All articles tagged with #intel
Intel has joined SpaceX, xAI, and Tesla in the Terafab chipmaking project, aiming to refactor silicon fab technology in Austin to produce ultra-high‑performance chips for AI inference, edge computing, and possibly space‑based training, with a target of about 1 TW/year of compute; the deal helped lift Intel’s stock and signals a strategic shift as it chips away at AI‑chip leadership.
Ars Technica’s review finds Intel’s Core Ultra 270K Plus and 250K Plus offer strong multi‑core performance and efficiency for their price, outperforming some AMD rivals in CPU-heavy tasks, and the 250K Plus is singled out as the better value. However, the overall system cost remains driven up by current RAM/SSD/GPU prices, and the LGA 1851 socket provides no upgrade path, making these CPUs a hard sell for budget-conscious builders who must also factor memory costs. Gaming performance lags AMD’s non‑X3D chips, though the chips stay cooler and more power-efficient than many predecessors.
Intel’s Core Ultra 7 270K Plus and Core Ultra 5 250K Plus bring more cores, higher clocks, and improved efficiency with DDR5 memory support and Binary Optimization Technology. They deliver strong multi‑threaded and content‑creation performance, compete well on efficiency, and offer solid value in gaming where titles vary. The LGA‑1851 platform remains, with no new motherboard necessary beyond a BIOS update; the 270K Plus targets demanding workloads (starting around $299) and the 250K Plus sits in the upper mainstream (starting around $199).
Intel's Core Ultra 5 250K Plus launches at $200 as part of the Arrow Lake Refresh Plus lineup, packing 6 P-cores + 12 E-cores, 30 MB L3 cache, and a higher 3.0 GHz die-to-die interconnect. The Plus introduces Intel Binary Optimization Technology (via IPPP), an opt-in runtime code modification aimed at optimizing game binaries for Intel architecture with claimed up to ~20% performance gains in ideal cases. Built on Arrow Lake’s disaggregated tile architecture (Compute tile on 3 nm, memory/NPU/PCIe tile on 6 nm, graphics on 5 nm) and keeping 125 W base/159 W turbo, it competes in the budget segment against AMD’s Ryzen 5 9600X and Intel’s i5-14600K. The update prioritizes core-count and cache gains alongside software-driven optimizations to deliver improved value and frame rates.
Intel’s Core Ultra 9 290HX Plus and Core Ultra 7 270HX Plus bring modest performance gains (FPS boosts vary by game: ~2% in Red Dead Redemption 2, ~10% Cyberpunk 2077, up to ~24% Borderlands 3) to 2026 gaming laptops, but the standout upgrade is brighter OLED and other high-refresh displays on models like Alienware’s Area-51. That glory comes with a price: base 16-inch Area-51 LCD starts around $3,150, while an OLED version with RTX 5090 and 290HX Plus can hit about $4,900; Asus and Acer are refreshing with higher-refresh OLED/mini-LED panels as costs climb. Availability spans 2026 with updates announced for Q2 and beyond, signaling premium pricing for incremental gains.
Intel’s Core Ultra 200S Plus (Arrow Lake Refresh) adds more cores, higher clock speeds, faster memory support, and improved inter‑chip communication to boost gaming performance by about 15%. The lineup includes 24‑core 270K Plus/270KF Plus (8 P‑cores, 16 E‑cores) at $299 and 6 P‑cores/12 E‑cores 250K/250KF Plus at $199, with prices roughly aligned to predecessors. They use the LGA 1851 socket and 800‑series chipsets and arrive March 26. Intel also touts the Binary Optimization Tool to enhance performance for games not optimized for newer CPUs.
Intel unveils the Core Ultra 7 270K Plus and 250K Plus as its fastest gaming desktops yet, claiming the 24‑core 5.5 GHz Turbo 270K Plus can beat the i9-14900K and Arrow Lake 285K in gaming for a $300 price, while the 250K Plus targets budget buyers at $200 with strong multicore gains. The processors add four efficiency cores, higher base clocks, DDR5-7200 support, early 4‑rank memory compatibility, and a Binary Optimization Tool that boosts select games, all while maintaining 125 W TDP and motherboard compatibility with 800‑series boards; shipping March 26.
Intel and SambaNova Systems agreed to a multiyear technical partnership to build a cloud AI service, with Intel contributing to SambaNova’s $350 million funding round and its Xeon chips powering SambaNova systems, as part of a broader push to challenge Nvidia in AI hardware; Intel’s CEO Lip-Bu Tan recused himself from SambaNova talks due to potential conflicts, with negotiations overseen by Kevork Kechichian under established governance. SambaNova remains independent, CEO Rodrigo Liang noting strong demand supports continued growth.
Intel's Nova Lake processor line is delayed to after CES 2027, causing about a 2% drop in INTC shares. Some Nova Lake variants may release this year, but the full Nova Lake-S line won’t appear until CES 2027. Analysts maintain a Hold rating with roughly 11% upside to a target around $48.21, even as Oregon faces concerns about Intel's long-term impact in the Silicon Forest amid ongoing job cuts. The delay highlights ongoing supply-demand tightness in the chip industry amid AI-driven demand.
Intel is reportedly moving away from a split P-Core/E-Core hybrid toward a unified core design for its next-generation CPUs, with a dedicated design team and an unclear launch timeline, potentially impacting both consumer and Xeon lines and leaving room for new differentiation techniques.
Nvidia is expanding from data-center AI chips into consumer laptops, teaming with MediaTek on ARM-based system-on-a-chip designs and with Intel to pair processors with Nvidia’s graphics and AI tools for laptops from Dell, Lenovo, and HP. The goal is a 150 million annual laptop market, emphasizing thin, long-lasting AI-enabled machines rather than immediate profits, with price points likely in the $1,000–$1,500 range and a focus on long-term ecosystem integration; NVDA stock rose modestly ahead of earnings.
A leak suggests Intel's Nova Lake-S Core Ultra CPUs will launch at CES 2027, potentially alongside AMD's Zen 6 “Olympic Ridge,” pushing back from a late-2026 window; early rumors indicate Nova Lake-S could feature up to 52 cores (16 P-cores, 32 E-cores, 4 LP-cores) with AI improvements, but specifics remain unconfirmed amid ongoing silicon and DRAM shortages.
Nvidia will deploy its Arm-based Grace CPUs in Meta’s data centers, signaling a broader shift toward Arm in the data-center market and putting pressure on Intel, with analysts predicting more Arm-based and custom-silicon deals in the near term.
Micron Technology is riding an AI‑driven upcycle in high‑bandwidth memory that could last longer than past cycles, fueling debate over whether it will resemble Nvidia or Intel. While competition from Samsung and SK Hynix and memory cycles pose risks, the article argues that sustained AI demand for Micron's HBM could keep the growth alive—potentially making Micron more Nvidia‑like than Intel, at least for now.
Notebookcheck’s early tests of Intel’s Panther Lake Core Ultra 5 338H with Arc B370 (10 Xe cores) show solid CPU performance comparable to Ryzen AI 9 465 at similar TDP, while the Arc B370 GPU delivers a notable uplift but trails the Arc B390 by a small margin. At 20W the B370 is virtually identical to the B390, and at 35W the gap grows to about 6%. The unit tested was an engineering sample running at 35W, with retail units expected in the coming weeks.