Tag: Intel

  • US Pressures TSMC to Acquire 49% of Intel for Tariff Relief

    US Pressures TSMC to Acquire 49% of Intel for Tariff Relief

    Key Takeaways

    1. The US government, under President Trump, has used trade tariffs inconsistently to address the trade deficit, significantly impacting Taiwan as a trading partner.

    2. Taiwan faces a 20% tariff on exports to the US, higher than Japan’s 15%, prompting negotiations for potential tariff reductions.

    3. TSMC is under pressure from Trump to meet two strict conditions for tariff relief, which includes a substantial financial investment and a stake in Intel.

    4. Intel is struggling financially, with a 33% revenue drop from 2021 to 2024, raising concerns about its ability to meet the US’s semiconductor manufacturing goals.

    5. TSMC’s potential reluctance to comply with US demands regarding Intel could affect the future of both companies and the US chip supply chain.

    The US government, under President Donald Trump, has primarily utilized trade tariffs to tackle the nation’s trade deficit with its partners. However, this tariff strategy has been inconsistent and often unexpected. Taiwan, a significant trading ally of the US, has also felt the impact of these tariffs.

    Tariff Rates and Their Implications

    With tariffs set at 20%, Taiwan faces a higher rate than nations like Japan, which has a baseline of 15%. Unsurprisingly, this 20% tax on Taiwanese exports to the US can severely impact Taiwanese companies. Therefore, Taiwan is negotiating with the US to lower these tariffs or potentially remove them completely. Nevertheless, a recent report from Taiwan indicates that President Trump has laid down two tough conditions for any tariff relief.

    According to mnews.tw, an industry insider revealed that Trump has required TSMC to meet these two strict conditions in exchange for tariff reductions. Ignoring the implications of TSMC owning 49% of Intel, these requirements are significant financial obligations. TSMC is already pouring a considerable amount into the US, with one of its fabrication plants expected to begin volume production by 2024. The company is also in the process of constructing two additional fabs in Arizona, along with an R&D center and a packaging facility. In total, TSMC has committed to investing $165 billion in the US.

    Financial Challenges Ahead

    Trump’s supposed requirement for TSMC to invest an additional $400 billion in the US while acquiring a share in Intel appears unrealistic from a financial perspective. Intel is experiencing a downward trend. All its sectors, from fabrication to consumer goods, are facing challenges. This decline is reflected in its reported annual revenue, which has dropped dramatically from $79 billion in 2021 to $53 billion in 2024, marking a 33% decrease.

    Intel plays a crucial role in the US strategy for domestic semiconductor manufacturing. The company has already secured billions in federal funding. However, this financial aid hasn’t resolved Intel’s issues, as the timeline for its Ohio fab has been pushed back from 2025 to between 2030 and 2031. This delay appears to be tied to Intel’s attempts to conserve capital in light of insufficient support from the US CHIPS Act and external partners.

    The Future of TSMC and Intel

    In summary, it seems that the US administration is pressuring TSMC to take a significant stake in Intel to provide much-needed capital and sustain the US’s domestic chip supply chain ambitions. However, TSMC might not agree to this condition.

    It will be interesting to watch Intel in the next few months. There are some promising products in development, like the Panther Lake and Nova Lake CPUs. If these products perform well, it could lead to improvements for Team Blue.

    Source:
    Link

     

    Tags: , ,
  • Intel Panther Lake: Revolutionizing Handheld Gaming Experience

    Intel Panther Lake: Revolutionizing Handheld Gaming Experience

    Key Takeaways

    1. Panther Lake is Intel’s first mobile CPU using the 18A manufacturing process, aimed at competing with TSMC’s advanced 2nm technology.
    2. The Panther Lake chips may feature 12 Xe3 graphics cores, a 50% increase compared to the Lunar Lake series, promising enhanced graphics performance.
    3. The improved integrated GPU (iGPU) is expected to greatly enhance gaming experiences, particularly for handheld devices.
    4. Panther Lake is anticipated to support AI upscaling, potentially providing superior gaming performance compared to current handhelds.
    5. The release of Panther Lake is expected in late 2025, with potential delays pushing notebook launches to 2026.

    Panther Lake is the upcoming mobile CPU from Intel, marking a significant step as the first chip utilizing the company’s 18A manufacturing process. This development is crucial for Intel, as it aims to demonstrate its competitiveness against TSMC, which is known for its advanced 2nm processes. Although real-world performance metrics are not yet available, recent leaks suggest that the improved integrated GPU (iGPU) in the Panther Lake chips could greatly enhance graphics capabilities, making them ideal for gaming handhelds.

    Leaked Information

    The details come from a supposed shipping manifest shared by the user X86 is dead&back on X. According to the document, some Panther Lake chips are equipped with 12 Xe3 graphics cores. In contrast, the Lunar Lake series only has eight Xe2 cores. This means that Panther Lake offers a 50% increase in cores, and since Xe3 cores are enhancements over the Xe2, each core is expected to deliver better performance.

    Gaming Potential

    If properly integrated into handheld devices, this could result in remarkable gaming experiences. Additionally, at Computex 2025, Intel showcased its Panther Lake chips, demonstrating their ability to render graphics in real-time and run AI applications. This suggests that the iGPU may have support for AI upscaling, a feature that appears to be exclusive to the Panther Lake series for now. A handheld powered by a Panther Lake chip with a robust iGPU and AI-upscaling capabilities could theoretically achieve gaming performance that surpasses what current generation handhelds can provide.

    Panther Lake is expected to be released in the latter half of 2025. However, earlier reports indicated that the launch of Panther Lake notebooks has been pushed to 2026 due to delays in mass production, moving from early September 2025 to mid-Q4 2025. Nevertheless, this still aligns with Intel’s original timeline for production in the second half of 2025.

    Conclusion

    The information comes from the NBD database, as noted by the user X86 is dead&back on X, and was also referenced by Tom’s Hardware.

    Source:
    Link

    Tags: ,
  • Arrow Lake Refresh: More Disappointing Than Anticipated

    Arrow Lake Refresh: More Disappointing Than Anticipated

    Key Takeaways

    1. Intel’s new CEO, Lip-Bu Tan, is implementing operational changes to improve the company’s performance.
    2. The Arrow Lake Refresh CPUs are expected to have quicker clock speeds but will continue using the outdated NPU 3, rather than upgrading to NPU 4.
    3. The updates for the Arrow Lake Refresh are minor, with no changes to core counts and only specific SKUs (‘K’ and ‘KF’) expected to receive enhancements.
    4. Intel will not introduce a new series name for the Arrow Lake Refresh; it will remain within the Core Ultra 200 series.
    5. The exact increase in clock speeds for the refreshed CPUs is unclear, with the launch expected in the latter half of the year.

    Intel has been facing challenges for some time, but new CEO Lip-Bu Tan is making efforts to improve the situation. Significant changes in operations have occurred since his leadership began, impacting future product launches. The company plans to update its Arrow Lake range of CPUs, but if the latest leaks are correct, the upgrades may be less significant than what was initially anticipated.

    Upcoming Changes

    A report from early July indicated that the Arrow Lake Refresh CPUs would feature quicker clock speeds and an updated NPU. The Arrow Lake chips, which include the Core Ultra 200S desktop CPUs, currently utilize NPU 3—an older version that debuted with Meteor Lake chips in late 2023. This particular NPU has a performance capability of just 11.5 TOPS, which makes it ineligible for Copilot+ certification. The report suggested that the Arrow Lake Refresh would introduce NPU 4, similar to the Lunar Lake chips, which can achieve 48 TOPS. However, it seems that Intel will continue using NPU 3 for the refresh instead.

    Leaker Insights

    Known leaker Jaykihn has been providing updates about the Arrow Lake Refresh for some time. In his most recent post, responding to one of his earlier insights, he claims neither the desktop nor the high-end laptop CPU SKUs will see any changes to the NPU. This implies that the new CPUs will merely receive a boost in clock speeds, resulting in yet another minor update. It also indicates that the core counts will likely stay the same.

    Naming and Specifications

    Regarding naming conventions, Intel is not expected to introduce a new series for the Arrow Lake Refresh, meaning it will still fall under the Core Ultra 200 series. Additionally, it’s disappointing that not all SKUs in the lineup will be updated; only the ‘K’ and ‘KF’ SKUs are anticipated to receive the special enhancements.

    The exact increase in clock speeds remains uncertain. For context, the flagship Arrow Lake processor, Core Ultra 9 285K, has a boost clock of 5.5 GHz, which is lower than that of the 14th Gen Core i7 and Core i9 processors. Nonetheless, Intel is projected to unveil the refreshed lineup sometime in the latter half of this year, although a specific date has yet to be confirmed.

    Source:
    Link

    Tags: , ,
  • Intel Titan Lake 2028: No More P-Cores, All-E-Core Design

    Intel Titan Lake 2028: No More P-Cores, All-E-Core Design

    Key Takeaways

    1. Intel plans to launch the Titan Lake series in 2028, featuring an impressive 100-core design focused on a single-core architecture based entirely on E-cores.

    2. The Razer Lake series, arriving in 2027, will be the last to use a mixed architecture of P-cores and E-cores, leading to a shift in Intel’s core design philosophy.

    3. Titan Lake’s all-E-core setup may include variations like dense 4C clusters and 2C clusters, optimizing performance-per-area (PPA) and performance-per-Watt (PPW).

    4. Nova Lake is expected to have a mixed core setup with 52 cores, while Titan Lake could potentially consist of two 48-core clusters and four low-power E-cores (LPE).

    5. The trend in the CPU/SoC market is shifting towards unified core designs, as seen with competitors like AMD and Qualcomm, which could influence Intel’s future architectures.

    Rumors about Intel’s future Nova Lake and Panther Lake architectures have started to emerge, and new information regarding the company’s plans for 2028 is also coming to light.

    Titan Lake’s Exciting Features

    If the whispers are correct, Intel is set to launch the Titan Lake series in 2028 featuring an astonishing…100 cores! Yes, you heard that right—one hundred cores! Unlike the current designs that use both P-cores and E-cores, Intel seems to be shifting towards a single-core architecture for this new lineup.

    This scoop comes from @Silicon_fly on X, referencing a leaked roadmap found on Zhihu.com that outlines Intel’s processor plans through 2028. The roadmap indicates that Razer Lake, arriving in 2027, will be the last to utilize the mixed P-core and E-core setup. Razer Lake will include Griffin Cove P-cores and Golden Eagle E-cores, serving as a small upgrade to Nova Lake, which is expected to launch next year.

    A New Direction for Core Design

    With the Titan Lake series, Intel appears to be moving towards an all-unified core design. Interestingly, the new cores will not include P-cores; rather, they will be entirely based on E-cores, likely evolved from the larger Arctic Wolf E-cores found in Nova Lake. This change could lead to improved performance-per-area (PPA) and performance-per-Watt (PPW).

    Any increase in die size and power consumption from using Arctic Wolf E-cores can be balanced out by enhanced PPA achieved through switching to a 14A process, without significantly raising the thermal design power (TDP). However, the new unified E-cores would still be more compact than the Coyote Cove and Griffin Cove P-cores used in Nova Lake and Razer Lake, respectively.

    Potential Variations in Core Clusters

    Although a unified core suggests an all-E-core setup, there could still be some differences within the architecture. @Silicon_fly theorizes that Intel might implement a mix of dense 4C clusters with shared L2 caches and 2C clusters that either share or have dedicated L2/L3 caches. Additionally, Titan Lake might incorporate a specialized core aimed at enhancing single-core performance.

    Given that Nova Lake is projected to have a 52-core setup featuring 16 P-cores, 32 E-cores, and 4 low-power E-cores (LPE), if Titan Lake truly abandons the large P-cores, it’s conceivable to see a 100-core design consisting of two 48-core clusters along with four LPE cores.

    All this remains speculative, of course, and there are still several generations to go before Titan Lake becomes a reality, so it’s wise to take this information lightly for now.

    Still, the details align with the current trends we observe in the CPU/SoC market. AMD has already adopted a mix of Zen 5 classic and Zen 5c compact cores in its Ryzen Strix Point APUs. MediaTek has shifted to an all-big core design, starting with the Dimensity 8400, which features a 1+3+4 configuration without efficiency cores. The flagship Dimensity 9400 SoC also utilizes a big core cluster in a similar fashion.

    Furthermore, Qualcomm has opted for a 2 Prime + 6 Performance all-Oryon architecture for its Snapdragon 8 Elite. It remains uncertain what Apple has planned for its upcoming A and M series chips, but the industry seems to be leaning more towards unified core clusters designed for various tasks rather than mixed architectures.

    Source:
    link

    Intel Titan Lake 2028: No More P-Cores, All-E-Core Design
    Tags: , ,
  • Intel to Cut 2,392 Jobs in Oregon by Mid-July for Cost Savings

    Intel to Cut 2,392 Jobs in Oregon by Mid-July for Cost Savings

    Key Takeaways

    – Intel plans to cut 2,392 jobs in Oregon, significantly higher than the previously announced 500 layoffs, as part of a cost-reduction strategy under CEO Lip-Bu Tan.
    – The job losses will negatively affect Oregon’s economy, where Intel is the largest employer, potentially impacting local businesses and income-tax revenue.
    – Intel is struggling to compete in the semiconductor market, trailing behind TSMC and missing opportunities in key areas like AI training chips.
    – The layoffs will impact all levels within the company, particularly in Intel Foundry, with significant cuts to technicians and process engineers, and the disbanding of the automotive division.
    – Laid-off employees will receive severance pay and healthcare benefits, but Intel’s future operations in Oregon remain uncertain despite ongoing tax incentives.

    Intel has informed Oregon officials that it plans to cut 2,392 jobs in the state by mid-July, a significant increase from the earlier figure of just over 500 layoffs. This decision is part of a larger effort to reduce costs under the new CEO, Lip-Bu Tan, who is reacting to decreased sales and ongoing manufacturing problems.

    Economic Impact on Oregon

    Oregon is home to Intel’s largest facility globally, employing around 20,000 people. The semiconductor industry in the state typically pays an average salary of nearly $180,000 per year. Consequently, such job losses are likely to have a negative effect on local businesses and decrease income-tax revenue. State economist Carl Riccadonna has already cautioned lawmakers about a weakening job market.

    Competitive Struggles

    Intel’s current difficulties can be traced back almost a decade, when delays in its technology allowed TSMC to take the lead. Competitors like AMD, Nvidia, and Apple now depend on TSMC’s advanced technology, while Intel fights to reclaim its position and is mostly missing out on thriving markets like AI training chips. “In terms of training, I fear it may be too late for us,” Tan mentioned to employees, noting that Nvidia’s hold is “too strong.”

    Specific Job Reductions

    The workforce cuts affect every level within the company, but they are particularly severe in Intel Foundry, which focuses on manufacturing and research and development. About one in five positions in this area will be eliminated. Many of the layoffs will impact technicians and process engineers, though around eight percent of those let go hold management roles. In addition to manufacturing cuts, Intel has disbanded its automotive division and shifted most marketing responsibilities to Accenture.

    Laid-off employees will receive 13 weeks of their base salary plus an additional 1.5 weeks for each year they have worked, along with a year’s worth of healthcare benefits. Even though Intel still benefits from over $300 million in annual tax incentives from Oregon, its reduced revenues and changing priorities make the future scale of its operations in the Pacific Northwest uncertain.

    Source:
    Link

    Tags: , ,
  • Intel Arrow Lake Refresh CPUs: Faster Clock Speeds for Desktops

    Intel Arrow Lake Refresh CPUs: Faster Clock Speeds for Desktops

    Key Takeaways

    1. Launch Timeline: Intel’s Arrow Lake Refresh CPUs are expected to be released in the second half of 2025.

    2. Performance Improvements: The new CPUs will feature higher clock speeds and will include NPU 4, though specific clock speed details are not yet available.

    3. Comparison to Previous Generations: The flagship Core Ultra 9 285K has a boost clock of 5.5 GHz, which needs to increase by over 8% to match the performance of the 14th-gen Core i9-14900K.

    4. NPU Integration: The Arrow Lake Refresh will incorporate NPU 4, which has seen limited practical applications on Windows, raising questions about its utility in desktop CPUs.

    5. Gaming Landscape: The second half of 2025 is shaping up to be exciting for gamers, with new Intel CPUs and anticipated launches of desktop GPUs from AMD and Nvidia.

    While Intel is gearing up to introduce Nova Lake as the next significant CPU architecture for desktops, the company is also working on the Arrow Lake Refresh chips. Similar to previous updates, the Core Ultra 200 Arrow Lake Refresh CPUs are not anticipated to deliver drastic changes. Nevertheless, there are certain important aspects where Intel aims to enhance performance.

    Upcoming Release

    As reported by ZDNET Korea (via @harukaze5719 on X), the Intel Arrow Lake Refresh CPUs are set to launch sometime in H2 2025, featuring quicker clock speeds and NPU 4. The specifics regarding the improvement in clock speeds for the Arrow Lake Refresh chips remain unclear, as the report does not provide any numerical details.

    Key Specifications

    To put it into perspective, the Core Ultra 9 285K, which is the leading Arrow Lake processor, has a boost clock of 5.5 GHz. This represents a significant reduction compared to the 14th-gen Core i9-14900K (Available on Amazon). In order for the Arrow Lake Refresh flagship to match the 14th-gen models, it will need to achieve a boost clock increase of over 8%.

    NPU Enhancements

    Currently, the Core Ultra 200S desktop CPUs utilize NPU 3, an earlier version that Intel first rolled out with Meteor Lake chips in late 2023. Intel later integrated the much-improved NPU 4 into the Core Ultra 200V Lunar Lake mobile chips. This is the NPU that Intel is reportedly planning to implement in the Arrow Lake Refresh CPUs.

    Despite the potential power of the NPU that Intel includes in its new processors, there are not many truly practical applications for an NPU on Windows at this moment. Thus, it raises the question of whether it’s wise to allocate valuable silicon for something as underused as an NPU in desktop CPUs.

    Exciting Times Ahead

    Overall, H2 2025 seems to be quite promising, as both AMD and Nvidia are also rumored to be launching new desktop GPUs within a few months. Alongside the new Intel CPUs, gamers will have plenty to look forward to this upcoming holiday season.

    Source:
    Link

    Tags: , ,
  • Intel Shifts Focus to 14A Production After 18A Setbacks

    Intel Shifts Focus to 14A Production After 18A Setbacks

    Key Takeaways

    1. Intel reported a significant loss of $18.8 billion for fiscal year 2024, leading to CEO Pat Gelsinger’s departure.
    2. Delays in the Intel 18A manufacturing process hinder Intel’s competitiveness against TSMC, which is advancing its N2 production.
    3. Intel plans to limit the use of Intel 18A for its own chip production, with minimal output for major clients like Amazon and Microsoft.
    4. Upcoming “Panther Lake” laptop processors are expected to use the Intel 18A process starting in late 2025.
    5. Intel is pivoting towards Intel 14A manufacturing to compete with TSMC and Samsung, aiming to attract clients like Apple and Nvidia, with strategy discussions planned for July and fall.

    The current state of Intel is deteriorating significantly. The company has revealed a staggering loss of $18.8 billion for the fiscal year 2024, leading to the departure of CEO Pat Gelsinger. As a result, the delays in the Intel 18A manufacturing process are making it outdated and unable to compete effectively. A report from Reuters indicates that while Intel is struggling, TSMC’s N2 production is progressing as planned.

    Shift in Production Strategy

    Intel’s recent plans involve limiting the use of Intel 18A and Intel 18A-P exclusively for its own chip production, rather than attracting new clients. There will be a “small quantity” of chips produced for major clients like Amazon and Microsoft using the Intel 18A process. The upcoming “Panther Lake” laptop processors are anticipated to utilize the 18A process starting in the latter half of 2025. As reported by Reuters, the Intel 18A process is probably similar to TSMC’s N3, although the N3 technology has been in mass production since 2022.

    Future Directions

    In an effort to enhance its competitive stance against TSMC and Samsung Foundry, Intel is reportedly shifting its focus to Intel 14A, the next-generation manufacturing process that supposedly provides benefits over TSMC’s offerings. The intention is to attract major customers such as Apple and Nvidia, encouraging them to transition from TSMC. Nevertheless, Intel has yet to finalize its strategy regarding both Intel 18A and Intel 14A. Discussions about potential strategies will take place during a supervisory board meeting in July, with final decisions anticipated by a fall meeting at the latest.

    Source:
    Link

    Intel Shifts Focus to 14A Production After 18A Setbacks

     

    Tags: ,
  • Intel to Lay Off Up to 20% of Chip Production Workforce

    Intel to Lay Off Up to 20% of Chip Production Workforce

    Key Takeaways

    1. Intel plans to cut 15 to 20% of jobs in chip manufacturing, with potential reductions in other divisions.
    2. Job losses could reach between 16,000 to 22,000, reducing the workforce from 124,800 to around 108,900.
    3. The company’s goal is to become “leaner, faster, and more efficient” by simplifying its organization.
    4. New CEO Lip-Bu Tan aims to streamline operations and focus more on engineers and customer demands.
    5. Intel’s stock price has dropped nearly 70% since April 2021, currently sitting just over $21 per share.

    Intel has recently informed its employees via email about plans to cut 15 to 20% of jobs in chip manufacturing. Additional reductions in other divisions are also possible, as the company is looking to downsize in various areas of its business.

    Job Reductions Ahead

    Back in December 2023, Intel employed around 124,800 people. Fast forward to December 2024, and that number has dropped to 108,900. With the new cuts coming, estimates suggest that the total job losses might reach between 16,000 to 22,000. A spokesperson from the company stated that Intel’s goal is to become a “leaner, faster and more efficient company.” They also plan on simplifying the organization and putting more focus on engineers and customer demands. These choices were made after thorough consideration to cut costs and improve Intel’s future standing.

    Leadership Changes and Market Pressures

    Lip-Bu Tan, the new CEO of Intel, had previously shared plans for streamlining the company about two months ago. Still, the exact number of job cuts wasn’t clear until now, showing that the situation is getting serious. The company has been feeling significant pressure lately; since April 2021, Intel’s stock price has plummeted nearly 70%. Although there was a slight recovery in 2023, the stock price has resumed its decline and is currently just over $21 per share, showing no signs of recovery in the near future.

    Source:
    Link

    Tags: ,
  • Intel Arc Battlemage Review: Big Challenges, Small Margins

    Intel Arc Battlemage Review: Big Challenges, Small Margins

    Key Takeaways

    1. Initial Availability and Driver Issues: The Arc B580 faced poor initial availability, primarily in the US, and driver issues caused CPU slowdowns on older systems.

    2. Silicon Costs Impacting Profitability: The larger size of the BMG-G21 chip increases production costs and risks, leading to significant financial losses for Intel despite aggressive pricing.

    3. Performance Limitations: Battlemage underperformed in productivity benchmarks compared to its predecessor, Alchemist A770, except in specific Unreal Engine 5 tests.

    4. Market Competition and Product Timing: The future of the Arc Pro B60 is uncertain due to its delayed market entry, risking overshadowing by potential RTX 5000 series refreshes and competitive AMD pricing.

    5. Need for Stronger Competition: Despite challenges, there’s hope for Intel Arc to emerge as a strong contender in the GPU market, as competition is crucial for industry health.

    On the surface, the Arc B580 (and to a lesser degree, the B570) seemed like a dream come true for gamers. The drivers promised a near plug-and-play setup, the performance was superior to the outgoing RTX 4060 and Radeon RX 7600, and the increased vRAM would ensure a longer lifespan compared to those GPUs. However, reality turned out to be more complex. Initial availability was poor, primarily favoring the US market, and driver issues caused CPU slowdowns on older, budget-friendly systems. Plus, there was the significant issue of silicon costs. The BMG-G21 chip inside Battlemage measures approximately 272 mm², which is much larger than Nvidia’s AD106 or AMD’s Navi33—its main competitors from the previous generation. Once again, Intel managed to stay ahead mainly due to its aggressive pricing strategy.

    Silicon and Pricing Connection

    While my earlier thoughts on Arc Alchemist may have stirred some debate, there is a clear logical relationship between the silicon used in a GPU and its production costs. Bigger chips occupy more space on a wafer and have a higher risk of defects, which can lower yields or necessitate overdesign for reliability, both of which can drive up costs or reduce profit margins. Similar to Alchemist, Intel opted to sell a chip equivalent to an RTX 5070 for merely half the price, all while delivering performance akin to an RTX 5060. This approach led to significant financial losses, especially as the overall company was already facing economic challenges, including concerning trends in the laptop segment.

    Battlemage’s Performance

    Battlemage didn’t excel in productivity either, lagging behind the Alchemist A770 in benchmarks like Puget and Blender, with the exception of certain Unreal Engine 5 tests from Puget. Yet, there appeared to be a glimmer of hope during Computex last month, thanks to the Arc Pro B60, which boasts an impressive 24 GB of vRAM. Although many are weary of seeing LLMs integrated into every product, the demand for GPUs with over 16 GB for larger AI models remains high. The only current options are the prohibitively priced RTX 5090, the aging RTX 3090, and the 4090. If there’s a chance for Intel to market BMG-G21 chips with decent profit margins, helping Arc contribute positively to Intel’s finances and even fund the next-gen Celestial, this could be it.

    Intel, however, wasn’t actually selling Arc Pro B60s; instead, they were showcasing them and promoting their inclusion in preconfigured multi-GPU “Battlematrix” systems, expected to launch later this year. The standalone card availability remains uncertain, contingent upon Intel’s timeline for enhancing Battlemage’s software. If the Pro B60 takes too long to hit the market, it risks being overshadowed by a rumored refresh of the RTX 5000 series, which may utilize 3GB memory chips to offer additional vRAM. Once having 24 GB of vRAM ceases to be a unique feature in the mid-market segment, the Pro B60 could be eclipsed by a possible RTX 5070 Ti Super, while its attractiveness for budget-conscious gamers might diminish if AMD decides to slightly reduce the price of the RX 9060 XT (currently $349.99 on Amazon), which is already quite competitive at its suggested retail price.

    Future Prospects for Intel Arc

    Is it still possible for Intel Arc to turn the tide? Could a rumored B770 rival the midrange offerings from GeForce and Radeon? Personally, I really hope so—having a third strong contender in the GPU market is crucial for healthy competition. However, the current state of Arc makes me cringe, just like when I penned my previous article. Similar to the GPU shortages of previous years, the local AI surge has provided Intel with an opportunity to thrive. Yet, the immature software hindered Alchemist long enough for the initial chance to be lost, forcing Intel to sell high-cost silicon in lower-priced graphics cards, ultimately starving their next generation of research and development funding.

    Source:
    Link

    Tags: , ,
  • Intel Nova Lake Core Count and TDP Details Leaked

    Intel Nova Lake Core Count and TDP Details Leaked

    Key Takeaways

    1. Intel’s Core Ultra 400 Nova Lake processors will focus on increasing core count to enhance multicore performance, with hyperthreading largely removed.
    2. The Core Ultra 9 processor is expected to feature a high core count and improved performance with new Coyote Cove P-cores and Arctic Wolf E-cores.
    3. Anticipated IPC (Instructions Per Cycle) increase of 20-40% compared to the previous Arrow Lake generation.
    4. Nova Lake processors are projected for release in 2026, with potential X3D-like models incorporating additional L3 cache.
    5. The Nova Lake CPU tile is likely to be manufactured using TSMC’s N2 node instead of Intel’s own advanced 18A/18AP process.

    Earlier predictions suggested that Intel would significantly boost the core count in its Core Ultra 400 Nova Lake processors. With hyperthreading largely removed, increasing the number of cores is the main way to enhance multicore performance in the new chips. A reliable leaker, @g01d3nm4ng0, has now disclosed the core counts for each CPU model.

    Tentative CPU Names

    The CPU names mentioned above are provisional and may be altered by the time they launch. However, the Core Ultra 9 processor is expected to stand out due to its massive core count. Moreover, its new Coyote Cove P-cores and Arctic Wolf E-cores are anticipated to deliver an IPC increase of around 20-40% compared to Arrow Lake. Additionally, the integrated GPU will combine elements from Celestial and Druid designs.

    A Long Wait Ahead

    Regardless, it’ll take some time before we get definitive details, as Nova Lake is not expected to be available until 2026. Still, the anticipation might be justified since Intel could introduce X3D-like models by pairing a CPU tile with additional L3 cache through Foveros 3D packaging technology. Lastly, it is widely believed that the Nova Lake CPU tile will be produced using TSMC’s N2 node rather than Intel’s advanced 18A/18AP process.

    Source:
    Link

    Tags: , ,