Latest news with #SamsungFoundry
Phone Arena
2 hours ago
- Business
- Phone Arena
Wall Street research firm surpisingly says that this foundry is the only alternative to TSMC
Many of you are familiar with TSMC, the Taiwan-based foundry that has a dominating 67.6% market share in the contract chip manufacturing industry. Its client list sounds like a who's who of tech leaders. Apple is its biggest client, and TSMC also produces chips for companies like Nvidia, AMD, Qualcomm, MediaTek, Sony, and Broadcom. Its top competitor is Samsung Foundry, but calling the latter a rival to TSMC is like calling the San Jose Sharks rivals of the back-to-back Stanley Cup Champion Florida Panthers. The other day, we told you about the first sign indicating that Samsung Foundry could be starting a turnaround. Tesla placed an order for $16.5 billion worth of chips, and the contract runs until 2033. But Samsung has a huge gap to fill as its 7.7% market share is nearly 6,000 basis points behind TSMC. Samsung Foundry has had serious issues with low yields, which could end up raising the prices of chips sold to clients. Samsung Foundry recently announced a $16.5 billion deal with Tesla to make its AI6 chip. | Image credit-X A foundry's yield measures the percentage of usable chips divided by the maximum number that could have been made on a silicon wafer. A low yield means that the customer has to pay more to get the number of non-defective chips it needs. With it weighed down by low yields, many in the industry wondered how Samsung Foundry could get a new customer, and whether Samsung would sell off the unit. Even though Intel's experiment to create a contract chip manufacturing unit to challenge TSMC and Samsung hasn't exactly paid off as expected, Intel is still going ahead with the production of advanced chips. But most of these components will be used by Intel itself. So far, Intel Foundry Services (IFS) has been a major flop as the company has zero significant external customers and the foundry's market share isn't in the top ten globally. Making matters worse, Intel is losing huge amounts of money. It has even taken hundreds of millions in write-offs for equipment it bought but has no use for. With construction delays slowing down the building of fabs in Ohio, and the cancellation of projects in Germany and Poland, it is hard to understand why some analysts still see Intel as a reasonable alternative to TSMC. One Wall Street research firm, Bernstein, says that Samsung Foundry is the only logical choice to be an alternative to TSMC. In a note to its clients, Bernstein, after the $16.5 billion deal between Samsung Foundry and Tesla was announced, wrote, "The world needs a leading-logic semiconductor producer in addition to TSMC, and that Samsung…is better positioned than Intel to be the alternative." Bernstein says that the announced deal for the AI6 custom-designed AI chips for Tesla will result in peak annual revenue for Samsung Foundry of $2 billion-$2.5 billion. The research firm says that for the length of the deal, the contract will contribute only $8 billion to Samsung Foundry's revenue. But this isn't necessarily bad news for Samsung Foundry. Bernstein says that the value of the contract Tesla has with Samsung is much lower than the disclosed US$16.5B." But the Wall Street firm notes that Tesla CEO Elon Musk says that the potential valuation of the contract could be "several times higher" than the initially reported $16.5 billion indicating that besides building chips for Tesla's vehicles, some of the contract could cover the manufacturing of chips for robotics or other products. Bernstein says that the $16.5 billion deal with Tesla could raise Samsung Foundry's revenue by 30% to 40% and the improvement to Samsung Foundry's profitability should be higher. The Wall Street research firm says that Samsung Foundry's fab in Taylor, Texas, is currently running at low capacity; this deal should help that fab's business grow.. As for what will happen to TSMC in light of the deal between Tesla and Samsung Foundry, Bernstein says that "The impact on TSMC is negligible." The research firm estimates that the amount of revenue Tesla generates for TSMC is very low.
Phone Arena
2 days ago
- Phone Arena
Exynos 2600 could fix Samsung's biggest chip problem
Power efficiency and heat management are super important (if not, the most important) aspects of chip design. Samsung's own Exynos processors are unfortunately known for their poor efficiency and heat management when compared with Qualcomm's. However, a new report from South Korea indicates that Samsung is trying new technologies to work on its upcoming in-house chips and to improve heat management. The Exynos 2600 processor, which is expected to be used in some Galaxy S26 versions, reportedly sports a newer thermal management tech. Samsung is said to use a new chip packaging technology, which should improve the performance of the Exynos 2600 chip. The company may reportedly use Heat Pass Block (otherwise known as HPB), a way of inserting heat-dissipating materials inside the semiconductor chip package. The Exynos 2600 is said to be a 2nm chip designed by Samsung and fabricated using Samsung Foundry. It would be the first chip by the company to use HPB. HPB is a copper-based heat sink that is placed on top of the application processor and DRAM component. Image Credit - Samsung This heat sink is intended to absorb the heat that is generated by the CPU, GPU, RAM, and other components found in smartphone SoCs. Samsung is expected to finish testing the Exynos 2600 by October of this year. If it gets favorable results, the mass production of the processor will soon follow, with the intention of it to power some of the Galaxy S26 phones, expected for an early 2026 launch. The Galaxy S26 Ultra may be the only phone sporting exclusively the Snapdragon 8 Elite Gen 2, which is a processor made by Qualcomm. In the past few years, Samsung has been trying new chip packaging methods for Exynos chips, looking for a way to improve their performance and have them more on par with Qualcomm's counterparts. For example, the Exynos 2400 used the Fan-Out Wafer Level Packaging (FOWLP) technology, which had both the input and output terminals outside the semiconductor chip, which offered better heat dissipation. The Exynos 2600 is also said to use FOWLP. I personally think this is a smart move by Samsung. If they can really improve heat management and efficiency, that could make Exynos chips much more competitive. A cooler phone means better performance and longer battery life – things everyone wants. It's good to see Samsung learning from past mistakes and trying new tech like HPB and FOWLP. If testing goes well, the Galaxy S26 with the Exynos 2600 might surprise us in a good way.
Phone Arena
5 days ago
- Phone Arena
This is why your experience with the Galaxy S26 Ultra could be so much different than mine
Qualcomm's new flagship application processor (AP), the Snapdragon 8 Elite Gen 2, is expected to power the top-of-the-line Galaxy S26 Ultra when the phone is released in late January or early February next year. However, a fresh rumor reveals that two Galaxy S26 Ultra owners could have differing experiences with their phones depending on the foundry that produced the Snapdragon 8 Elite Gen 2 application processor (AP) powering their units. It's a strange situation to think that the Galaxy S26 Ultra in my hand could run faster than the one in yours, and deliver better battery life even though the AP that both are using is the same. Without trying to sound repetitive, the difference would be in the foundry that manufactured the two chipsets and the process node employed by each. In plain English, Qualcomm is rumored to be using "dual-sourcing" for the Snapdragon 8 Elite Gen 2, which means that some APs will be made by TSMC and some by Samsung Foundry. As of this moment, Qualcomm will dual-source the Snapdragon 8 Elite Gen 2 AP. | Image credit-Weibo Posting on Chinese social media site Weibo, tipster Digital Chat Station says that the Samsung-built version of the Snapdragon 8 Elite Gen 2 AP has not yet been cancelled. If true, at this moment, we are looking at two different variants of the chip. TSMC will build the chipset using its third-generation 3nm process node which is the same process TSMC will use to build the A19 APs for the upcoming iPhone 17 series. Samsung Foundry, on the other hand, will use its 2nm Gate-All-Around (GAA) process node, which in theory would be technologically superior to TSMC's 3nm version of the chipset. That's because as the process node number drops, so does the size of the transistors employed. Smaller transistors mean an increase in transistor density, which typically measures the number of transistors in a die per square millimeter. This means that a chip carries more transistors in a given area. The transistor density figure is important because as that number rises, it usually means that a foundry can pack more transistors into a given space, which means that smaller transistors are being used. Smaller transistors use less power, switch states more quickly, and reduce the manufacturing cost per function. Samsung Foundry also has another advantage as its 2nm process node includes the use of GAA transistor architecture which results in the gate surrounding the channel on all four sides. This reduces current leaks and improves the drive current, which results in better-performing APs that use less power. The pricing of the Snapdragon 8 Elite Gen 2 is rumored not to be much higher than the price of the Snapdragon 8 Elite. Part of that is due to Qualcomm's use of its 3nm process (albeit its third-generation version), and Samsung Foundry's involvement. This could be a big deal for the latter, which has a market share in the industry of only 7.7% compared to TSMC's 67.6% (both figures are Q1 2025 numbers). Also, it would be a positive sign that Samsung Foundry has been able to improve its awful yield figures and was able to hit its goal of 50%. Whatever happens with the Snapdragon 8 Elite Gen 2, there is speculation that Qualcomm will copy Apple next year and offer two different versions of the Snapdragon 8 Elite Gen 3. With a much higher price expected to be charged by TSMC for its 2nm wafers, a pricier, more cutting-edge version of the Snapdragon 8 Elite Gen 3 made by TSMC could be used in more premium versions of Samsung's flagship Galaxy S27 series in 2027. The non-premium version of the AP would be made by Samsung Foundry using its 2nm process and would have lower CPU and GPU clock speeds and reduced cache. Samsung Foundry's Snapdragon 8 Elite Gen 2 is rumored to have a model number of SM8850s and have the codename 'Kaanapali S'. The Snapdragon 8 Elite Gen 3 "Pro" could have a model number of SM8950, with the "non-Pro" model carrying a number of SM8945. Will Qualcomm dual-source the Snapdragon 8 Elite Gen 2? As soon as we have an update, we will let you know.
Geeky Gadgets
18-07-2025
- Geeky Gadgets
Flagship Showdown: Exynos 2500, Snapdragon 8 Elite, Tensor G5, & A18 Pro Compared
When assessing flagship chipsets, performance remains the definitive metric for determining their value. The video below from Techmo gives us more insights into the latest processors from leading manufacturers: Samsung's Exynos 2500, Apple's A18 Pro, Qualcomm's Snapdragon 8 Elite and 8 Gen 3, and Google's Tensor G4. By analyzing CPU, GPU, and thermal performance through standardized benchmarks and real-world applications, you'll gain a detailed understanding of how these chipsets stack up against one another. Watch this video on YouTube. Key Chipsets and Manufacturing Processes The chipsets under review represent the pinnacle of mobile processing technology, using advanced manufacturing techniques that significantly influence their performance and efficiency. Two key players, TSMC and Samsung Foundry, dominate the fabrication of these processors: Qualcomm's Snapdragon 8 Elite and 8 Gen 3: Both are produced using TSMC's innovative fabrication nodes, making sure high performance and energy efficiency. Both are produced using TSMC's innovative fabrication nodes, making sure high performance and energy efficiency. Samsung's Exynos 2500 and Google's Tensor G4: These rely on Samsung Foundry's advanced processes, showcasing the company's efforts to compete in the high-performance chipset market. These rely on Samsung Foundry's advanced processes, showcasing the company's efforts to compete in the high-performance chipset market. Apple's A18 Pro: Manufactured using TSMC's 3nm process, this chipset represents the most advanced fabrication technology available today, offering unparalleled efficiency and power. The choice of manufacturing process directly impacts not only raw performance but also thermal management and battery efficiency, making it a critical factor in evaluating these chipsets. CPU Performance: Single-Core and Multi-Core Benchmarks CPU performance is a cornerstone of chipset evaluation, as it dictates how well a device handles tasks ranging from app responsiveness to multitasking and complex computations. Standardized benchmarks such as Geekbench provide a clear picture of how these chipsets perform: Single-Core Performance: Apple's A18 Pro leads the field, delivering unmatched speed and efficiency, making it ideal for tasks requiring quick processing. Apple's A18 Pro leads the field, delivering unmatched speed and efficiency, making it ideal for tasks requiring quick processing. Multi-Core Performance: Qualcomm's Snapdragon 8 Elite Galaxy Edition excels in handling demanding workloads, securing its position as a top performer in multi-threaded tasks. Qualcomm's Snapdragon 8 Elite Galaxy Edition excels in handling demanding workloads, securing its position as a top performer in multi-threaded tasks. Exynos 2500: While it shows significant improvement over its predecessor, the Exynos 2400, it still lags behind Snapdragon variants in both single-core and multi-core benchmarks. These results highlight the varying strengths of each chipset, with Apple dominating single-core tasks and Qualcomm excelling in multi-core scenarios. GPU Performance: Graphics and Gaming Graphics performance is a critical consideration for gaming enthusiasts and users of GPU-intensive applications. The chipsets demonstrate varying levels of capability in this area: Exynos 2500: Stands out in Android-specific GPU tests, occasionally outperforming Snapdragon chipsets in certain scenarios, particularly in optimized gaming environments. Stands out in Android-specific GPU tests, occasionally outperforming Snapdragon chipsets in certain scenarios, particularly in optimized gaming environments. Apple's A18 Pro: Maintains its position as the gold standard for GPU performance, though direct comparisons with Android chipsets are complicated by platform differences. Maintains its position as the gold standard for GPU performance, though direct comparisons with Android chipsets are complicated by platform differences. Snapdragon 8 Elite Variants: Deliver consistent and reliable results, making them a strong choice for gaming and other graphics-heavy applications. Benchmarking tools such as 3DMark Wild Life and Geekbench provide detailed insights into these differences, showcasing the strengths and weaknesses of each chipset in handling graphical workloads. Thermal Management and Stress Testing Thermal performance plays a crucial role in determining a chipset's usability, particularly in compact devices like foldable smartphones, which are more prone to overheating. Stress tests reveal how well these processors manage heat under sustained workloads: Snapdragon 8 Elite Variants: Demonstrate excellent thermal management, maintaining stable performance even during prolonged stress tests. Demonstrate excellent thermal management, maintaining stable performance even during prolonged stress tests. Exynos 2500: Shows resilience but is more susceptible to thermal throttling, especially in compact devices such as the Galaxy Z Flip 7, where space for heat dissipation is limited. Shows resilience but is more susceptible to thermal throttling, especially in compact devices such as the Galaxy Z Flip 7, where space for heat dissipation is limited. Apple's A18 Pro: Effectively balances performance and heat generation, making sure consistent results without significant throttling. Efficient thermal management is essential for maintaining performance and preventing overheating, particularly in modern devices with limited internal space. Real-World Application Performance While benchmarks provide valuable data, real-world performance tests offer practical insights into how these chipsets handle everyday tasks. These tests reveal how well the processors perform under typical usage scenarios: Video Export: Apple's A18 Pro consistently completes tasks the fastest, showcasing its ability to handle complex workloads efficiently. Apple's A18 Pro consistently completes tasks the fastest, showcasing its ability to handle complex workloads efficiently. Exynos 2500: Delivers competitive results, occasionally outperforming Snapdragon variants in specific real-world scenarios. Delivers competitive results, occasionally outperforming Snapdragon variants in specific real-world scenarios. Tensor G4: Struggles to keep up, underperforming in nearly all tested applications, highlighting its limitations in optimization and raw power. These findings emphasize the importance of optimization and platform-specific tuning in achieving superior real-world performance. Benchmarking Tools and Key Observations To ensure a comprehensive evaluation, multiple benchmarking tools were employed, including Geekbench, AnTuTu, and 3DMark Wild Life. Key observations from these tests include: Apple's A18 Pro: Leads overall, particularly excelling in single-core CPU performance and real-world application tests. Leads overall, particularly excelling in single-core CPU performance and real-world application tests. Snapdragon 8 Elite Variants: Consistently rank high across all benchmarks, with the standard edition achieving the highest AnTuTu score. Consistently rank high across all benchmarks, with the standard edition achieving the highest AnTuTu score. Exynos 2500: Shows significant progress over its predecessor but struggles with thermal throttling in compact devices. Shows significant progress over its predecessor but struggles with thermal throttling in compact devices. Tensor G4: Underperforms across all categories, highlighting its challenges in competing with other flagship chipsets. These observations underscore the diverse capabilities of each chipset and their suitability for different use cases. Ranking the Flagship Chipsets Based on the analysis, the flagship chipsets can be ranked as follows: 1st Place: Apple's A18 Pro – Excels in both benchmarks and real-world applications, setting the standard for performance and efficiency. Apple's A18 Pro – Excels in both benchmarks and real-world applications, setting the standard for performance and efficiency. 2nd Place: Qualcomm's Snapdragon 8 Elite Variants – Offer a strong balance of CPU, GPU, and thermal performance, making them a reliable choice for most users. Qualcomm's Snapdragon 8 Elite Variants – Offer a strong balance of CPU, GPU, and thermal performance, making them a reliable choice for most users. 3rd Place: Samsung's Exynos 2500 – A significant improvement over previous generations but still trails Snapdragon in most tests. Samsung's Exynos 2500 – A significant improvement over previous generations but still trails Snapdragon in most tests. 4th Place: Google's Tensor G4 – Struggles to compete, underperforming in nearly every category. This comparison highlights the rapid advancements in chipset technology and the varying levels of performance across manufacturers. Whether your priority is raw power, thermal efficiency, or real-world usability, understanding these differences can guide you in selecting the right device for your needs. Stay informed about the latest in Exynos 2500 by exploring our other resources and articles. Source & Image Credit: Techmo Filed Under: Android News, Apple iPhone, Guides Latest Geeky Gadgets Deals Disclosure: Some of our articles include affiliate links. If you buy something through one of these links, Geeky Gadgets may earn an affiliate commission. Learn about our Disclosure Policy.
Phone Arena
12-07-2025
- Business
- Phone Arena
Questionable image of Tensor G5 prototype surfaces
Pixel 10 Pixel 10 Receive the latest Google news Subscribe By subscribing you agree to our terms and conditions and privacy policy On the other hand, there is some indication that Samsung Foundry felt that it would get the job of building the Tensor G5. As we told you last month, Korean financial newsletter "The Bell" said that Samsung Foundry is investigating why it lost the business of building the Tensor G5 to TSMC . While that doesn't prove that Samsung foundry built a Tensor G5 prototype, it does indicate that Samsung might have been surprised by Google's decision to have TSMC build the AP although most analysts were not. Samsung Foundry's yield problems have resulted in the Samsung unit's market share declining to 7.7% during Q1 2025 from 8.1% during the fourth quarter of last year. Meanwhile, TSMC's share has risen from 67.1% to 67.6% from Q4 2024 to Q1 2025, respectively. Galaxy S26 Galaxy S26 Pixel 10 Grab Surfshark VPN now at more than 50% off and with 3 extra months for free! Secure your connection now at a bargain price! We may earn a commission if you make a purchase Check Out The Offer As we've mentioned before, the Pixel 10 will be powered by the Tensor G5 application processor (AP). This will be the first Tensor chip designed from the ground up by Google ; previous Tensor chips used a Samsung Exynos AP as a foundation while Google added some customized parts of the chipset. The change will allow Google to design exclusive Pixel features right into the chipset. The Tensor G5 is also the first Tensor SoC built by the world's largest foundry, TSMC. The four previous Tensor APs were built by Samsung the struggles that Samsung Foundry is having with its yield at 3nm, it is no surprise that TSMC ended up with the job. But a tipster posted an image on "X" showing what he says is a prototype of the Tensor G5 AP. The tipster, @Jukanlosreve, wrote, " I received a photo of aprototype chip, and it's marked as SEC (Samsung). It certainly seems that, at least for the early versions of the, they were considering having the Tensor G5 manufactured at Samsung Foundry."@Jukanlosreve might have thought that the SEC marking on the chip stood for Samsung Electronics Company. However, while Samsung uses that abbreviation in press releases and news articles; it would not be a marking Samsung would typically use on a finished chip to indicate that it was made by Samsung Foundry. Also, there are signs that certain numbers on the chip have been concealed, possibly to perpetuate the myth that Samsung Foundry built a prototype of the Tensor most important project that Samsung Foundry is working on is the 2nm Exynos 2600 AP, expected to power the Galaxy S26 and+ in most markets outside of the U.S., Canada, and China. If Samsung uses the Exynos 2600 AP in those models, in most markets worldwide, themodels using Sammy's homegrown AP will be the first equipped by a chip produced using the 2nm process thing that is clear is that the Tensor G5 AP will not be manufactured by Samsung Foundry and will be produced by TSMC. Theseries is rumored to be unveiled on August 20th and released on August 28th.
