Tag: NVIDIA

  • Nvidia Changes Review Embargoes Before RTX 50 Series Launch

    Nvidia Changes Review Embargoes Before RTX 50 Series Launch

    Nvidia’s presentation of the GeForce RTX 50 Series at CES 2025 was full of bold promises but also left some doubts. CEO Jensen Huang claimed the RTX 5090 would be twice as fast as the RTX 4090. Furthermore, the RTX 5080, RTX 5070, and RTX 5070 Ti are set to offer significant improvements over older RTX models. However, gamers might need to be patient for real-world reviews, as Team Green has adjusted some of the review embargo timelines.

    Review Embargo Updates

    According to VideoCardz.com, the review embargo for the RTX 5080 will not be lifted until January 30th. Nvidia had originally planned for this to happen on January 21st, but delays in releasing a BIOS affected the manufacturers. Although leaks are still coming out, full reviews for the RTX 5080 can’t be published until its official release on January 30th, which means potential buyers will have little time to consider third-party evaluations of the desktop GPU before making a decision.

    The RTX 5090 Launch

    The GeForce RTX 5090 stands as the flagship of Nvidia’s Blackwell series. Nvidia has set the review embargo date for this high-end graphics card to January 24th. The RTX 5090 will be on store shelves by January 30th, giving customers the chance to think carefully about whether its hefty $1999 price is worth it. On the other hand, the RTX 5070 and 5070 Ti will not make their appearance until later in February.

    Pricing and Performance Considerations

    With the steep price of the RTX 5090, many gamers are turning their attention to the more affordable $999 RTX 5080. Instead of offering substantial raw performance over the RTX 4080, this card will significantly depend on DLSS 4. This new multi-frame generation feature is exclusive to the RTX 50 series and allows for even greater frame rates through AI technology. At launch, 75 games will support this feature. However, the value of upgrading might be questionable if the games you enjoy don’t fully support DLSS 4.

    The adjustment in the embargo dates has led to speculation that the RTX 5080 may not significantly outperform older generation GPUs. Regardless, reviews are expected soon to shed light on why Nvidia is emphasizing the RTX 5090.

    Source:
    Link


    Image 1
  • Biden Administration Imposes AI Export Controls on China

    Biden Administration Imposes AI Export Controls on China

    The Biden administration revealed a major change to U.S. export controls on sophisticated AI chips and technologies today. This new framework establishes a tiered approach that allows preferred access for crucial allies while mandating special licenses for other countries. Dubbed the “AI Diffusion Rule,” the regulations are designed to restrict access to the most powerful AI chips and models, especially for nations viewed as adversaries, with China at the forefront.

    Importance of U.S. Leadership in AI

    Gina Raimondo, the U.S. Commerce Secretary, emphasized the need to maintain America’s lead in AI development and chip design, noting, “It’s critical that we keep it that way.” This rule, managed by the Commerce Department’s Bureau of Industry and Security, is a landmark move, as it is the first instance where the U.S. limits the transfer of powerful AI models alongside advanced chips.

    Access for Trusted Allies

    The updated policy allows 18 trusted allies, which includes the UK, Canada, Japan, Germany, and South Korea, to have near-limitless access to cutting-edge U.S. AI technology. In contrast, other nations will face a range of restrictions and licensing hurdles. Countries that are not under arms embargoes can still acquire AI chips with a total computing power equivalent to about 1,700 of the newest GPUs without needing a license. This provision provides some leeway for nations developing their own AI sectors, as chip orders won’t count against specific national limits. Additionally, there is a cap of 50,000 GPUs per country, but government-to-government agreements could potentially raise that number to 100,000 units.

    Special Licensing for Greater Needs

    Certain institutions in specific countries may apply for permission to purchase as many as 320,000 advanced graphics processing units over a two-year timeframe. However, there are constraints on how much AI computational power could be sent abroad by firms and other entities. The exception for the 1,700 GPUs is likely aimed at supporting universities and medical facilities rather than data centers.

    Countries like China, Russia, Iran, and North Korea, along with other embargoed nations, are already prohibited from acquiring advanced AI chips. The new regulations aim to close a gap by restricting access to powerful AI models and hinder China’s capability to develop advanced AI systems in partnership with organizations from other countries.

    Addressing Parallel Imports

    The updated rules also address the issue of parallel imports of advanced chips into arms-embargoed nations. For instance, despite earlier restrictions, AI chips were supplied to Russia via parallel imports through third countries that did not enforce sanctions. The new quotas for these third nations now pose a significant challenge to such imports of AI chips into Russia since companies will be limited in their ability to purchase chips in bulk for resale.

    Gina Raimondo remarked, “The semiconductors that power AI and the model weights are, as we all know, a dual-use technology.” They serve many commercial functions, but can also be utilized by adversaries for military advancements, nuclear simulations, and bioweapon development.

    Industry Concerns

    Nvidia, the leading manufacturer of AI chips worldwide, has expressed strong disapproval of the announcement. In a blog post, the company described the rule as “unprecedented and misguided,” arguing that it would diminish America’s competitiveness on the global stage. Nvidia claims that while these regulations are presented as an “anti-China” measure, they wouldn’t really boost U.S. security. Instead, they would control technologies that are already accessible in mainstream gaming PCs and consumer products, ultimately harming American innovation.

    The new export regulations will be subject to a 120-day consultation period, wherein the incoming Trump administration will need to consider input from industry leaders and international allies. After this timeframe, the Trump administration could either amend or enforce the new rules. This timeline has raised concerns about consistency and the possibility of major policy changes with the new leadership.

    Biden officials have defended the need for these rules, pointing to the fast pace of global AI advancements. National Security Advisor Jake Sullivan cautioned about the potential “transformative impacts” that AI could have on both the economy and national security in the near future.

    Source:
    Link

  • RTX 5090 vs RTX 4090: Rasterization vs Ray Tracing Performance

    RTX 5090 vs RTX 4090: Rasterization vs Ray Tracing Performance

    The RTX 4090 stands as the top consumer desktop GPU within the RTX 40 series. With AMD not offering a rival to the RTX 4090, as the RX 7900 XTX is more aligned with the RTX 4080, Nvidia has maintained its dominance in the high-end gaming segment. Looking ahead, Team Green is expected to keep this top spot with the upcoming RTX 5090, especially since AMD has opted out of competing in the premium market with its RDNA 4 architecture.

    Performance Comparison

    Now, the big question is how does the RTX 5090 stack up against the RTX 4090? Nvidia has shared some performance figures for the RTX 5090, but didn’t provide in-depth insights into its raw rasterization or ray tracing performance. Instead, the focus has been on DLSS 4 and its impressive 4x Frame Generation.

    Benchmark Insights

    Fortunately, the RTX 5090 might already be in the hands of some reviewers, leading to the emergence of synthetic benchmark scores. Two notable results indicate that the RTX 5090 outperforms the RTX 4090 in both rasterization and ray tracing tests.

    According to a thread on Chiphell, which is likely to be taken down soon, the RTX 5090 achieved scores exceeding 24,000 in the 3DMark Time Spy Extreme test and over 13,500 in the Speed Way benchmark.

    Detailed Score Analysis

    Focusing on the Time Spy Extreme score, the RTX 4090’s results from 3DMark show values ranging from about 22,800 to 23,300. If we take a score of 22,800 for the RTX 4090 and 24,000 for the RTX 5090, it suggests that the Blackwell GPU is around 5.3% faster than its earlier version. Of course, results may vary with different high scores, but for our discussion, we can assume lower figures for both GPUs.

    When considering the 3DMark Speed Way test, which evaluates ray tracing, the RTX 5090’s reported score of 13,500+ positions it 8% ahead of the best RTX 4090 recorded in the 3DMark database. Again, since the RTX 4090 has scores that fall between about 11,430 and 12,489, the performance difference in favor of the RTX 5090 could be significantly larger if we base it on the lower RTX 4090 scores. For example, using the 11,430 Speed Way score for the RTX 4090, the RTX 5090’s performance is 18% superior.

    Conclusion

    Thus, it seems reasonable to conclude that the RTX 5090 could deliver nearly a 20% improvement in ray tracing capabilities compared to the RTX 4090.

    In summary, if the rasterization performance of the RTX 5090 is indeed less than 10% faster than the RTX 4090 as noted earlier, this raises questions about the $400 price hike. Such a slight improvement in rasterization might also clarify Nvidia’s strong emphasis on DLSS 4 during the RTX 50 launch.

    Nonetheless, we should hold off on forming any conclusions until we see third-party reviews for the RTX 5090. Until then, take any performance speculations regarding the RTX 5090 with a hefty dose of skepticism.

  • Nvidia Slams US Last-Minute Restrictions on AI Chip Exports

    Nvidia Slams US Last-Minute Restrictions on AI Chip Exports

    Nvidia is strongly against the forthcoming chip export limitations from the White House, characterizing them as a hurried initiative that might exceed its intended objectives. The anticipated regulations—expected to be announced imminently—aim to establish a three-tiered system to manage AI chip exports, tailored both to specific countries and companies.

    Concerns Over Policy Impact

    Ned Finkle, who leads Nvidia’s government relations, expressed his worries regarding the extensive implications of this policy. He stated that the proposed “extreme country cap policy” would negatively affect everyday computing worldwide without genuinely enhancing national security. These new restrictions would particularly influence AI accelerators, an area where Nvidia holds a dominant position.

    Proposed Access Levels

    According to the suggested rules, American semiconductors would be allocated based on different access categories. Some of the U.S.’s closest allies would receive unrestricted import permissions, while many other countries would face new limitations on overall computing capacity. These restrictions would not only target specialized AI chips but also general-purpose GPUs utilized in various devices, including gaming PCs and data centers.

    Timing and Economic Concerns

    The timing of this announcement is significant, occurring less than two weeks prior to the presidential changeover. “This last-minute policy from the Biden administration could become a legacy that will draw criticism from both the U.S. industry and the global community,” Finkle warned, cautioning that it may damage American economic interests.

    Meanwhile, Nvidia’s CEO, Jensen Huang, has expressed his willingness to work with the new administration coming in. He showed interest in meeting with Trump and even offered his assistance. Speaking at CES in Las Vegas, Huang seemed optimistic about the possibility of reduced regulations under Trump, stating, “As an industry, we want to move fast.”

    Market Implications for Nvidia

    These new regulations could significantly impact Nvidia’s position in the market, especially considering the company’s remarkable growth: its stock price nearly tripled last year, following a 239 percent increase in 2023, largely fueled by a rise in AI investments.

    Source:
    Link

  • Nvidia DLSS 4: 8x Gaming Performance Boost for GeForce RTX GPUs

    Nvidia DLSS 4: 8x Gaming Performance Boost for GeForce RTX GPUs

    Nvidia has unveiled the fourth version of its Deep Learning Super Sampling (DLSS) technology. This new update claims to offer a frame rate that is over eight times higher along with latency that is about half of what native rendering provides. A comparison video from Nvidia demonstrates that the frame rate achieved by the new DLSS is set to double compared to DLSS 3.

    Enhanced Frame Generation

    To achieve these impressive results, Nvidia has included “Multi Frame Generation” within DLSS Frame Generation. Unlike previous versions, DLSS 4 now generates three-quarters of all displayed frames using AI, instead of just every second one. Nvidia notes that this was not feasible before due to the strain it would place on GPUs, which would reduce output frame rates. However, with new advancements in the GeForce RTX 5000, Multi Frame Generation has become achievable.

    Improved Performance

    The new AI model for Frame Generation operates 40% faster while consuming 30% less graphics memory. For instance, in the game Warhammer 40K: Darktide, Nvidia asserts that DLSS 4 can deliver a frame rate increase of 10%, even when using conventional Frame Generation, saving 400 MB of graphics memory in the process. When paired with the GeForce RTX 5000, DLSS 4 can compute nearly 94% of all pixels using AI instead of relying on native rendering.

    Significant Frame Rate Boosts

    The impact on performance is substantial. Benchmarks indicate that, depending on the title, DLSS 4 can boost frame rates by anywhere from just under 370% to as much as 720%, translating to nearly four times to over eight times the original rate.

    Even though Multi Frame Generation is exclusive to the GeForce RTX 5000, users with older graphics cards, even those as far back as the GeForce RTX 2000, will still find advantages with DLSS 4. This is because the AI models for standard Frame Generation, DLSS Ray Reconstruction, DLSS Super Resolution, and DLSS Edge Smoothing (DLAA) have all seen enhancements.

    Broad Game Support

    Nvidia reports that more than 75 games now embrace DLSS, featuring major titles like Alan Wake 2, Black Myth: Wukong, and Cyberpunk 2077.

    Source: Link


    Image 1
    Image 1
    Image 1
    Image 1
  • Nvidia GeForce RTX 5090, 5080, 5070 Ti GPUs Unveiled at CES 2025

    Nvidia GeForce RTX 5090, 5080, 5070 Ti GPUs Unveiled at CES 2025

    Nvidia has finally unveiled its latest range of desktop graphics cards. The company has introduced four different models and made some bold statements regarding their performance. At the top of the line, the GeForce RTX 5090 is equipped with 32 GB of VRAM, just like was suggested before. The other models, including the GeForce RTX 5080, RTX 5070 Ti, and RTX 5070, also launch with the expected VRAM amounts. No big surprises here.

    Performance Claims

    Nvidia asserts that the GeForce RTX 5070 can match the performance of the GeForce RTX 4090. Such assertions are not a new thing; they’ve been made with every new generation. However, this time it could be a reality thanks to advancements like DLSS 4, GDDR7 VRAM, and Nvidia’s fresh texture compression technology. What’s even better is that these cards might not break the bank. Below is the complete pricing breakdown for each model.

    Pricing Details

    The pricing of the GeForce RTX 5090 seems justified, considering its VRAM upgrade. The other models are also priced quite fairly, particularly the GeForce RTX 5080, which was rumored to be around $1,500. Nvidia has stated these GPUs will hit the market in January, though they didn’t provide any more details beyond that.

    Source: Link


    Image 1
  • Nvidia Patents Holographic AR Glasses with No Backlight System

    Nvidia Patents Holographic AR Glasses with No Backlight System

    Nvidia has recently submitted a patent for a completely new design of augmented reality glasses that eliminates the standard backlight system and instead incorporates digital holographic technology. The patent is officially named “Backlight-free Augmented Reality Digital Holographic Technology” (US20250004275A1), and it was made public by the United States Patent and Trademark Office on January 2, 2025.

    Innovative AR Visuals

    This new invention makes use of a spatial light modulator paired with ambient light interference to generate AR visuals. Unlike traditional AR glasses that darken the whole lens, Nvidia’s method only shades the areas of the lens necessary for displaying AR content. This advancement could lead to improved energy efficiency and a more pleasant experience for users.

    Enhanced Control System

    The spatial light modulator is powered by a neural network-based control system, which guarantees sharp and lifelike virtual images. The design as a whole seeks to solve prevalent problems found in current AR glasses while keeping a style and feel akin to conventional eyewear.

    Competitive Market Landscape

    The augmented reality glasses market is experiencing growth, with major tech companies increasing their product lines. Last year, Meta launched its Orion glasses, and there are rumors that Samsung will present its own AR headset at the Galaxy Unpacked event on January 22. Apple is also developing its AR glasses, although they may not be available for another three to five years.

    Naturally, as is common with patents, there’s no assurance that this technology will be brought to market anytime soon. It may require significant adjustments before it’s suitable for everyday consumers.

    Source: Link


    Image 1
  • RTX 5090 Laptop VRAM Leak: 50% Increase Over RTX 4090

    RTX 5090 Laptop VRAM Leak: 50% Increase Over RTX 4090

    We shared news in October regarding Nvidia’s GeForce RTX 5090 laptop GPU, which is the top mobile graphics card in the RTX 50 Blackwell series, boasting 24 GB of VRAM. At that time, we mentioned that achieving 24 GB of VRAM on a 256-bit bus would necessitate using 3 GB GDDR7 modules. Thanks to the leaker Golden Pig Upgrade from China, we now have confirmation that the RTX 5090 laptop will indeed come with 24 GB of VRAM.

    RTX 5090’s Unique Features

    According to Golden Pig Upgrade, the RTX 5090 laptop will be the first in the RTX 50 series to utilize these 3 GB GDDR7 modules. This setup allows the GPU to have 24 GB of VRAM on a 256-bit bus. Interestingly, it seems that the RTX 5090 laptop will be the sole RTX 50 mobile card to adopt these 3 GB modules, while less powerful variants such as the RTX 5080 laptop will continue with standard memory configurations. For instance, the RTX 5080 laptop is expected to have a memory buffer of 16 GB.

    Release Timeline

    In other news, Golden Pig Upgrade suggests that gaming laptops featuring the RTX 5090 and RTX 5080 will launch in February. Meanwhile, models equipped with the mid-range RTX 5070 and RTX 5060 are expected to arrive a month later, around the end of March 2025. This aligns with CES, where manufacturers typically showcase their upcoming gaming laptops.

    Performance Expectations

    When the RTX 5090 laptop GPU is released, it may utilize the GB203 Blackwell GPU, the same chip found in the desktop RTX 5080. This GPU is thought to have 10,752 CUDA cores, a 256-bit bus, and a total TGP of 175 W. For comparison, the RTX 4090 laptop has 9,728 CUDA cores, a 256-bit bus, and a max TGP of 150 W. While the difference in CUDA cores is only about 10.5%, the architectural upgrades and increased power consumption should enable the RTX 5090 laptop to deliver a significant boost in performance.

    Anticipated Performance

    So, what kind of performance can we expect from the RTX 5090 laptop GPU? If we look at the RTX 4090 laptop’s performance, we might anticipate that the RTX 5090 laptop will perform similarly to the mid-range RTX 5070 desktop GPU, as the RTX 4090 mobile is like the RTX 4070 Super. Of course, this is a rough estimate of performance, so it’s likely that the RTX 5090 laptop could be even faster.

    Source: Link


    Image 1
  • RTX 5080 Release Date Leaked: Launch Expected in January

    RTX 5080 Release Date Leaked: Launch Expected in January

    After a long wait of over two years with the RTX 40 series, which showcased remarkable ray tracing capabilities but had some unsatisfactory price trends, we are approaching the reveal of the next-gen RTX 50 “Blackwell” GPUs. Nvidia is expected to introduce the first models, which include the RTX 5090, RTX 5080, and possibly the RTX 5070/Ti, during CES 2025 in just a few days. The lower-tier variants are likely to be released sometime after that.

    Upcoming Release Dates

    Similar to the RTX 40 series, it’s anticipated that the RTX 5090 and RTX 5080 will be available in stores shortly after their official announcement. However, Nvidia might switch things up this time around by launching the RTX 5080 ahead of the RTX 5090. Last month, we reported that the RTX 5080 might hit the shelves around mid-January 2025. Thanks to HKEPC on X, we now have a specific release date for the RTX 5080, set for January 21.

    Timelines and Expectations

    Moreover, according to HKEPC’s insights and earlier leaks, the launch date for the RTX 5090 is expected to be after January 21. If we look back at the release pattern of the RTX 40 series, there might be about a month gap between the launches of the RTX 5080 and RTX 5090. The RTX 4090 debuted on October 12, 2022, and the RTX 4080 followed a month later on November 16, 2022.

    Nvidia must ensure that the RTX 5080’s pricing remains reasonable.

    Pricing Speculations

    There are rumors suggesting that the RTX 5080 will come at a higher price than the RTX 4080. If Nvidia sets the RTX 5080’s price above the RTX 4080’s launch MSRP of $1,200, it might significantly diminish the card’s attractiveness, especially considering the negative effect on its price/performance ratio.

    Even if we expect the RTX 5080 to offer a significant performance boost and achieve rasterization performance comparable to the RTX 4090, a price tag exceeding $1,200, or even worse, nearing the RTX 4090’s level, would be quite disheartening.

    However, just like with any GPU launch, the final price of the RTX 5080 won’t be confirmed until Jensen Huang makes the announcement during the event. Until then, it’s wise to approach any pricing rumors and discussions with a healthy skepticism.

    Source: Link

  • RTX 5060 vs RTX 4060: 20% Faster but VRAM Limit Concerns

    RTX 5060 vs RTX 4060: 20% Faster but VRAM Limit Concerns

    CES 2025 is nearly here, and it’s bringing news of the new RTX 50 and RDNA 4 GPUs. Currently, we are anticipating Nvidia to showcase the RTX 5090 and RTX 5080 desktop graphics cards during the event. Additionally, it’s likely that Team Green will introduce some RTX 50 variants for laptops, as CES tends to be when laptop manufacturers like MSI reveal new gaming devices equipped with the latest technology from Intel, AMD, and Nvidia.

    Performance Expectations

    Once the next-gen gaming laptops hit the market, players can expect the new RTX 50 laptop graphics cards to provide significant performance improvements. In fact, a benchmark result from 3DMark Time Spy for the RTX 5060 laptop GPU has been shared on Bilibili, and it looks quite impressive.

    The RTX 5060 laptop is said to achieve a score of 13,821 points in the 3DMark Time Spy Graphics test. According to our records, the RTX 4060 and RTX 4070 laptop GPUs have maximum scores of 11,451 and 13,391, respectively. This suggests that the RTX 5060 laptop GPU is approximately 21% and 3% faster than the highest-performing RTX 4060 and RTX 4070 laptop models we’ve evaluated.

    Uncertainty in Performance

    However, since we lack information about the TGP of the RTX 5060 laptop, we can’t make definitive claims about its ultimate performance. The RTX 5060 laptop could potentially perform even better if the tested version wasn’t allowed to fully utilize the reported 115 W TGP.

    As things stand, the RTX 5060 laptop seems to provide a noticeable improvement over the RTX 4060 laptop variant without increasing power consumption. Still, this performance boost may not meet the demands of all games, as Nvidia is again equipping the card with just 8 GB of VRAM.

    Concerns About VRAM

    Having only 8 GB of VRAM is already becoming a significant limitation in the latest AAA games. The fact that the RTX 5060 laptop includes only 8 GB of VRAM raises concerns for the long-term viability of gaming systems that rely on this card.

    Additionally, the presence of 8 GB of memory in the mobile RTX 5060 laptop suggests that the desktop RTX 5060 will likely have the same memory specification. If Nvidia launches a next-gen desktop GPU in 2025 priced over $300 with just 8 GB of VRAM, it could be quite disappointing. This is especially relevant now, as we’ve seen with the Arc B580 that Nvidia is capable of producing a sub-$300 card with 12 GB of VRAM if they choose to.

    Luckily, we won’t have to wait long to learn about Nvidia’s strategy for the RTX 50 series since CES is almost here.

    Source: Link,,,,


    Image 1