Hasan Taufique

It's an exciting time at Texas Instruments as we announce an award agreement of up to $1.6B in direct funding through CHIPS for America, which will support three new fabs under construction in Sherman, Texas, and Lehi, Utah. https://lnkd.in/g_6bc4Tp ​

132

1 Comment

It's an exciting time at Texas Instruments as we announce an award agreement of up to $1.6B in direct funding through CHIPS for America, which will support three new fabs under construction in Sherman, Texas, and Lehi, Utah. https://lnkd.in/gYTcf_FF ​

566

7 Comments

TSMC could charge up to $45,000 for 1.6nm wafers — rumors allege a 50% increase in pricing over prior-gen wafers 🤯💡 As TSMC is gearing up to start making #chips on its N2 (2nm-class) process #technology later this year, rumors have emerged regarding the pricing of N2 #wafers, as well as the pricing of subsequent nodes. We already knew that TSMC reportedly plans to charge up to $30,000 per #wafer processed using its N2 technology, but now #Taiwan-based China Times reports that the company will charge up to $45,000 per wafer for 'more advanced nodes,' which allegedly points to the company's A16 (1.6nm-class) node. "TSMC's wafer foundry price for 2nm chips has soared to $30,000 per wafer, with [more] advanced nodes rumored to reach up to $45,000," the report by China Times reads. This claimed price is 45% higher than the rumored cost of TSMC's N2 node, which is also rumored to have risen up to $30,000. Thanks again to Anton Shilov and Tom's Hardware for the full article with more background and insights via the link below 💡🙏👇 https://lnkd.in/eMzGz8kA #semiconductorindustry #semiconductors #technology #chip #chips #ic #iot #geopolitics #foundry #ai #nvidia #apple #amd #intel #innovation #it

211

8 Comments

Intel sells 51% of Altera FPGA business to Silver Lake for $4.46 billion Intel bought Altera in 2015 for $16.7 billion. Intel Corporation has reached an agreement to transfer a majority stake in its Altera division to Silver Lake for $4.46 billion, which values the business at $8.75 billion. The deal is part of Intel's effort to streamline its operations and improve its financial position, while making Altera the world's largest pure-play #FPGA provider. Under the terms of the agreement, Silver Lake will acquire 51% of Altera, while Intel retains 49%, ensuring it continues to benefit from the unit's future performance. The transaction enables Altera to function as an independent company focused on programmable logic technologies. Specifically, Altera aims to double down on established fields such as #automotive, #aerospace, and #communications, while also targeting growth in areas like #artificialintelligence, #cloud platforms, #edge systems, and next-generation #wireless networks. Intel, on the other hand, will reduce operational complexity and focus on its primary business areas: #CPUs, #GPUs, supporting platforms, and #chip production. Thanks again to Anton Shilov and Tom's Hardware for the full article with more background and insights via the link below 💡🙏👇 https://lnkd.in/eQQxFW6C #semiconductorindustry #foundry #semiconductormanufacturing #ic #ai

112

4 Comments

the Intel Corporation’s stock dropped after Broadcom’s comments about not pursuing a deal with the company. Broadcom had been rumoured to be interested in acquiring Intel’s #chipDesign division. In an earnings conference call, Broadcom’s CEO, Tan, addressed the rumours. He explained that the company is too busy focusing on #AI and #VMware. At this moment, they have no interest in considering any acquisition of Intel. Intel has faced growing competition and challenges, including production delays and the need for innovation. The potential bid from Broadcom had been seen as an opportunity for Intel to realign strategically. Broadcom’s clarification caused a negative reaction from investors. For a deeper understanding of the semiconductor industry and the forces shaping it, check out my new book, Semiconductor for All. It covers these developments and much more in detail. Get your copy now on Amazon Kindle! https://lnkd.in/gzVV2fJm For more updates and insights on AI and semiconductor, check out my blog. https://lnkd.in/gUmGrQh4 #semiconductor #ai #chips #manufacturing #semiconductorforall

25

1 Comment

TSMC is right on track in Phoenix supplying U.S. customers and collaborating seamlessly with President Trump to restore advanced semiconductor manufacturing domestically. The company is investing $165 billion to build six leading edge wafer fabs (the first is already operational fabricating chips with high yields), two packaging facilities, and a major R&D center. After completion, 30% of our 2 nanometer and more advanced capacity will be made onshore, creating an independent semiconductor manufacturing cluster. TSMC is honored to join hands with the Trump Administration and extend American AI leadership producing the fastest chips in the world in the great state of Arizona.

293

8 Comments

Big news from Micron today—this marks a pivotal moment in our mission to advance U.S. semiconductor leadership. We’re expanding our U.S. investments to approximately $200 billion, reinforcing our long-term commitment to domestic memory manufacturing and R&D. This includes: - A second leading-edge memory fab in Boise, Idaho—complementing the high-volume fab already under construction - Expansion and modernization of our Manassas, Virginia facility - Plans to bring advanced HBM packaging capabilities to the U.S. - A $50 billion investment in domestic R&D - Continued progress toward our planned megafab in New York These initiatives are expected to create around 90,000 direct and indirect jobs and will help us meet the surging demand for AI-driven innovation. As someone closely involved in our U.S. expansion strategy, I’m incredibly proud of what this means for our communities, our workforce, and the future of American-made memory. This is about more than just scale—it’s about building resilience, driving innovation, and securing leadership in a critical industry. Thank you to our federal, state, and local partners for your support—and to the Micron team, whose dedication makes this possible. Read more about the announcement: https://lnkd.in/gqGUNQRX See the scale of our Idaho construction: https://lnkd.in/g_Cd4FXA

1,649

46 Comments

Intel just peeled back the curtain on its Panther Lake (PC) and Clearwater Forest (Server) processors—the first on the critical 18A process node. Our analysis cuts through the marketing, discussing: * 18A Risk Mitigation--Intel's calculated defense against poor yields. * A second think of E-cores. * What really would make us confident in 18A. In contrast to everything else you've read, this is a pragmatic look at Intel's manufacturing gamble. https://bit.ly/46OHEe3

21

TSMC Reportedly Leads Talks on Intel Foundry JV with NVIDIA, AMD, and Broadcom Is a new U.S. chip alliance taking shape? According to Reuters, TSMC is in discussions with NVIDIA, AMD, Broadcom, and Qualcomm to invest in a joint venture managing Intel’s foundry operations. While TSMC would oversee the division, it would hold less than a 50% stake, with investors expected to be Intel’s advanced-node customers. Key Points: 🔹 NVIDIA and Broadcom are already testing Intel’s 18A process, while AMD is evaluating its viability. However, 18A remains a contentious topic—Intel has previously claimed it outperforms TSMC’s 2nm node. 🔹 Intel’s Panther Lake chips, set to use 18A, are still on track for H2 2025 production. 🔹 Intel reportedly outsources 30% of its wafers to TSMC, marking a shift from its prior goal of reducing reliance on external foundries. Any potential deal would face regulatory scrutiny to prevent full foreign ownership of Intel’s foundry. Additionally, challenges remain, including differences in manufacturing processes and concerns over trade secrets. With the semiconductor landscape rapidly evolving, could this JV redefine industry dynamics? Thanks again to TrendForce Corporation for the full article with more background and insights click the source link in the comments below. #Semiconductors #TSMC #Intel #Foundry #ChipManufacturing #ChipDesign,#SemiconductorManufacturing, #VLSI, #Lithography, #ICPackaging

265

13 Comments

As highlighted in Intel Corporation’s Q4 and 2025 earnings call yesterday, AI continues to drive demand across everything we do - reinforcing the critical importance of x86 and CPUs.   2025 was a solid year, but there is more work ahead. In 2026, demand for traditional servers remains strong, and we are focused on ramping available capacity to support our customers.   And, as we look at the immense opportunity with AI, our teams are fully focused on improving our position and delivering for our customers. We have clear priorities for 2026, and I am confident in the direction we are headed. https://lnkd.in/gs8EVUkz

581

12 Comments

𝗧𝗼 𝗘𝘃𝗲𝗿𝘆𝗼𝗻𝗲 𝗪𝗵𝗼 𝗕𝘂𝗶𝗹𝗱𝘀 𝗦𝗶𝗹𝗶𝗰𝗼𝗻 ➖ "𝗖𝗵𝗶𝗽 𝗪𝗮𝗿" 𝗜𝘀 𝘁𝗵𝗲 𝗦𝘁𝗼𝗿𝘆 𝗬𝗼𝘂 𝗖𝗮𝗻’𝘁 𝗔𝗳𝗳𝗼𝗿𝗱 𝘁𝗼 𝗠𝗶𝘀𝘀 I just finished Chris Miller’s Chip War, and it hit differently when you’ve spent years chasing timing closure, debugging DRCs, or optimizing for PPA. We often see our work as purely technical, but this book reframes it entirely. It reminds us that we’re not just designing circuits; we’re building the foundation of global power. 🌍 𝗦𝗶𝗹𝗶𝗰𝗼𝗻 𝗧𝗵𝗲 𝗡𝗲𝘄 𝗢𝗶𝗹 𝗼𝗳 𝘁𝗵𝗲 𝟮𝟭𝘀𝘁 𝗖𝗲𝗻𝘁𝘂𝗿𝘆 In a world increasingly reliant on technology, the importance of semiconductors cannot be overstated. Every smartphone, car, satellite, and defense system depends on these tiny pieces of silicon the invisible engines driving our digital age. After diving deep into Chris Miller's "Chip War," I've come to realize: whoever controls chip manufacturing controls the future. 𝗧𝗵𝗲 𝗠𝗼𝘀𝘁 𝗣𝗼𝘄𝗲𝗿𝗳𝘂𝗹 𝗣𝗮𝗿𝘁𝗻𝗲𝗿𝘀𝗵𝗶𝗽 𝗬𝗼𝘂'𝘃𝗲 𝗡𝗲𝘃𝗲𝗿 𝗛𝗲𝗮𝗿𝗱 𝗢𝗳 Two companies. Two countries (Netherlands & Taiwan). One irreplaceable partnership. Without ASML's $350 million EUV lithography machines the most complex devices ever built advanced chips simply cannot exist. Only one company manufactures them. Without TSMC's manufacturing mastery, those machines are useless. Only one company can turn silicon into 3nm miracles at scale. Here's the reality: Every AI breakthrough, every smartphone, every advanced weapon system passes through these two companies. Not Google. Not Apple. Not even America or China can bypass them. 90%+ of the world’s advanced logic chips (<7nm) come from a single company on a single island. the quiet powerhouse that fabricates world’s most advanced chips. Whether your design’s at 5nm, 3nm, or the upcoming 2nm GAA odds are, it’s going to TSMC. 𝗧𝗵𝗲 𝗜𝗻𝗱𝗶𝗮 𝗔𝗻𝗴𝗹𝗲 India’s story feels different this time: 𝗙𝗮𝗯𝘀 𝗮𝗿𝗲 𝗰𝗼𝗺𝗶𝗻𝗴 (Micron ATMP, Tower partnership, Tata proposals) 𝗗𝗲𝘀𝗶𝗴𝗻 𝗲𝗰𝗼𝘀𝘆𝘀𝘁𝗲𝗺 is already World Class 𝗧𝗮𝗹𝗲𝗻𝘁 𝗽𝗶𝗽𝗲𝗹𝗶𝗻𝗲 is deep we just need manufacturing scale This could redefine local tape outs, reduce timezone bottlenecks, and open global opportunities from Indian soil. India's immediate success is not about competing with TSMC at the 2nm edge. It is about establishing resilient domestic capacity in crucial mature and specialty nodes for automotive, power, communications and industry automation. which currently accounts for the bulk of India’s import demand while simultaneously investing heavily in the R&D and talent needed to close the advanced node gap over the next decade. 💭 I’d like to hear your thoughts. #ChipWar #SemiconductorIndustry #VLSI #TechGeopolitics #SemiconIndia #SiliconPower #ChipDesign #FutureOfTech

20

Not exciting, but important. AMD's published PyTorch Docker images have slimmed down -- from over 20GB for images based on ROCm 6.4 to around 9GB for images based on ROCm 7.0. Container size has been an ongoing source of feedback to AMD, and eliminating components that were redundant or needed only for our CI testing allows the same functionality in a much smaller artifact. While there's still room for improvement, this is a significant step. https://lnkd.in/gtAv3wTp

274

7 Comments

AI isn’t slowing down, and Broadcom continues expanding its optical interconnect solutions portfolio for next-gen AI infrastructure. Today, we’re proud to announce the expansion of our industry-leading 200G/lane DSP PHY portfolio with the introduction of Sian3 and Sian2M, purpose-built for the demanding connectivity requirements of AI/ML clusters. Learn more: https://dy.si/HM9MHj2

124

Update - NEW SILICON PARETO CURVE, EVP: It’s been 8 months since I first posted about the concept of EVP: "I've debated this problem for a while, and many people have looked at replacing the infamous PPA Pareto curve. I have experimented with adding interconnects (including the pins) to the equation, as well as 4-letter and 5-letter replacements. After playing with several alternatives, I think EVP is probably the most effortless future transition and more straightforward to pronounce than the alternatives.” ** ENERGY, E: because our community should care not only about Performance but the "cost" of Performance, i.e., energy-in and the heat-out. Power does not accurately reflect the cost. Yes, time is a function of Performance, but it is also a function of cost. I am concerned that the industry has lost its Moral Imperative on efficiency. Reduction of logic or geometry shrinkage is a minor factor in the efficiency game and not particularly useful (2025 update, giving us 3% to 5% per shift). ** VOLUME, V: Yes, we will be moving to physical 3D silicon, no longer concerned with 2D floorplans (very 1970s). We will choose computational cubes over pancakes. We will care more about making 3D circuits cool enough to drive future computational cubes that avoid thermal nuclear meltdowns. ** PERFORMANCE, P: does not change. Performance is the metric for a satisfied consumer. Again, linked to Energy, but how much is too much? How close can we push the modern computing world towards the Margolus-Levitin limit with the available Energy without destroying the world's resources in the process?" 2025 UPDATE: That was my initial thought on replacing PPA, but there is a more fundamental consideration as well: we need to escape Euclidean geometry (2D). Yes, FinFETs and GAA transistors are 3D, but the design of logic remains 2D.  Tensors, Vectors, and Quaternions (plus alternatives) aren’t just crucial for Artificial Intelligence; they will be essential for future silicon computational design. #VaireComputing #AdiabaticReversibleComputing #PhysicsAwareComputing #NearZeroEnergyChips #Physics #Semiconductor #Future #PPA #EVP

52

9 Comments