-generation Axion processors, set to launch around 2028, and has asked Samsung to build part of its key Tensor Processing Units (TPUs) for AI computing workloads as early as 2028.

BYD, the world’s largest electric vehicle maker, is also in discussion with Samsung for its next-generation autonomous driving chips, while Tesla has already confirmed that its next-generation AI6 chip will be produced at Samsung’s Texas facility. NVIDIA-backed Groq, which develops language processing units, is already using Samsung’s foundry for chip production and may also use it for its next version of specialized AI chips.

Samsung is witnessing a sharp rise in manufacturing inquiries from both existing and new clients, people familiar with the matter told Nikkei Asia. Companies are adopting dual-sourcing strategies, splitting orders between TSMC and Samsung, rather than relying on the foundry to reduce supply-chain risk.

With tech giants like Google, Tesla, BYD, and AMD aggressively securing what’s left of TSMC’s capacity, smaller companies are left with little choice but to look elsewhere, making Samsung their strongest alternative.

Samsung’s chip business gains momentum as demand outpaces TSMC

There are only three companies in the world that can produce advanced chips: TSMC, Samsung, and Intel.

TSMC makes the most advanced chips and therefore dominates the market, followed by Samsung, which has struggled to win over major outside customers, but things have started to shift. While Intel is a massive company, its contract manufacturing business is still in its infancy regarding external scale.

Due to unprecedented demand for AI and high-performance chips, TSMC’s production capacity is unable to keep pace with the demand, forcing companies to diversify orders across multiple foundries, and that’s where Samsung shines. It’s emerging as the main alternative for companies that are not able to secure capacity at TSMC and for those who want to diversify to ensure it’s not affected by the supply-chain crisis.

Also read:
1. Huawei aims to launch lithography equivalent to TSMC’s 1.4nm by 2031 in a major defiance of U.S. sanctions
2. Apple turns to Intel to manufacture its next-gen MacBook Neo chips as TSMC supply tightens
3. TSMC’s 2nm supply shortage forces smartphone brands to save the best for top-end models amid growing DRAM crisis

For more daily updates, please visit our News Section.

And, don’t forget to join our Telegram community and sign up for our daily newsletter for top stories.

The post TSMC’s capacity crisis is handing Samsung a massive victory appeared first on Gizmochina.

Samsung has reportedly achieved a 3D transistor design milestone, which could reshape and improve future chip performance. The company continues to expand its technological footprint, along with nurturing the Foundry and LSI divisions.

According to ZDNet (via SemiconductorsX), Samsung walked away from VLSI 2026 with the Best Paper award after presenting a 3D Stacked FET design that squeezes more transistors into less space than anything the industry has managed before.

Samsung’s design piles transistors on top of each other, slashing the footprint in half and theoretically doubling density in a single architectural move.

The score was 8.29 out of 10 across a pool of over 1,000 submissions. The company also pushed the gate pitch, the horizontal width of each transistor, down to 42nm from 48nm.

Samsung’s own V-NAND flash and HBM memory both live in three dimensions. The company’s Foundry business has taken serious hits over the past two years.

Yield problems, lost clients, and a general sense that the gap between Samsung and TSMC was widening rather than closing. A Best Paper at VLSI doesn’t fix a fab, but it signals that the engineering talent is still in the building.

The post Samsung’s new 3D transistor design could improve future chip performance appeared first on Sammy Fans.

Samsung is quietly gearing up for the next big step in memory technology. According to recent reports (via @SemiconductorsX), the company is working with partners to introduce equipment for its 7th-generation 10nm-class DRAM, called “1d,” with mass production targeted for late 2027.

Right now, Samsung’s latest is the 1c (6th-gen) process at around 11-12nm. The new 1d shrinks that to about 10-11nm. Smaller lines mean better performance, lower power use, and higher density – exactly what AI systems need.

The company has already been testing samples internally. The current plan is to bring in the new equipment in the first half of 2027 and start real production toward the end of the year.

Why does this matter?

This 1d DRAM is expected to become the core die for HBM5E – Samsung’s high-bandwidth memory planned for 2029. HBM is the super-fast memory that powers AI training and inference in data centers. This stronger manufacturing process will help Samsung remain competitive with SK Hynix in the booming AI memory market.

For everyday users, this new tech will eventually reach phones, laptops, and servers, bringing better speed and lower power use. For the industry, it proves Samsung is still pushing hard on shrinking memory nodes, even as the overall memory market heats up.

It’s a steady, long-term play. No sudden big breakthrough, but these small improvements are exactly what help build real leadership in the AI memory race. We will keep an eye on updates by the end of this year.

The post Samsung’s next DRAM leap: 1d Process coming in 2027 – What it means for AI memory appeared first on Sammy Fans.

Samsung has reportedly built a data-sharing ecosystem called DSEP, the Data Sharing Eco Platform, and it’s now pulling in more than 60 chip equipment and materials partners.

The newly formed Data Sharing Eco Platform reportedly is the operational spine of Samsung’s push to run fully unmanned semiconductor fabs by the end of the decade, 2030.

Here’s what DSPE stands for

For years, Samsung kept that data locked inside its own walls. Security protocols made it nearly impossible to export error codes or processing times outside the factory.

DSEP opens a controlled slice of that process data to partners in real time, feeding it into AI models that can flag irregularities and predict failures without anyone boarding a flight.

This helps everyone (via SemiconductorsX):

• Find defects quicker
• Reduce bad chips (higher yield)
• Improve machines and materials using real factory info
• Use AI to predict problems before they happen

Samsung’s HPC Center inside its Device Solutions division already handles the infrastructure side. DSEP sits on top, connecting the partner ecosystem to that compute backbone.

Partners bring deep knowledge of their own equipment’s failure signatures. Samsung shares production-line data, partners sharpen their AI, and the whole system gets smarter.

The post Samsung targets unmanned chip fabs by 2030, DSEP is a key part of AI strategy appeared first on Sammy Fans.