News World That's Worth

OpenAI Reveals Test Results for Custom 'Jalapeño' AI Chip Developed with Broadcom and Celestica OpenAI has officially disclosed performance benchmarks for Jalapeño , its custom artificial intelligence chip co-designed in partnership with Broadcom and Celestica . Initially unveiled in June without granular technical specifications, new data showcases significant operational efficiency gains in handling large-scale inference workloads. According to Richard Ho , Vice President of Hardware at OpenAI, the defining architectural advantage of Jalapeño is its ability to break the traditional engineering trade-off between throughput (processing volume per unit) and latency (response speed). Typically, AI chips must sacrifice execution speed to maximize processing capacity, or vice versa. To evaluate real-world performance, OpenAI utilized the InferenceX benchmark suite across three open-source AI models: GPT-OSS 120B , DeepSeek R1 , and Kimi K2.5 1T . Benchmark results demonstrated...

TSMC Reportedly Raising Wafer Prices by Up to 10% in 2027 Driven by Surging Raw Material and Equipment Costs Pure-play foundry leader TSMC is reportedly planning a broad-based price hike of up to 10% across its semiconductor manufacturing nodes starting in 2027 , according to industry sources cited by Nikkei Asia . The price adjustments will impact both leading-edge processes (nodes under 6nm) and mature trailing nodes, including 12nm, 16nm, and 28nm technologies. Industry insiders attribute the upcoming price escalation directly to rising input costs, specifically the soaring prices of specialized raw materials and next-generation lithography equipment required for advanced chipmaking. While TSMC maintains a strict corporate policy against disclosing specific client pricing structures, the foundry has repeatedly emphasized that its commercial strategy is not based on opportunistic short-term surge pricing. TSMC CEO C.C. Wei previously affirmed the company's commitment to long-te...

OpenAI Unveils Custom 'Jalapeño' AI ASIC Co-Designed with Broadcom and Celestica to Slash Compute Costs In a major stride toward hardware independence, OpenAI has officially introduced its proprietary artificial intelligence microchip, codenamed Jalapeño . Developed through an elite, multi-party engineering alliance with semiconductor titan Broadcom and manufacturing solutions provider Celestica , Jalapeño is an Application-Specific Integrated Circuit ( ASIC ) custom-tailored exclusively to power OpenAI's specialized generative AI workloads and large language model inference frameworks. While OpenAI kept the chip's raw compute benchmarks under wraps during the unveiling, the firm highlighted that Jalapeño’s internal architecture was engineered to strike an optimal structural equilibrium between matrix math processing units, memory sub-systems, and ultra-high-speed networking fabrics . By optimizing these specific interconnections according to OpenAI's real-world o...

The Silicon Elite: Semiconductor Degrees Rival Medical School Admission Scores in South Korea Amid AI Chip Boom In a striking cultural and academic shift, South Korea's prestigious corporate-sponsored Semiconductor Engineering degrees have risen to near-parity with the country’s traditionally untouchable medical school tracks. According to a comprehensive data analysis of this year’s university admissions cycle released by Jongro Academy , a premier educational consulting institution, minimum cut-off scores for elite semiconductor programs have surged to within a razor-thin 1-point margin behind top-tier medicine faculties. The explosive demand is heavily concentrated in specialized university departments operating in direct, bilateral partnerships with global memory juggernauts SK hynix and Samsung Electronics . An institutional breakdown of this year's highly competitive minimum admission cut-offs highlights the rising prestige of the tech sector: Hanyang University: Dep...

Intel Foundry Appoints Dr. Seok-Hee Lee to Lead Advanced Packaging as Apple Secures Landmark U.S. Manufacturing Deal In a major strategic restructuring of its semiconductor fabrication division, Intel has officially appointed industry veteran Dr. Seok-Hee Lee as the new executive leader of Intel Foundry . Reporting directly to Intel CEO Lip-Bu Tan , Dr. Lee will take absolute command of Intel’s advanced packaging, system integration, and backend technology development ecosystems a portfolio critical to scaling operations and meeting the rigorous, next-generation demands of global foundry clients. Dr. Lee brings an elite corporate pedigree back to Intel. He previously served as the CEO of EV battery innovator SK On and memory giant SK hynix , solidifying his reputation as one of the world's foremost semiconductor operations experts. Notably, this appointment marks a historical homecoming, as Dr. Lee spent his formative engineering years inside Intel’s research labs. "Joining...

Huawei Unveils "LogicFolding" and Tau Scaling Law to Bypass Moore’s Law Limitations In a major technical presentation at the IEEE International Symposium on Circuits and Systems (ISCAS), Huawei introduced a radical paradigm shift in semiconductor design labeled "LogicFolding." The newly unveiled framework is engineered specifically to counteract the stagnation of Moore's Law, which historically dictated that the number of transistors on a microchip doubles roughly every two years. As physical limits make traditional silicon shrinking increasingly unviable, Huawei’s new design methodology pivots away from pure transistor density. Instead, it focuses heavily on minimizing data propagation delay a principle the tech giant defines as "Tau ( τ ) Scaling Law." The Four Pillars of Tau Scaling Law The Tau Scaling Law framework optimizes chip performance across four interconnected architectural layers: Device Level: Re-engineering physical components to dras...

India Semiconductor Leap: Tata Electronics Partners with Dutch Giant ASML for Historic $11 Billion Dholera Fab In a monumental stride toward tech sovereignty, India Tata Electronics has signed a strategic Memorandum of Understanding (MoU) with Dutch lithography titan ASML . The partnership aims to build and scale India's very first commercial front-end semiconductor fabrication facility (fab) in the Dholera Smart City of Gujarat, a western coastal state. Under the agreement, ASML will supply its highly specialized, cutting-edge lithography systems and advanced software solutions. These tools are critical to successfully bootstrapping operations for the upcoming 300 mm (12-inch) silicon wafer fab , firmly placing India on the global hardware manufacturing map. An $11 Billion Blueprint for Global Supply Chains Tata Electronics has earmarked a massive $11 billion investment to construct this state-of-the-art facility. The pure-play foundry is designed to mass-produce analog and logi...

Amazon Unleashes $200B Investment Surge: Andy Jassy Bets Big on AI Infrastructure and Custom Silicon In his annual letter to shareholders, Amazon CEO Andy Jassy outlined a bold, high-stakes roadmap for the company’s future. Despite a sharp decline in free cash flow, Jassy defended the company’s aggressive spending, revealing a massive $200 billion investment plan for the year, with a primary focus on scaling AI infrastructure. The AI Revenue Engine The justification for this heavy capital expenditure lies in the explosive growth of Amazon's AI services. Jassy noted that this segment alone has already achieved an annualized revenue run rate of $15 billion . To sustain this momentum, Amazon is prioritizing the development of data centers and specialized hardware. The Hidden Semiconductor Giant One of the most striking revelations in the letter was the scale of AWS’s custom chip business . Jassy highlighted that if AWS’s chip division (responsible for Graviton, Trainium, and Nitro ...

Intel Joins Elon Musk’s "Terafab" Initiative: A Strategic Alliance to Power the Next Generation of AI and Robotics Intel has officially announced its participation in Terafab , the ambitious semiconductor mega-factory project led by Elon Musk . Intel joins an elite consortium that already includes Musk’s core ventures: Tesla , SpaceX , and xAI . Scaling to 1 Terawatt of Compute Intel stated that its contribution will be pivotal in accelerating the development of the Terafab project. By leveraging its world-class expertise in chip design , advanced manufacturing (Foundry) , and large-scale chip packaging , Intel aims to help the project reach its staggering production goal: 1 Terawatt (TW) of computing power annually . This massive output is specifically earmarked for advanced Artificial Intelligence and humanoid robotics. The partnership was solidified following a high-profile visit by Elon Musk to Intel’s headquarters last weekend, where the two parties finalized the strate...

Raspberry Pi Announces Second Price Hike for 2026 Amid Surging Memory Costs Raspberry Pi has officially announced another round of price increases across its high-performance product lines. Following a previous hike earlier this year, the foundation cites the continuing global surge in RAM prices as the primary driver behind this adjustment. The price changes primarily affect models with 4GB of RAM or higher , including the latest flagship boards and industrial modules. Affected Products and New Pricing The price adjustments vary by model and memory capacity, with higher-spec versions seeing the most significant increases: Raspberry Pi 4 & 5: The 4GB model increases by $25 , the 8GB model by $50 , and the top-tier 16GB (exclusive to Pi 5) jumps by $100 . Raspberry Pi 500 Series: Both the standard 500 and the advanced 500+ (standalone and kits) see increases of $50 and $150 , respectively. Compute Modules (CM4, CM4S, CM5): Prices have climbed across the board, ranging from an...

  TAALAS HC1: The AI-on-Silicon Breakthrough Delivering 17,000 Tokens per Second The Canadian chip design startup TAALAS has demonstrated a stunning leap in AI inference technology. Their new HC1 chip features the Llama 3.1 8B model etched directly into its silicon a "hardwired" approach that sacrifices modifiability for unprecedented speed and efficiency. Mind-Blowing Speed: Instant Results By embedding the model directly into the hardware, the HC1 achieves an astonishing inference speed of 16,960 tokens per second . At this rate, entire pages of text are generated almost instantaneously. To achieve this, TAALAS utilized a 3-bit quantized version of the Llama 3.1 8B model. While this compression results in a slight trade-off in output quality compared to the full-precision model, the performance gain is transformative. The Power of "Hardwired" AI The HC1 chip is a massive engineering feat, boasting 53 billion transistors and a power draw of 2.5kW . Although th...

Congress Eyes Veto Power Over AI Chip Exports: The "AI Overwatch Act" Aims to Counter Trump’s Trade Deals A bipartisan group of U.S. Senators, led by Mark Warner (D-VA) and Thom Tillis (R-NC) , has introduced a significant piece of legislation titled the " AI Overwatch Act " This bill seeks to grant Congress the authority to veto any executive branch decision that attempts to ease export controls on high-performance AI chips to high-risk nations. National Security vs. Trade Leverage

The Global Memory Chip Crisis: How the AI Boom is Starving Other Industries The tech world is facing a severe supply crunch in memory chips, driven by an insatiable appetite from AI data centers. Leading tech giants are now willing to pay exorbitant premiums for high-performance memory, forcing manufacturers to shift their production lines away from consumer staples. As a result, prices for basic components like RAM, Flash memory, and SSDs have skyrocketed, with the upward trend expected to persist throughout the current quarter.