According to industry sources, MediaTek is reportedly planning to integrate an Nvidia GPU into its next-generation flagship mobile processor, potentially as early as 2024. This collaboration between MediaTek and Nvidia goes beyond enhancing AI and gaming capabilities on MediaTek’s application processors for mobile handsets. The two companies are also expected to cooperate in developing Windows on Arm (WOA) platform products for notebooks.
Despite Nvidia’s 21% decline in revenue for the fourth quarter of the previous year, the US-based GPU supplier has been securing large-volume orders for its A100/H100 chips from general customers, as well as the A800/H800 GPUs specifically tailored for the Chinese market.
These orders, along with Nvidia’s partnership with Taiwan Semiconductor Manufacturing Company (TSMC), which has secured follow-up foundry orders for the mentioned chips, indicate strong production capacity utilization in TSMC’s 7/5nm process nodes. Additionally, Nvidia has reserved CoWoS packaging capacity at TSMC.
With increasing momentum in GPU orders driven by seasonal demand in the gaming market, Nvidia is expected to witness performance improvements quarterly in the latter half of the year.
credit: digitimes
The collaboration between MediaTek and Nvidia extends to the development of WOA platform products, leveraging Nvidia GPUs and AI technologies.
This partnership is anticipated to enable both companies to penetrate the notebook market. MediaTek, which currently focuses on the entry-level Chromebook market with a 20% market share, aims to strengthen its presence in the midrange to high-end notebook segments through the WOA platform products.
While global Chromebook shipments are projected to reach around 20 million units in 2023, maintaining a similar level to the previous year, the integration of Nvidia GPUs and AI technologies in MediaTek’s products could contribute to the expansion of their market reach and competitiveness in the notebook industry.
The prices of AMD Radeon RX 7900 XT and NVIDIA GeForce RTX 4080 graphics cards are currently experiencing significant declines in China, much to the delight of consumers. The overall trend of falling prices below the manufacturer’s suggested retail prices (MSRP) has been observed across various hardware components, presenting a fantastic opportunity for buyers.
Notably, both official price reductions and retailer discounts have been seen for NVIDIA, Intel, and AMD products over the past few months. In China, these price drops have been even more remarkable, with specific variants of graphics cards being sold at prices up to 20% below MSRP.
Take, for instance, the AMD Radeon RX 7900 XT 20 GB graphics card, which recently hit its lowest price ever in the US at $762 US.
credit: wccftech
This represents a 15% decrease from its original MSRP of $899 US and a 10% drop from the new MSRP of $849 US. In China, where the graphics card initially had a high starting MSRP of 7399 RMB ($1063 US), prices have seen a significant decline. The XFX MERC 618 variant, for example, can now be found starting at 5899 RMB or $847 US, making it slightly cheaper than the US MSRP and reflecting a 20% reduction compared to the Chinese MSRP.
However, the most enticing deal can be found on Taobao, where the AMD Radeon RX 7900 XT is available for 5499 RMB, equivalent to $790 US. This astonishing price represents a remarkable 25% drop below the MSRP. Chinese gamers are indeed in for an incredible offer, and it is anticipated that prices will continue to decrease for the AMD Radeon RX 7900 XT until they match the deals seen in the US in the coming days.
credit: baidu
Turning to the NVIDIA GeForce RTX 4080 graphics card, the Manli Gallardo variant is currently available in China for as low as 7399 RMB, representing a substantial 22.1% discount from the MSRP of 9499 RMB. This brings the price down to $1060 US, nearly $50 US lower than the lowest observed price for US buyers at $1109 US. Consequently, Chinese prices exhibit an impressive 11.5% reduction compared to the US MSRP of $1199 US.
Moreover, Intel’s Arc graphics cards have also seen some price reductions. The Arc A750 8 GB variant, for example, starts at 1599 RMB or $229 US. This represents a $20 US decrease from the new US MSRP of $249 US. The Arc A750 offers an excellent option for value-oriented gamers and can be found in the US market for a similar price of around $220 US.
All in all, the current market trend in China presents a unique opportunity for consumers to acquire AMD Radeon RX 7900 XT, NVIDIA GeForce RTX 4080, and Intel’s Arc graphics cards at highly attractive prices, signaling a promising time for tech enthusiasts and gamers alike.
NVIDIA has placed additional orders for wafer supply at TSMC in response to the surging demand for its top AI GPUs, such as the A100 and H100 models. This increased demand has the potential to disrupt the supply of gaming chips, as NVIDIA is focusing more resources on AI, which it considers a revolutionary technology for the PC and tech industry. Nevertheless, NVIDIA is working tirelessly to ensure a sufficient chip supply for its major partners who are willing to pay a premium for these world-class AI chips.
According to reports from DigiTimes, NVIDIA is specifically requesting chips that utilize the CoWoS packaging technology.
CoWoS, also known as Chip-on-Wafer-on-Substrate, is a packaging technology deployed in NVIDIA’s high-end data center and cloud GPUs that utilize HBM memory. The current Ampere line of A100 GPUs and the upcoming Hopper line of H100 GPUs, which are primarily used for AI and machine learning applications, employ this technology.
TSMC, with its monthly capacity to produce around 8000-9000 CoWoS wafers, will need to accommodate an additional 10,000 wafers throughout 2023 due to NVIDIA’s increased order. This heightened demand has also sparked optimism at TSMC about the growth potential of its CoWoS technology.
Another company using the CoWoS technology is AMD, which will be supplying its Instinct-class chips to Microsoft. While there were rumors of a new chip under the codename “Athena” to counter AMD’s offerings, it has been clarified that Microsoft will instead utilize existing and upcoming AMD accelerators to power its AI initiatives.
Bloomberg updated the article
Frank Shaw, a Microsoft spokesman, denied that AMD is part of Athena. “AMD is a great partner,” he said. “However, they are not involved in Athena.”
China is also experiencing significant demand for the green teams latest AI GPUs, despite the fact that the variants offered to them may lack certain advanced interconnect fabrics. The prices offered to Chinese customers are notably higher than those in regions not affected by US technology sanctions. It remains to be seen whether the green team will reduce the production of A800 and H800 GPUs in favor of the standard A100 and H100 GPUs.
Overall, the growing demand for AI GPUs has led NVIDIA to secure additional wafer supply and tighten the partnership with TSMC, while also fueling optimism for the future of CoWoS technology.
NVIDIA is reportedly planning to discontinue the production of the GeForce RTX 3060 Ti graphics card in order to make room for its upcoming GeForce RTX 4060 Ti models. The highly anticipated NVIDIA GeForce RTX 4060 Ti is expected to hit store shelves later this month on May 24th.
Serving as the successor to the popular GeForce RTX 3060 Ti 8 GB, which was launched towards the end of 2020 and currently holds the 7th position in popularity on the Steam Hardware Survey, the new graphics card is generating excitement among users.
The GeForce RTX 3060 Ti was originally priced at $399 US, and early reports indicate that the RTX 4060 Ti may follow a similar pricing pattern, with two different variants available.
The initial variant of the NVIDIA GeForce RTX 4060 Ti, set to be released later this month, will come equipped with 8 GB of memory. However, there are also plans for a second variant to be launched in July, featuring an impressive 16 GB of memory.
As a result, there will likely be a variation in price, with the standard model expected to retain a similar price point to its predecessor, around $399 US, while the higher memory variant may be priced at $499 US.
credit: wccftech
According to information from Chinese Board Forums, NVIDIA has instructed its partners to clear their existing inventory of GeForce RTX 3060 Ti graphics cards. This may explain why Colorful, a manufacturer, has recently introduced a new GDDR6X variant of the card.
Despite being over three years old, the GeForce RTX 3060 Ti cards are still being sold close to their original manufacturer’s suggested retail price (MSRP) of around $390-$400 US. However, it is anticipated that selling them at the same prices will become increasingly challenging once the RTX 4060 Ti hits the market at a similar price range.
Both the 16 GB and 8 GB variants of the NVIDIA GeForce RTX 4060 Ti graphics cards will feature a similar core configuration, including the AD106-350 GPU, 34 SMs, 4352 cores, and 16 Gbps memory running on a 128-bit bus interface, delivering a bandwidth of 288 GB/s. These specifications promise impressive performance and are likely to excite gaming enthusiasts and professionals alike.
The release dates of the next 8 GB graphics cards, the AMD Radeon RX 7600 and NVIDIA GeForce RTX 4060 Ti, are becoming more evident as we approach closer to their release. According to sources, the AMD Radeon RX 7600 8 GB graphics card will be released on May 25th, a day after NVIDIA releases its GeForce RTX 4060 Ti 8 GB graphics cards.
Both cards will be geared at mainstream gamers, but early data indicate that the 4060 Ti 8 GB will outperform the AMD Radeon RX 7600, with the Navi 33 option competing against the non-Ti 4060, which will be available in July.
According to recent rumours, NVIDIA will release three GeForce RTX 4060 series cards. The first card, the 8 GB GeForce RTX 4060 Ti, will be available this month, with two more cards due in July, including the 16 GB GeForce RTX 4060 Ti and a Non-Ti 8 GB edition. Meanwhile, according to the latest leak, AMD will introduce their entire Navi 33 GPU on the AMD Radeon RX 7600.
A Singapore retail outlet has also put the AMD Radeon RX 7600 8 GB graphics card’s shelf release date as May 26th, however due to time zone variations, that’s May 25th for US residents.
credit: wccftech
These two launches will make the end of May pretty packed, as we will not only have these two cards to choose from, but also several special models from AIB. Because the debuts are so close to the Computex 2023 event, we can expect to see some new custom variants on the exhibit floor, though we don’t expect a full new set of cards to be announced.
According to previous rumours, the NVIDIA GeForce RTX 4060 Ti will use the AD106-350-A1 GPU core, a scaled-down version of the full AD106 graphics chip, and will have 34 SMs or 4352 CUDA cores, 16/8 GB GDDR6 memory running at 18 Gbps across a 128-bit bus interface, providing the card with 288 GB/s of bandwidth. The GPU also has 32 MB of L2 cache, which is an 8x improvement over the GeForce RTX 3060 Ti.
The PG190 SKU 363 PCB is used in the 16 GB variant, while the PG190 SKU 361 PCB is used in the 8 GB form. The NVIDIA GeForce RTX 4060 Ti graphics card will be available in SFF and compact ITX configurations, making it ideal for small PC builders.
The card will also consume significantly less power, requiring close to 150-160W or less while gaming, which is 25% less than its predecessor, the RTX 3060 Ti. The cards are set to be released at the end of May and will cost between $399 and $499 USD.
The AMD Navi 33 GPU will power the AMD Radeon RX 7600 series graphics cards and will be the third and only monolithic processor in the RDNA 3 portfolio. The Navi 33 GCD contains two Shader Engines, each with two Shader Arrays (2 per SE / four in total). This equates to 16 WGPs or 32 Compute Units for a total of 2048 cores, the same as the Navi 23 GPU.
Despite the fact that AMD’s Radeon RX 7900 series has already been released, AMD’s Radeon RX 6800 remains one of the top graphics cards on the retail market. Sasa Marinkovic, AMD’s senior director of game marketing, wants to make sure buyers don’t forget this by posting a chart demonstrating the Radeon RX 6800’s superiority over the GeForce RTX 3070.
The Radeon RX 6800 and GeForce RTX 3070 are both last-generation models from late 2020. Back then, comparing the two cards made little sense because their MSRPs placed them in distinct tiers.
The Radeon RX 6800 was released with a $579 MSRP, whereas the GeForce RTX 3070 was released with a $499 MSRP. Of course, neither graphics card was available at their respective MSRPs between 2020 and 2022.
Since the conclusion of Ethereum mining, things have finally calmed down. The lowest Radeon RX 6800 is now $479, while the GeForce RTX 3070 is now $456, making them direct competitors.
While Nvidia has announced the GeForce RTX 4070 to replace the GeForce RTX 3070, AMD has yet to release a replacement to the Radeon RX 6800.
(Image credit: Sasa Marinkovic/Twitter)
The GeForce RTX 4070 is likewise priced at $599, which is 20% higher than the MSRP of its predecessor while providing GeForce RTX 3080-level performance with lower power needs. That means testing the GeForce RTX 4070 against the Radeon RX 6800 is somewhat pointless, especially since the RTX 4060 Ti is expected to come before the end of the month.
In Marinkovic’s assessment of 32 games, the Radeon RX 6800 performed 13.4% faster than the GeForce RTX 3070 on average. The performance delta ranges from -2% to +31%. Only in Metro Exodus, Grand Theft Auto V, and Dota 2 did the GeForce RTX 3070 outperform the Radeon RX 6800, and the difference was essentially a tie.
The tweeted image also emphasises how the Radeon RX 6800 has twice as much onboard memory as the GeForce RTX 3070, a selling point that AMD has made a point of emphasising.
AMD put the two graphics cards through their paces at a native 1440p (2560×1440) resolution. However, the chipmaker did not specify which graphical settings it utilised for the tests, which is critical information. Nonetheless, AMD’s assertions typically match our own findings, so there doesn’t appear to be any shady behaviour going on.
The capability of Nvidia’s mystery A800 compute GPU, which is developed for the Chinese market, has been exposed in a relatively brief narrative about overwhelming demand for Nvidia’s high-performance computing hardware in China. According to MyDrivers, the A800 operates at 70% the speed of A100 GPUs while adhering to tight US export requirements that limit the amount of processing power Nvidia can sell.
Nvidia’s A100, now three years old, is a powerhouse: it produces 9.7 FP64/19.5 FP64 Tensor TFLOPS for HPC and up to 624 BF16/FP16 TFLOPS (with sparsity) for AI tasks. Even if the values are reduced by roughly 30%, they are still impressive: 6.8 FP64/13.7 FP64 Tensor TFLOPS and 437 BF16/FP16 (with sparsity).
In terms of compute capabilities, despite ‘castration’ (performance limitations), as MyDrivers puts it, Nvidia’s A800 competes with fully-fledged China-based Biren’s BR104 and BR100 compute GPUs.
Meanwhile, Nvidia’s compute GPUs and CUDA architecture are extensively supported by its customers’ applications, whereas Biren’s CPUs have yet to be embraced. Because of the new regulations, even Biren cannot ship full-fledged computing GPUs to China.
The export limits implemented by the US in October 2021 prohibit the transfer to China of American technologies that enable supercomputers with performance exceeding 100 FP64 PetaFLOPS or 200 FP32 PetaFLOPS in a space of 41,600 cubic feet (1,178 cubic metres) or less. While the export restrictions do not directly limit the performance of each compute GPU sold to a Chinese business, they do limit throughput and scalability.
credit: biren technology
Following the implementation of the new laws, Nvidia lost the ability to sell its ultra-high-end A100 and H100 compute GPUs to Chinese clients without an export licence, which is difficult to obtain. In order to meet the performance demands of Chinese hyperscalers, the firm produced the A800, a scaled-down version of their A100 GPU. It was unclear how capable this GPU was until now.
As the use of artificial intelligence grows among both consumers and organisations, so does the demand for high-performance technology capable of handling acceptable workloads. Nvidia is one of the primary beneficiaries of the AI megatrend, which is why its GPUs are so popular that even the entry-level A800 is sold out in China.
Biren’s BR100 will be offered in an OAM form-factor and will be capable of using up to 550W of power. The chip supports the company’s unique 8-way BLink technology, which allows up to eight BR100 GPUs to be installed per machine.
The 300W BR104, on the other hand, will come in an FHFL dual-wide PCIe card form-factor and will allow up to 3-way multi-GPU setup. According to EETrend (via VideoCardz), both chips employ a PCIe 5.0 x16 interface with the CXL protocol for accelerators on top.
Biren claims that both of its chips are manufactured utilising TSMC’s 7nm-class fabrication process (without specifying whether N7, N7+, or N7P is used). The bigger BR100 has 77 billion transistors, compared to 54.2 billion in the Nvidia A100, which is also built on one of TSMC’s N7 nodes.
The company also claims that in order to overcome TSMC’s reticle size limitations, it had to use chiplet design and the foundry’s CoWoS 2.5D technology, which is entirely logical given that Nvidia’s A100 was approaching the size of a reticle and the BR100 is expected to be even larger due to its higher transistor count.
According to people with knowledge of the situation, Microsoft Corp. is collaborating with Advanced Micro Devices Inc. on the chipmaker’s entry into artificial intelligence processors as part of a multifaceted strategy to obtain more of the highly sought-after parts.
According to the people, who declined to be identified because the conversation is private, the companies are working together to provide a rival to Nvidia Corp., which currently dominates the market for graphics processing units with AI capabilities. According to the sources, the software giant is supporting AMD’s initiatives by offering engineering resources and collaborating with the chipmaker on the Athena processor, a custom-built Microsoft chip for AI workloads.
Microsoft gained about 1% on Thursday, while AMD stock increased by more than 6.5%. Reps for AMD declined to comment. The price of Nvidia fell by 1.9%.
With the explosion of chatbots like ChatGPT and other services based on the technology, there is a larger rush to increase AI processing power, which is in high demand. Microsoft is a leader in cloud computing and an innovator in the application of AI. The company promised to include such features in its entire software lineup and has invested $10 billion in OpenAI, the maker of ChatGPT.
credit: bloomberg
Additionally, the action reflects Microsoft’s growing involvement in the chip sector. Under the direction of former Intel Corp. executive Rani Borkar, the company has been developing a silicon division over the past few years, and the division now employs close to 1,000 people. The Athena artificial intelligence chip, which Microsoft is developing, was covered by The Information last month.
According to one of the people, Microsoft has invested about $2 billion in its chip initiatives, and several hundred of those employees are working on the Athena project. However, the project does not herald a rift with Nvidia. Microsoft plans to continue collaborating closely with the maker of the chips used to train and power AI systems.
Microsoft’s partnership with OpenAI and its own lineup of recently launched AI services are requiring more computing power than the company anticipated when it placed its order for chips and built its data centres. Businesses interested in integrating OpenAI’s ChatGPT service into their own goods or internal programmes have expressed interest, and Microsoft has unveiled a chat-based Bing and new AI-enhanced Office tools.
Older products like GitHub’s code-generation tool are also being updated by it. The expensive and potent processors Nvidia offers are necessary for all of those AI programmes, which are run in Microsoft’s Azure cloud.
It will be difficult to develop a lineup that can compete with Nvidia’s. Customers can quickly upgrade their capabilities by using the company’s integrated software and hardware, which includes servers, networking hardware, chips, and a programming language.
One of the causes of Nvidia’s rise to prominence is this. However, Microsoft is not the only company attempting to create its own AI processors. Amazon, a competitor in the cloud space, bought Annapurna Labs in 2016 and has since created two different AI processors. Google, a subsidiary of Alphabet Inc., also has its own training chip.
A brand-new NVIDIA GeForce RTX graphics card will be unveiled by ASUS at its ROG Pulse livestream on May 10. An image of a soldier sporting a “TUF Gaming” backpack and a Camu uniform was posted on the official ASUS Facebook page. According to the description, a brand-new GeForce RTX graphics card with the codename “TUF OG” will soon be available for purchase and be used in gaming PCs.
Here is ASUS’s complete statement, which it is asking its followers to guess the name of the chipset and the full card name:
Hey recruit!
Stand by for a new Geforce RTX™ graphics card! The TUF OG is poised for deployment in your gaming rig…
Guess the chipset and full card name for a chance to win it!
Watch the ROG Pulse livestream for the official unveil and winner announcement: May 10th, 9 am | New York May 10th, 3 pm | Berlin May 10th, 9 pm | Taipei
credit: wccftech
You’ll notice that ASUS doesn’t mention the NVIDIA 40 series specifically, but given that this new GPU will be released in 2023, it will undoubtedly be a member of the Ada Lovelace family. The card will also relate in some way to the TUF Gaming name.
This might be a new GeForce RTX 40 series card from ASUS that bears the “TUF OG” branding and has a camouflage pattern similar to the vest and uniform worn by the soldier in the image. ASUS already offers a number of GeForce RTX 40 series cards in TUF Gaming designs.
Additionally, NVIDIA’s GeForce RTX 4060 Ti graphics card is anticipated to be unveiled later this month, so this announcement may be related. An AIB hasn’t teased something like this in a while. The new family may only be introduced by ASUS at the event, and the products may not go on sale until a later time or until the RTX 4060 Ti is introduced around Computex.
The latest update of OBS Studio, version 29.1, brings exciting news for content creators and streamers. This release introduces support for AV1 streaming to YouTube over Enhanced RTMP, leveraging the capabilities of the next-generation video codec.
OBS 29.1 — With AV1 Encoding on GeForce RTX 40 Series GPUs — Now Available
The AV1 codec has been gaining momentum in the industry and with the AV1 encoding feature now available on all GeForce RTX 40 Series GPUs, including laptop GPUs and the recently launched GeForce RTX 4070, creators can take advantage of real-time AV1 hardware encoding.
This technology provides a significant boost, delivering approximately 40% more efficient encoding compared to the widely used H.264 codec. Not only does AV1 encoding offer better efficiency, but it also ensures higher-quality video output, surpassing competing GPUs in this aspect.
One of the major advantages of AV1 encoding is its ability to reduce the required upload bandwidth for streaming. This is particularly beneficial as streaming services and internet service providers often impose limitations on bandwidth. With AV1, creators can achieve greater efficiency, especially at higher resolutions.
For instance, streaming 4K content at 60 frames per second now only requires a 10 Mbps upload bandwidth, compared to the previous 20 Mbps required by H.264. This breakthrough makes 4K60 streaming accessible to a broader audience, enabling a more immersive and visually stunning experience.
AV1 — The New Standard
As a founding member of the Alliance for Open Media, NVIDIA has played a crucial role in the development of the AV1 codec. The company recognized the need to push the boundaries of outdated formats that were defined nearly two decades ago, driven by the demands of gamers and online content creators.
The previous standard for live streaming, H.264, often fell short when it came to delivering high-quality video, especially at higher resolutions and frame rates. The introduction of AV1 addresses these limitations by providing creators with a more efficient and advanced encoding option, resulting in improved image quality and smoother playback.
YouTube users can now benefit from AV1 support through the recent update to RTMP. This enhanced protocol also includes HEVC streaming support, expanding the range of formats available to users within the existing low-latency protocol commonly used for H.264 streaming. Although still in beta, the enhanced RTMP ingestion feature on YouTube unlocks new possibilities for content creators, paving the way for enhanced streaming experiences.
To configure OBS Studio for AV1 streaming using GeForce RTX 40 Series GPUs, detailed instructions can be found in the OBS setup guide, ensuring a seamless integration of this cutting-edge technology into your streaming workflow.
GeForce RTX 40 Series GPUs also introduce AV1 encoding support on the eighth-generation NVENC, taking high-quality streaming to new heights. This feature offloads compute-intensive encoding tasks from the CPU to the dedicated hardware on the GPU, resulting in improved performance and efficiency.
NVENC on GeForce RTX GPUs has been carefully designed to meet the demands of professional content creators, preserving video quality with exceptional accuracy. Creators can now achieve higher-quality streams at the same bitrate as competing products or reduce the bitrate while maintaining a similar level of picture quality.
Better Streams With NVENC, NVIDIA Broadcast
For an even more polished streaming experience, NVIDIA Broadcast, an integral part of the exclusive NVIDIA Studio suite of software, offers a range of AI-powered effects to enhance live streams, voice chats, and video calls. Features like eye contact, noise and room echo removal, virtual background, and more ensure that your content looks and sounds impeccable, transforming any space into a professional home studio.
With OBS Studio 29.1 and the power of AV1 encoding on GeForce RTX 40 Series GPUs, content creators can elevate their streaming quality, unlock new possibilities, and captivate audiences with visually stunning and high-performance streams
