Latest news with #chipproduction
Bloomberg
2 days ago
- Business
- Bloomberg
TSMC Affiliate VIS May Expedite Production at $8 Billion Singapore Fab
Taiwan Semiconductor Manufacturing Co. 's smaller affiliate Vanguard International Semiconductor Corp. may accelerate the chip production schedule at its new $7.8 billion joint venture in Singapore on greater customer demand for hedging against geopolitical risks. VIS may be able to push production at the new plant, which makes mature chips, to as soon as late 2026 versus the originally announced schedule of the first half of 2027, VIS Chairman Fang Leuh told reporters at a company event on Saturday in Taoyuan, Taiwan. VIS broke ground for the facility in the fourth quarter of 2024.
Phone Arena
3 days ago
- Business
- Phone Arena
Samsung Foundry plans to pay for top sales talent to take business away from TSMC
TSMC is the world's largest foundry with a market share in the 65%-67% range. Samsung Foundry is a distant second thanks to the low yields it has achieved on advanced chip production. This makes it more costly for Samsung's customers to use the foundry instead of TSMC since low yields mean fewer chips can be produced at the same price. This is the exact reason why Qualcomm dropped Samsung Foundry in 2022. At the time, Samsung Foundry was manufacturing the Snapdragon 8 Gen 1 and reportedly had a 35% yield. TSMC's 4nm yield rate was 70% at the time. As a result, Qualcomm picked up all of its marbles from Samsung's playground and moved them to TSMC. The latter built the slightly reconfigured Snapdragon 8+ Gen 1 and has manufactured all of Qualcomm's flagship Snapdragon application processors since then. There is no denying that the Taiwan-based chip foundry has a star-studded list of clients that are among the top names in tech including Apple, AMD, Nvidia, MediaTek, Qualcomm, and more. Thanks to Samsung Foundry's poor yields, Qualcomm switched to TSMC to make the Snapdragon 8+ Gen 1. | Image credit-PhoneArena Samsung Foundry might catch a break if it can improve yields on the production of Samsung's own Exynos 2600 application processor which will be built on the foundry's 2nm process node. If Samsung Foundry can get its yield up enough to ensure the production of the number of chips Samsung needs, the Exynos 2600 will power the Galaxy S26 and Galaxy S26 + in most markets and those models will become the first smartphones to feature a 2nm application processor. Usually, the iPhone has been the device that welcomed the next cutting-edge process node. But next year, if Samsung plays its cards right, this honor will belong to the Galaxy S26 and Galaxy S26 +. Ironically, the Galaxy S26 Ultra, the top-of-the-line model, will be powered by a 3nm Snapdragon 8 Elite 2 for Galaxy and won't be involved in any record-breaking historical moments. Samsung has been supposedly looking to spin off or sell its Foundry. Until a deal can be made to its liking, Samsung hopes to generate more revenue for the unit by trying to convince TSMC's customers to make the switch to Samsung Foundry. The latter hired away a former TSMC sales executive earlier this year and a report out of Korea states that Samsung is willing to pay as much as $300,000 to pros looking to join Samsung Foundry's sales team in the U.S. Both Samsung and TSMC have built U.S. fabs and TSMC has started producing 4nm chips in Arizona and will build 2nm components in 2028. Samsung's timeline has it ahead of TSMC in the States. In Taylor, Texas, Samsung Foundry will mass produce 2nm chips in 2026. Samsung needs a strong order flow to keep the Taylor, Texas line humming, churning out advanced chips. That is why Samsung Foundry will pay California-based salesmen as much as $300,000. A foundry sales director can earn as much as $319,800 and a senior manager inside a foundry's customer service unit, can draw a salary within the $180,950 to $289,050 range. The salaries Samsung is willing to pay in the U.S. top what it is paying for the same jobs in South Korea. The new hires will have the task of increasing Samsung Foundry's market share in the United States. Some of TSMC's customers in Taiwan will also take delivery of chips built in the U.S. including Apple, AMD, Nvidia, and Qualcomm. TSMC is already shipping 4nm chips from its Arizona fab to Apple and Nvidia. Semiconductors built in Arizona will not face any tariffs which is part of the appeal of purchasing domestically manufactured chips. Depending on which report you read, the third largest foundry globally is China's SMIC. The latter manufactures the chips designed by Huawei and is the largest foundry in the country. Secure your connection now at a bargain price! We may earn a commission if you make a purchase Check Out The Offer
Phone Arena
6 days ago
- Business
- Phone Arena
Samsung could produce cutting-edge chips in the U.S. two years before TSMC
At the start of this year, we told you that the world's largest foundry, TSMC, is expected to manufacture chips using its 2nm process node in the United States by 2028. The foundry's Arizona facilities will produce the cutting-edge chips and even though TSMC is going to be shipping 2nm silicon from its Taiwan fabs during the second half of this year or early next year, it will still be quite the feather in President Donald Trump's hat to have 2nm chips made in the states just 2 years after they are launched in Taiwan. In that same report from back in January, we also noted that TSMC won't be the only foundry to build 2nm chips in the States. Samsung will also be producing 2nm chips in the U.S. as the company hopes to join TSMC in taking advantage of President Trump's desire to have cutting-edge semiconductors built in the U.S. along with major consumer electronics like the iPhone. The article we published in January said that Samsung was building a fab for advanced chip production in Taylor, Texas. Now, five months later, Samsung has started preparing production lines at the Taylor facility. We could see 2nm chips roll off the line in Taylor starting next January or February which would be two years ahead of TSMC's 2028 target for 2nm production in the U.S. There has been plenty of interest from tech firms interested in buying U.S.-made advanced chipsets, especially in light of the tariffs imposed by President Trump. Semiconductors built in the U.S. would not be subject to tariffs which are simply import taxes placed on products imported into the U.S. Even with a head start as long as two years over TSMC's U.S. 2nm production, Samsung Foundry must get its yield up. | Image credit-RetailNewsAsia Globally, Samsung Foundry is well behind TSMC in global market share and part of the reason is Samsung Foundry's poor yields on advanced process nodes which drives up the prices customers pay for the chips they purchase. Samsung's plans originally called for it to produce 4nm chips in the U.S. at first. Despite spending heavily on the project, the Korean firm didn't manufacture one 4nm chip in the U.S. which has many questioning whether Samsung can build 2nm chips in the U.S. As it always seems to be when discussing Samsung Foundry, the Foundry's yield is key. Supposedly, Samsung has achieved a 40% yield with its SF2 (2nm) production compared to 60% for TSMC. Typically a yield of 70% or higher is needed to start mass production so while Samsung hopes that it can take advantage of the two-year head start it will have producing 2nm chips in America, the story remains the same: Samsung needs to get its yield up so that it can start mass production at the 2nm node in the U.S. Secure your connection now at a bargain price! We may earn a commission if you make a purchase Check Out The Offer
Yahoo
19-06-2025
- Business
- Yahoo
Texas Instruments commits $60B to domestic chip manufacturing
This story was originally published on CIO Dive. To receive daily news and insights, subscribe to our free daily CIO Dive newsletter. Texas Instruments plans to spend more than $60 billion to increase chip production domestically, according to a Wednesday announcement. The initiative represents the 'largest investment in foundational semiconductor manufacturing in U.S. history,' the company said. The nearly century-old technology provider will build seven semiconductor fabrication facilities across three manufacturing 'mega-sites' located in Texas and Utah. The plants will produce analog and embedded processing chips for Apple, Ford, Medtronic, Nvidia and SpaceX as part of the initiative, the company said. Texas Instruments was awarded $1.6 billion in CHIPS and Science Act funding last year. Despite President Donald Trump's efforts to undo the Biden-era legislation, his administration lent its support to the Texas Instruments initiative. 'President Trump has made it a priority to increase semiconductor manufacturing in America — including these foundational semiconductors that go into the electronics that people use every day,' U.S. Secretary of Commerce Howard Lutnick said in the Wednesday announcement. Texas Instruments' buildout plan comes amid compute consumption spikes spurred by AI adoption and concurrent enterprise modernization efforts. Hyperscaler infrastructure investments drove an 18% year-over-year increase in semiconductor revenue globally last year, according to Gartner. The analyst firm anticipates the massive market to climb at least another 11% this year to over $700 billion. The three largest cloud providers — AWS, Microsoft and Google Cloud — reported compute capacity constraints as enterprises ramped up AI projects last year, triggering a wave of multi-billion-dollar hyperscale data center investments. Oracle, a junior member of the hyperscale club, saw capital expenditures more than double to over $20 billion during its fiscal year, which ended May 31. Texas Instruments' manufacturing push has geopolitical dimensions, as well. President Trump made a domestic manufacturing rebuild central to his administration's policy objectives, and the technology sector has responded. In April, amid confusion and uncertainty triggered by Trump administration tariffs, IBM announced plans to invest $150 billion over the next five years in domestic research, development and manufacturing. Nvidia — one of the biggest beneficiaries of the data center building boom — pledged $500 billion to bolstering domestic chip manufacturing in April. The GPU giant has seen quarterly revenues skyrocket to more than $40 billion, up from under $6 billion two-and-a-half years ago. Texas Instruments hasn't been as fortunate. The company reported $4.1 billion in revenue during the first three months of the year, up 11% compared to Q1 2024. With the Trump administration threatening levies on semiconductor imports, Texas Instruments President and CEO Haviv Ilan indicated the company would shift manufacturing from foundries in Taiwan to a plant in Lehi, Utah, during a recent earnings call. Texas Instruments is ramping up production this year in Utah and Texas facilities as it continues to build capacity. The timeline for the $60 billion investment will be based on business demand, a spokesperson told CIO Dive. Recommended Reading Hyperscalers bet on costly new data centers to capture growing market Error in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data
Yahoo
18-06-2025
- Business
- Yahoo
Nvidia (NVDA) Target Raised to $200 by Barclays Amid Blackwell Ramp, AI Cloud Expansion
Nvidia (NVDA, Financials) received a price target upgrade from Barclays on Monday, with analysts citing stronger-than-expected Compute segment growth and expanding Blackwell chip production. Warning! GuruFocus has detected 4 Warning Signs with NVDA. Barclays raised its target to $200 from $170 and maintained an Overweight rating, forecasting Nvidia will beat consensus revenue estimates by $2 billion for the July quarter. The firm now expects Q3 revenue to reach $42 billion and Q4 to hit $48 billion, compared to Wall Street's estimates of $40.8 billion and $46.2 billion, respectively. The revision comes as Blackwell chip production scales, reaching 30,000 wafers per month in June. Though below the firm's earlier forecast of 40,000, Barclays said utilization rates remain healthy and expects ramped system sales to comprise nearly half of total revenue by October. The forecast boost aligns with positive supply chain feedback and accelerating enterprise AI demand. Nvidia's collaboration with Deutsche Telekom to build Europe's first industrial AI cloud also signals geographic expansion. The project involves 10,000 chips from Nvidia and is expected to complete by 2026. Nvidia is excluding China from revenue guidance amid U.S. restrictions, CEO Jensen Huang confirmed. Competitors like Advanced Micro Devices (AMD, Financials) are stepping up AI chip efforts, with new MI350 and MI400 series launches, but analysts suggest Nvidia's ecosystem remains dominant. Investors should monitor upcoming Blackwell Ultra volumes and progress on European cloud deployments for near-term catalysts. This article first appeared on GuruFocus. Error in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data
