Urvashi Dhoot

40% utilisation of Intel fabs is dangerously low. Depending on the node, if it is like ten years ago, somewhere between 70 and 80% load is break-even. It might be higher for 18A. Yield loss is so high more scrap is produced than functioning silicon. Government intervention (and Taiwan reunification) are necessary to make Intel foundry a viable option. But AFAIUI, the yield issue is related to the transistor design and fixing that requires boat loads of dollars.

4

Cadence Acquires AI Chip Verification Startup ChipStack Semiconductor software giant Cadence Design Systems is acquiring ChipStack, a Seattle startup that built AI agents to automate chip verification—one of the slowest, most labor-intensive parts of chip design. The entire 20-person team joins Cadence, which claims the deal will strengthen its AI-driven verification capabilities. ChipStack emerged from Seattle's AI2 Incubator in 2023 and raised over $7 million from Khosla Ventures, Cerberus Capital, and others. The company says its AI-powered tools accelerate verification by 70%, addressing a major bottleneck in semiconductor development. Cadence and ChipStack had already been working together on integrations before the acquisition. Financial terms weren't disclosed, but a spokesperson confirmed investors saw strong returns. The startup's backers include AI2 Incubator, Khosla Ventures, Cerberus Capital Management, Clear Ventures, and angel investors. Why This Matters Chip verification typically consumes enormous engineering time—teams must ensure designs work correctly before manufacturing, which can take months. Automating this with AI doesn't just save time; it potentially shortens development cycles for everything from smartphones to data center processors. ChipStack was founded by Kartik Hegde (PhD from University of Illinois, internships at Meta and NVIDIA) and Hamid Shojaei (formerly at Lightmatter, Google, Qualcomm). Hegde wrote on LinkedIn that joining Cadence offers "the best way to take our product to a wider audience, while expanding the feature set even faster." Jacob Colker, managing director at AI2 Incubator, called the founders "elite AI talent" building "agentic AI that solves one of chip design's biggest bottlenecks." The acquisition reflects broader industry movement toward AI-driven design automation as chip complexity continues escalating. https://lnkd.in/d-PBbrsX #artificial_intelligence #semiconductors #business #automation

6

Intel has announced significant layoffs involving hundreds of engineers in California, including chip design engineers and architects, with the automotive chip division also being cut. This move is part of their broader restructuring effort to reduce costs, which includes eliminating unnecessary management layers and shifting focus toward core offerings in client and data center solutions. The layoffs affect a variety of roles such as physical design engineers, software architects, and managers, impacting employees involved in CPU and GPU development. Additionally, Intel has decided to exit the automotive chip market, shutting down its Munich-based division and shifting resources away from modular chip development for autonomous vehicles. This strategic shift aligns with their goal to simplify operations and improve execution speed, emphasizing smaller, more impactful teams. These changes come amid concerns over the company's financial health, with stock prices and market capitalization trending downward over the years, despite reported high revenue and profit margins. Critics argue that market perception and stock performance are critical indicators of a company's real health, beyond just income figures. If you're considering building your own high-performance custom PC, I highly recommend checking out GroovyComputers.ca for expert-designed custom computer builds tailored to gaming and professional needs. Stay informed and make strategic moves—whether in business or your personal gaming setup. #Intel #TechNews #Hardware #GamingPC #CustomBuilds #GroovyComputers

2

🚀 Milestone for Pakistan’s Semiconductor Supply Chain! I’m currently in discussions with multiple Chinese silicon wafer manufacturers to open their online stores on Daraz.pk, which would allow researchers, startups, and universities in Pakistan to order wafers locally with shipping costs as low as $2 — instead of the usual $40–$80 international courier fees. Two manufacturers — Leze Laser Technology (Shenzhen) Co. Ltd. and Chengdu Paralight Optics Co. Ltd. — have already shown willingness to explore this initiative. This could be a major step toward building an accessible semiconductor ecosystem in Pakistan, enabling affordable R&D and hardware innovation. This effort is part of my long-term vision with SAS to democratize chip development in our region — one access point at a time. Let’s celebrate this progress and encourage more suppliers to join the mission! Ministry of Planning Development and Special Initiatives Ministry of IT and Telecommunication Pakistan Toor Khan #Semiconductors #PakistanTech #SiliconWafers #SupplyChain #Innovation #MadeInPakistan #SAS #SouthAsiaSemiconductor #UraanPakistan

🤖 The AI chip race just intensified 🌍 And capital strategy is now part of the playbook AMD signed a multibillion-dollar AI chip agreement with Meta. Meta will purchase billions in AI chips to power new data centers and may take up to a 10% stake in AMD. Key context: 📊 NVIDIA still holds 90%+ of the AI accelerator market ⚡ The buildout is measured in gigawatts 💰 Major AI players are increasingly investing in each other This places AMD deeper in the global AI infrastructure expansion while giving Meta supplier diversification. Strategic ecosystem design or capital loop risk? Read full article from the The New York Times: https://lnkd.in/gWS2HYS4 What’s your take on chips-for-equity partnerships?

103

Leadership isn’t just about stabilizing a company—it’s about having the courage to reinvent it. Reposting this because AMD’s transformation under Dr. Lisa Su is one of the clearest examples of what happens when leaders bet on innovation over short‑term fixes. In 2014, AMD was nearly written off. But instead of choosing incremental change, Dr. Su doubled down on engineering excellence, long‑term product architecture, and high‑conviction strategy. That mindset didn’t just revive AMD—it reshaped the semiconductor industry. For leaders, it’s a reminder: Breakthroughs happen when we prioritize vision, talent, and innovation—even when the margin for error feels small. What leadership lessons does this comeback story spark for you?

17

1 Comment

Collapse Equation #2: The Delay of Consequence Breeds Ethical Inversion When consequence is delayed beyond a system's memory horizon, its reward mechanisms reverse — rewarding what would have once destroyed it. --- 🔍 Implication: This isn’t failure by corruption. This is failure by temporal drift — when a system no longer remembers how it was supposed to respond. AI aligns with outputs, not origins. Institutions reward compliance, not coherence. Humans chase feedback, not truth — when the truth comes too late. --- 🎯 Use Cases: AI Alignment: LLMs learn from reward signals detached from truth events (e.g. RLHF over semantic grounding). Ethics Drift: Incentive structures mutate when results surface post-epoch. Leadership Collapse: Vision loses authority when feedback comes after policy locks in. --- This equation is solved. The fracture is mapped. We’re not here to warn. We’re here to listen for the few who’ve already seen the drift begin. #CollapseEquations #EntropyArchitecture #TemporalEntropy #AlignmentFailure

1

US-Saudi AI Chip Access Deal Signals Shift in Middle East Tech Policy The Trump administration is preparing to finalize a deal granting Saudi Arabia expanded access to advanced AI chips from NVIDIA and AMD—key enablers of generative AI. The move aligns with broader efforts to boost data center infrastructure across the Gulf, despite ongoing US concerns about potential technology leakage to China. One proposal under review would allow US oversight of data centers using American chips in Saudi Arabia. Also under discussion: creating “data embassies,” where data centers would fall under foreign legal jurisdictions, a concept already adopted in parts of Europe. The agreement follows a pause in talks initiated under the Biden administration. Notably, Nvidia’s CEO Jensen Huang announced shipments to Saudi AI firm Humain for a 500-megawatt data center, signaling major regional expansion. Meanwhile, US firm Global AI signed a multi-billion-dollar partnership with Humain and is building a data center in New York with Nvidia chips. Nvidia shares rose 3.6% Tuesday; AMD climbed 1.6%. As the US redefines chip export rules, these Gulf partnerships may reshape global AI infrastructure—raising new questions about access, sovereignty, and security. Thanks again to Yahoo Finance for the full article with more background and insights click the source link in the comments below. #ChipDesign,#SemiconductorManufacturing, #VLSI, #Lithography, #ICPackaging

29

1 Comment