A previous leak revealing that there could be potentially six versions of the Snapdragon 8 Elite Gen 6 Pro has been clarified by new findings showing the block diagrams of Qualcomm’s first 2nm chipset. To minimize the confusion, the San Diego firm is prepping just two variants, one with LPDDR5X RAM support, while the top-end one will offer LPDDR6. However, don’t count out Qualcomm introducing binned models of the Snapdragon 8 Elite Gen 6 Pro, because the way its handset business is hurting, it’ll need to adopt the “Apple” approach. Top-end Snapdragon 8 Elite Gen 6 Pro may also offer Qualcomm […]
The slightly advanced 2nm ‘N2P’ process from TSMC is reportedly being used by Qualcomm and MediaTek, with both companies aiming to gain an advantage over Apple later this year. However, the Cupertino firm is expected to match its rivals’ competitiveness by adopting the same manufacturing process next year for the A21 Pro, but only the latter will receive preferential treatment for the improved technology, not the A21. Standard A21 for the base iPhone 20 could stick with TSMC’s 2nm N2, as wafer and memory costs threaten to chew through Apple’s profits In a report from Commercial Times, Apple is reported […]
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.
Chip manufacturing giant Intel has partnered up with Taiwan's second-largest contract chip manufacturer, United Microelectronics Corporation (UMC), for advanced manufacturing process technology, suggests a report from FundaAI. Intel, under its CEO Lip-Bu Tan, is seeking to compete with Taiwan's TSMC in the contract chip manufacturing industry. According to the details, UMC is seeking to team up with Intel to gain a foothold in advanced chip manufacturing without having to expend significant amounts of capital for the machinery. Intel & Taiwan's Second Largest Chip Manufacturer Team Up For 12nm & 3nm Nodes - Report While most of the focus when it […]
ASML CEO, Christophe Fouquet, believes that the demand from Elon Musk's Terafab project will be similar to fabrication plants capable of manufacturing millions of wafers per month. The executive spoke to Bloomberg, and his discussion ranged from data centers in space to the current stage of the semiconductor industry in the context of major demand from the AI buildout and Europe's position in the AI race. ASML CEO Says Europe Is Quite Behind In The AI Race The Terafab project was announced by Musk earlier this year as part of his efforts to procure chips for Tesla's edge computing and […]
By now, everyone and their grandmother is familiar with ASML, the sole provider of EUV (Extreme Ultraviolet) capable lithography machines in the world. As hyperscalers continue to aggressively build out compute, we constantly discover new bottlenecks at almost every level of Jensen's five-layer AI cake, down to an insulating film supplied by, of all things, an MSG manufacturer. Yet for all the novel bottlenecks that crop up, there remains a perennial, critical constraint to AI build-out: the capacity of leading fabs to produce wafers on bleeding-edge process nodes. In other words, this bottleneck is EUV capacity, as all sub-7 nm […]
Korean memory manufacturer Samsung is gearing up to produce its next-generation 1d DRAM, suggests a report from the Korean media. 1d manufacturing technology is the latest in the industry, and it is equivalent to Micron's 1-delta node. As of now, Samsung is discussing equipment plans for the next-generation RAM with its manufacturing partners as part of its preparation to begin mass production as early as late 2027. Samsung Evaluating Internal Capabilities for 1d DRAM Production, Says Report Industry sources quoted by ZDNet Korea outline that Samsung is developing equipment with its partners to kick off the next generation DRAM production. […]
A standard Snapdragon 8 Elite Gen 6 followed by the premium Snapdragon 8 Elite Gen 6 Pro are slated to launch later this year, becoming Qualcomm’s first 2nm chipsets to arrive to the market. However, the advanced lithography is rumored to make the top-end SoC ludicrously expensive, but according to the latest leak, there could be six samples for Qualcomm’s partners to choose from, offering a diversified lineup for manufacturers to mix and match their launches and reduce their chipset expenditure. New leak reveals the Snapdragon 8 Elite Gen 6 Pro may also have varying CPU and GPU clock speeds to […]
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.
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.
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.
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 race to challenge Nvidia’s grip on AI hardware may be heading for another blockbuster deal. Qualcomm has been in discussions to acquire AI chip startup Tenstorrent in a transaction that could value the company between $8 billion and $10 […]
Nvidia already has more than $13 billion in cash. Now, the company is preparing to raise another $20 billion. The AI chip giant is returning to the corporate bond market for the first time in five years, according to a […]
Login
