NVIDIA

new-geforce-50-series-gpus:-there’s-the-$1,999-5090,-and-there’s-everything-else

New GeForce 50-series GPUs: There’s the $1,999 5090, and there’s everything else


Nvidia leans heavily on DLSS 4 and AI-generated frames for speed comparisons.

Nvidia’s RTX 5070, one of four new desktop GPUs announced this week. Credit: Nvidia

Nvidia’s RTX 5070, one of four new desktop GPUs announced this week. Credit: Nvidia

Nvidia has good news and bad news for people building or buying gaming PCs.

The good news is that three of its four new RTX 50-series GPUs are the same price or slightly cheaper than the RTX 40-series GPUs they’re replacing. The RTX 5080 is $999, the same price as the RTX 4080 Super; the 5070 Ti and 5070 are launching for $749 and $549, each $50 less than the 4070 Ti Super and 4070 Super.

The bad news for people looking for the absolute fastest card they can get is that the company is charging $1,999 for its flagship RTX 5090 GPU, significantly more than the $1,599 MSRP of the RTX 4090. If you want Nvidia’s biggest and best, it will cost at least as much as four high-end game consoles or a pair of decently specced midrange gaming PCs.

Pricing for the first batch of Blackwell-based RTX 50-series GPUs. Credit: Nvidia

Nvidia also announced a new version of its upscaling algorithm, DLSS 4. As with DLSS 3 and the RTX 40-series, DLSS 4’s flagship feature will be exclusive to the 50-series. It’s called DLSS Multi Frame Generation, and as the name implies, it takes the Frame Generation feature from DLSS 3 and allows it to generate even more frames. It’s why Nvidia CEO Jensen Huang claimed that the $549 RTX 5070 performed like the $1,599 RTX 4090; it’s also why those claims are a bit misleading.

The rollout will begin with the RTX 5090 and 5080 on January 30. The 5070 Ti and 5070 will follow at some point in February. All cards except the 5070 Ti will come in Nvidia-designed Founders Editions as well as designs made by Nvidia’s partners; the 5070 Ti isn’t getting a Founders Edition.

The RTX 5090 and 5080

RTX 5090 RTX 4090 RTX 5080 RTX 4080 Super
CUDA Cores 21,760 16,384 10,752 10,240
Boost Clock 2,410 MHz 2,520 MHz 2,617 MHz 2,550 MHz
Memory Bus Width 512-bit 384-bit 256-bit 256-bit
Memory Bandwidth 1,792 GB/s 1,008 GB/s 960 GB/s 736 GB/s
Memory size 32GB GDDR7 24GB GDDR6X 16GB GDDR7 16GB GDDR6X
TGP 575 W 450 W 360 W 320 W

The RTX 5090, based on Nvidia’s new Blackwell architecture, is a gigantic chip with 92 billion transistors in it. And while it is double the price of an RTX 5080, you also get double the GPU cores and double the RAM and nearly double the memory bandwidth. Even more than the 4090, it’s being positioned head and shoulders above the rest of the GPUs in the family, and the 5080’s performance won’t come remotely close to it.

Although $1,999 is a lot to ask for a graphics card, if Nvidia can consistently make the RTX 5090 available at $2,000, it could still be an improvement over the pricing of the 4090, which regularly sold for well over $1,599 over the course of its lifetime, due in part to pandemic-fueled GPU shortages, cryptocurrency mining, and the generative AI boom. Companies and other entities buying them as AI accelerators may restrict the availability of the 5090, too, but Nvidia’s highest GPU tier has been well out of the price range of most consumers for a while now.

Despite the higher power budget—as predicted, it’s 125 W higher than the 4090 at 450 W, and Nvidia recommends a 1,000 W power supply or better—the physical size of the 5090 Founders Edition is considerably smaller than the 4090, which was large enough that it had trouble fitting into some computer cases. Thanks to a “high-density PCB” and redesigned cooling system, the 5090 Founders Edition is a dual-slot card that ought to fit into small-form-factor systems much more easily than the 4090. Of course, this won’t stop most third-party 5090 GPUs from being gigantic triple-fan monstrosities, but it is apparently possible to make a reasonably sized version of the card.

Moving on to the 5080, it looks like more of a mild update from last year’s RTX 4080 Super, with a few hundred more CUDA cores, more memory bandwidth (thanks to the use of GDDR7, since the two GPUs share the same 256-bit interface), and a slightly higher power budget of 360 W (compared to 320 W for the 4080 Super).

Having more cores and faster memory, in addition to whatever improvements and optimizations come with the Blackwell architecture, should help the 5080 easily beat the 4080 Super. But it’s an open question as to whether it will be able to beat the 4090, at least before you consider any DLSS-related frame rate increases. The 4090 has 52 percent more GPU cores, a wider memory bus, and 8GB more memory.

5070 Ti and 5070

RTX 5070 Ti RTX 4070 Ti Super RTX 5070 RTX 4070 Super
CUDA Cores 8,960 8,448 6,144 7,168
Boost Clock 2,452 MHz 2,610 MHz 2,512 MHz 2,475 MHz
Memory Bus Width 256-bit 256-bit 192-bit 192-bit
Memory Bandwidth 896 GB/s 672 GB/s 672 GB/s 504 GB/s
Memory size 16GB GDDR7 16GB GDDR6X 12GB GDDR7 12GB GDDR6X
TGP 300 W 285 W 250 W 220 W

At $749 and $549, the 5070 Ti and 5070 are slightly more within reach for someone who’s trying to spend less than $2,000 on a new gaming PC. Both cards hew relatively closely to the specs of the 4070 Ti Super and 4070 Super, both of which are already solid 1440p and 4K graphics cards for many titles.

Like the 5080, the 5070 Ti includes a few hundred more CUDA cores, more memory bandwidth, and slightly higher power requirements compared to the 4070 Ti Super. That the card is $50 less than the 4070 Ti Super was at launch is a nice bonus—if it can come close to or beat the RTX 4080 for $250 less, it could be an appealing high-end option.

The RTX 5070 is alone in having fewer CUDA cores than its immediate predecessor—6,144, down from 7,168. It is an upgrade from the original 4070, which had 5,888 CUDA cores, and GDDR7 and slightly faster clock speeds may still help it outrun the 4070 Super; like the other 50-series cards, it also comes with a higher power budget. But right now this card is looking like the closest thing to a lateral move in the lineup, at least before you consider the additional frame-generation capabilities of DLSS 4.

DLSS 4 and fudging the numbers

Many of Nvidia’s most ostentatious performance claims—including the one that the RTX 5070 is as fast as a 4090—factors in DLSS 4’s additional AI-generated frames. Credit: Nvidia

When launching new 40-series cards over the last two years, it was common for Nvidia to publish a couple of different performance comparisons to last-gen cards: one with DLSS turned off and one with DLSS and the 40-series-exclusive Frame Generation feature turned on. Nvidia would then lean on the DLSS-enabled numbers when making broad proclamations about a GPU’s performance, as it does in its official press release when it says the 5090 is twice as fast as the 4090, or as Huang did during his CES keynote when he claimed that an RTX 5070 offered RTX 4090 performance for $549.

DLSS Frame Generation is an AI feature that builds on what DLSS is already doing. Where DLSS uses AI to fill in gaps and make a lower-resolution image look like a higher-resolution image, DLSS Frame Generation creates entirely new frames and inserts them in between the frames that your GPU is actually rendering.

DLSS 4 now generates up to three frames for every frame the GPU is actually rendering. Used in concert with DLSS image upscaling, Nvidia says that “15 out of every 16 pixels” you see on your screen are being generated by its AI models. Credit: Nvidia

The RTX 50-series one-ups the 40-series with DLSS 4, another new revision that’s exclusive to its just-launched GPUs: DLSS Multi Frame Generation. Instead of generating one extra frame for every traditionally rendered frame, DLSS 4 generates “up to three additional frames” to slide in between the ones your graphics card is actually rendering—based on Nvidia’s slides, it looks like users ought to be able to control how many extra frames are being generated, just as they can control the quality settings for DLSS upscaling. Nvidia is leaning on the Blackwell architecture’s faster Tensor Cores, which it says are up to 2.5 times faster than the Tensor Cores in the RTX 40-series, to do the AI processing necessary to upscale rendered frames and to generate new ones.

Nvidia’s performance comparisons aren’t indefensible; with DLSS FG enabled, the cards can put out a lot of frames per second. It’s just dependent on game support (Nvidia says that 75 titles will support it at launch), and going off of our experience with the original iteration of Frame Generation, there will likely be scenarios where image quality is noticeably worse or just “off-looking” compared to actual rendered frames. DLSS FG also needed a solid base frame rate to get the best results, which may or may not be the case for Multi-FG.

Enhanced versions of older DLSS features can benefit all RTX cards, including the 20-, 30-, and 40-series. Multi-Frame Generation is restricted to the 50-series, though. Credit: Nvidia

Though the practice of restricting the biggest DLSS upgrades to all-new hardware is a bit frustrating, Nvidia did announce that it’s releasing a new transformer module for the DLSS Ray Reconstruction, Super Resolution, and Anti-Aliasing features. These are DLSS features that are available on all RTX GPUs going all the way back to the RTX 20-series, and games that are upgraded to use the newer models should benefit from improved upscaling quality even if they’re using older GPUs.

GeForce 50-series: Also for laptops!

Nvidia’s projected pricing for laptops with each of its new mobile GPUs. Credit: Nvidia

Nvidia’s laptop GPU announcements sometimes trail the desktop announcements by a few weeks or months. But the company has already announced mobile versions of the 5090, 5080, 5070 Ti, and 5070 that Nvidia says will begin shipping in laptops priced between $1,299 and $2,899 when they launch in March.

All of these GPUs share names, the Blackwell architecture, and DLSS 4 support with their desktop counterparts, but per usual they’re significantly cut down to fit on a laptop motherboard and within a laptop’s cooling capacity. The mobile version of the 5090 includes 10,496 GPU cores, less than half the number of the desktop version, and just 24GB of GDDR7 memory on a 256-bit interface instead of 32GB on a 512-bit interface. But it also can operate with a power budget between 95 and 150 W, a fraction of what the desktop 5090 needs.

RTX 5090 (mobile) RTX 5080 (mobile) RTX 5070 Ti (mobile) RTX 5070 (mobile)
CUDA Cores 10,496 7,680 5,888 4,608
Memory Bus Width 256-bit 256-bit 192-bit 128-bit
Memory size 24GB GDDR7 16GB GDDR7 12GB GDDR7 8GB GDDR7
TGP 95-150 W 80-150 W 60-115 W 50-100 W

The other three GPUs are mostly cut down in similar ways, and all of them have fewer GPU cores and lower power requirements than their desktop counterparts. The 5070 GPUs both have less RAM and narrowed memory buses, too, but the mobile RTX 5080 at least comes closer to its desktop iteration, with the same 256-bit bus width and 16GB of RAM.

Photo of Andrew Cunningham

Andrew is a Senior Technology Reporter at Ars Technica, with a focus on consumer tech including computer hardware and in-depth reviews of operating systems like Windows and macOS. Andrew lives in Philadelphia and co-hosts a weekly book podcast called Overdue.

New GeForce 50-series GPUs: There’s the $1,999 5090, and there’s everything else Read More »

rumors-say-next-gen-rtx-50-gpus-will-come-with-big-jumps-in-power-requirements

Rumors say next-gen RTX 50 GPUs will come with big jumps in power requirements

Nvidia is reportedly gearing up to launch the first few cards in its RTX 50-series at CES next week, including an RTX 5090, RTX 5080, RTX 5070 Ti, and RTX 5070. The 5090 will be of particular interest to performance-obsessed, money-is-no-object PC gaming fanatics since it’s the first new GPU in over two years that can beat the performance of 2022’s RTX 4090.

But boosted performance and slower advancements in chip manufacturing technology mean that the 5090’s maximum power draw will far outstrip the 4090’s, according to leakers. VideoCardz reports that the 5090’s thermal design power (TDP) will be set at 575 W, up from 450 W for the already power-hungry RTX 4090. The RTX 5080’s TDP is also increasing to 360 W, up from 320 W for the RTX 4080 Super.

That also puts the RTX 5090 close to the maximum power draw available over a single 12VHPWR connector, which is capable of delivering up to 600 W of power (though once you include the 75 W available via the PCI Express slot on your motherboard, the actual maximum possible power draw for a GPU with a single 12VHPWR connector is a slightly higher 675 W).

Higher peak power consumption doesn’t necessarily mean that these cards will always draw more power during actual gaming than their 40-series counterparts. And their performance could be good enough that they could still be very efficient cards in terms of performance per watt.

But if you’re considering an upgrade to an RTX 5090 and these power specs are accurate, you may need to consider an upgraded power supply along with your new graphics card. Nvidia recommends at least an 850 W power supply for the RTX 4090 to accommodate what the GPU needs while leaving enough power left over for the rest of the system. An additional 125 W bump suggests that Nvidia will recommend a 1,000 W power supply as the minimum for the 5090.

We’ll probably know more about Nvidia’s next-gen cards after its CES keynote, currently scheduled for 9: 30 pm Eastern/6: 30 pm Pacific on Monday, January 6.

Rumors say next-gen RTX 50 GPUs will come with big jumps in power requirements Read More »

nvidia-partners-leak-next-gen-rtx-50-series-gpus,-including-a-32gb-5090

Nvidia partners leak next-gen RTX 50-series GPUs, including a 32GB 5090

Rumors have suggested that Nvidia will be taking the wraps off of some next-generation RTX 50-series graphics cards at CES in January. And as we get closer to that date, Nvidia’s partners and some of the PC makers have begun to inadvertently leak details of the cards.

According to recent leaks from both Zotac and Acer, it looks like Nvidia is planning to announce four new GPUs next month, all at the high end of its lineup: The RTX 5090, RTX 5080, RTX 5070 Ti, and RTX 5070 were all briefly listed on Zotac’s website, as spotted by VideoCardz. There’s also an RTX 5090D variant for the Chinese market, which will presumably have its specs tweaked to conform with current US export restrictions on high-performance GPUs.

Though the website leak didn’t confirm many specs, it did list the RTX 5090 as including 32GB of GDDR7, an upgrade from the 4090’s 24GB of GDDR6X. An Acer spec sheet for new Predator Orion desktops also lists 32GB of GDDR7 for the 4090, as well as 16GB of GDDR7 for the RTX 5080. This is the same amount of RAM included with the RTX 4080 and 4080 Super.

The 5090 will be a big deal when it launches because no graphics card released since October 2022 has come close to beating the 4090’s performance. Nvidia’s early 2024 Super refresh for some 40-series cards didn’t include a 4090 Super, and AMD’s flagship RX 7900 XTX card is more comfortable competing with the likes of the 4080 and 4080 Super. The 5090 isn’t a card that most people are going to buy, but for the performance-obsessed, it’s the first high-end performance upgrade the GPU market has seen in more than two years.

Nvidia partners leak next-gen RTX 50-series GPUs, including a 32GB 5090 Read More »

nvidia’s-new-ai-audio-model-can-synthesize-sounds-that-have-never-existed

Nvidia’s new AI audio model can synthesize sounds that have never existed

At this point, anyone who has been following AI research is long familiar with generative models that can synthesize speech or melodic music from nothing but text prompting. Nvidia’s newly revealed “Fugatto” model looks to go a step further, using new synthetic training methods and inference-level combination techniques to “transform any mix of music, voices, and sounds,” including the synthesis of sounds that have never existed.

While Fugatto isn’t available for public testing yet, a sample-filled website showcases how Fugatto can be used to dial a number of distinct audio traits and descriptions up or down, resulting in everything from the sound of saxophones barking to people speaking underwater to ambulance sirens singing in a kind of choir. While the results on display can be a bit hit or miss, the vast array of capabilities on display here helps support Nvidia’s description of Fugatto as “a Swiss Army knife for sound.”

You’re only as good as your data

In an explanatory research paper, over a dozen Nvidia researchers explain the difficulty in crafting a training dataset that can “reveal meaningful relationships between audio and language.” While standard language models can often infer how to handle various instructions from the text-based data itself, it can be hard to generalize descriptions and traits from audio without more explicit guidance.

To that end, the researchers start by using an LLM to generate a Python script that can create a large number of template-based and free-form instructions describing different audio “personas” (e.g., “standard, young-crowd, thirty-somethings, professional”). They then generate a set of both absolute (e.g., “synthesize a happy voice”) and relative (e.g., “increase the happiness of this voice”) instructions that can be applied to those personas.

The wide array of open source audio datasets used as the basis for Fugatto generally don’t have these kinds of trait measurements embedded in them by default. But the researchers make use of existing audio understanding models to create “synthetic captions” for their training clips based on their prompts, creating natural language descriptions that can automatically quantify traits such as gender, emotion, and speech quality. Audio processing tools are also used to describe and quantify training clips on a more acoustic level (e.g. “fundamental frequency variance” or “reverb”).

Nvidia’s new AI audio model can synthesize sounds that have never existed Read More »

facebook,-nvidia-push-scotus-to-limit-“nuisance”-investor-suits-after-scandals

Facebook, Nvidia push SCOTUS to limit “nuisance” investor suits after scandals


Facebook, Nvidia ask SCOTUS to narrow legal paths to retrieve investor losses.

The Supreme Court will soon weigh two cases that could potentially make it harder for misled investors to sue Big Tech companies after major scandals.

One case involves one of the largest tech scandals of all time, the Facebook-Cambridge Analytica data breach. In 2019, Facebook agreed to pay “more than $5 billion in civil penalties to settle charges by the Federal Trade Commission (FTC) and the Securities and Exchange Commission (SEC) that it had misled its users and investors over the privacy and security of user data on its platform,” a Supreme Court filing said.

The other case involves an allegation that Nvidia intentionally hid how much of its 2017–2018 GPU demand was due to a volatile cryptocurrency boom and not Nvidia’s core gaming business—allegedly misleading investors ahead of a crypto crash. After the bust, Nvidia suddenly had to slash half a billion dollars from its earnings projection, and market experts later estimated that the firm had understated its crypto-related revenue by more than a billion. In 2022, Nvidia paid a $5.5 million SEC penalty over the inadequate disclosures that one SEC chief said “deprived investors of critical information to evaluate the company’s business in a key market.”

Investors, however, have not yet settled their own legal challenges. In both cases, investors suing convinced the 9th Circuit that the companies were guilty of misleading investors. But now, the tech companies have appealed to the Supreme Court, hoping to reverse those rulings.

In case documents, each claimed that their investors have not satisfied high legal bars, which Nvidia argued Congress designed to prevent “frivolous” or “nuisance” lawsuits from going on “fishing expeditions” to claim securities “fraud by hindsight.” Both warned that SCOTUS upholding the 9th Circuit rulings risked flooding courts with frivolous suits, with Nvidia cautioning that such lawsuits can be “used to injure the entire US economy.”

The Supreme Court will hear arguments in the Facebook case on Wednesday, November 6, then the Nvidia case on November 13.

SCOTUS may be persuaded by tech companies still stuck coping with the aftermath of scandals. A former SEC lawyer, Andrew Feller, told Reuters that the Supreme Court’s conservative majority may continue its “recent track record of handing down business-friendly decisions that narrowed the authority of federal regulators” in these cases. Both cases give justices opportunities to “rein in the power of private plaintiffs to enforce federal rules aimed at punishing corporate misconduct,” Reuters reported.

Facebook defends describing risk as hypothetical

The Facebook case centers on an SEC disclosure where Facebook said that its business may be harmed by a data breach, posing that as a hypothetical, without mentioning the ongoing Cambridge Analytica data breach. Specifically, Facebook wrote, “[a]ny failure to prevent or mitigate . . . improper access to or disclosure of our data or user data . . . could result in the loss or misuse of such data, which could harm our business and reputation and diminish our competitive position.”

Investors felt misled, accusing Facebook of hiding the breach by only presenting the risk as a hypothetical that implied no breach had ever occurred in the past and certainly did not disclose the present risk.

However, in a SCOTUS filing, Facebook insisted that “no reasonable investor would interpret a risk disclosure using probabilistic, forward-looking language as impliedly representing that the specified triggering event had never occurred in the past.”

Facebook is now arguing that SCOTUS agreeing that the company should have disclosed the major data breach “would result in a regime under which companies would be required to disclose every previous material incident they have experienced—effectively creating a sweeping regime of omissions liability.”

According to Facebook, news broke about the Cambridge Analytica data breach in 2015, and its business wasn’t immediately harmed. Following that logic, the social media company hopes that SCOTUS will agree that Facebook was only required to disclose the data breach in its SEC filing if Facebook knew its business would likely be harmed from the ongoing breach.

By affirming the 9th Circuit ruling, Facebook alleged, SCOTUS would be “vastly expanding the circumstances in which risk disclosures are deemed false or misleading,” exposing to legal challenges “a wide range of previously immune forward-looking statements—revenue projections, future business plans or objectives, and the like.”

But investors suing argue that Facebook is still being misleading about the data scandal in its court filings.

“The only reason Facebook has ever given to explain why the misappropriation risked no harm was that the event was allegedly disclosed to the public in 2015 and no one cared,” investors’ SCOTUS brief said. But in 2015, a report exposing a data breach tied to a Ted Cruz campaign was denied by Cambridge Analytica and prompted a Facebook investigation that concluded no damage had been done.

“Facebook actively misled the public about its investigation, ‘represent[ing] that no misconduct had been discovered,'” investors alleged, and “Facebook’s deception extended to its public filings with the SEC.”

According to investors, the real damage was done when the true extent of the Cambridge Analytica scandal was exposed in 2018. That caused substantial revenue losses that Facebook likely understood it was risking while allegedly leaving investors blind to those risks for years.

Investors argue that disclosure should not be required of every data breach that hits Facebook, whether it harms its business or not, but that the Cambridge Analytica data breach was significant and should have been disclosed as a material risk. The 9th Circuit agreed, holding that “publicly treating such a material adverse event as a merely hypothetical prospect can be misleading even if the event has not yet produced follow-on business harm because the company has kept the truth from the public.”

They further argued that requiring so-called overdisclosure wouldn’t trigger unwarranted litigation, as Facebook suggests, because Congress has always “given considerable attention to concerns over abusive private litigation.”

If Facebook wins, investors alleged, SCOTUS risks giving any tech company “a license to intentionally mislead investors about the occurrence of hugely material events by describing those events as purely hypothetical prospects.” Siding with Facebook would allegedly give “companies an incentive to stuff their annual reports with boilerplate, generic warnings that reveal little about the company’s actual business and to cover up events that could give rise to corporate scandals, as Facebook did here.”

Facebook argued that if the SEC is concerned about specific disclosures connected to the data breach, “the SEC can invoke the rulemaking process to impose” a requirement that companies must disclose all “past material adverse events.”

Nvidia disputes expert’s crypto data

While the Facebook case involved a bigger scandal, the Nvidia case could have bigger legal implications if Nvidia wins.

In the Nvidia case, investors argued that Nvidia CEO Jensen Huang made public statements allegedly misleading investors by downplaying the high demand for GPUs tied to volatile crypto markets. To plead their case, investors relied on statements from Nvidia employees, internal documents like meeting slides, industry research, as well as an expert opinion crunching general market numbers and estimating that Nvidia “underreported its crypto revenues by $1.126 billion.”

Nvidia claimed it’s far more plausible that the company simply made an “honest miscalculation” while navigating a complex emerging market.

To defend against the suit, Nvidia is arguing that the Private Securities Litigation Reform Act (PSLRA) imposes “special burdens on plaintiffs seeking to bring federal securities fraud class actions” through “heightened pleading requirements” to deter frivolous lawsuits arguing fraud by hindsight.

According to Nvidia, the PSLRA requires investors to allege particular facts based on particular contents of internal Nvidia documents, which goes beyond relying on an expert opinion. The tech company has urged SCOTUS that the 9th Circuit “‘significantly erode[d]” the PSLRA requirements by allowing Plaintiffs to “simply” hire “an expert who manufactured data to fit their allegations.”

“They hired an expert to create data and then filed a class action alleging that Nvidia and its CEO committed securities fraud by failing to disclose the data invented by Plaintiffs’ expert,” Nvidia argued.

This allegedly “eviscerates the guardrails that Congress erected to protect the public from abusive securities litigation” and creates a “dangerous” and “easy-to-replicate ‘roadmap’ for plaintiffs to sidestep the PSLRA in this recurring context.”

“Far from serving Congress’s goal of guarding against fishing expeditions by vexatious litigants, the Ninth Circuit’s opinion declares it open season so long as a plaintiff has funding to hire an expert,” Nvidia alleged.

Investors are hoping SCOTUS will uphold the 9th Circuit’s judgment. Instead of seeing their suit as frivolous, they argued that the SEC fine over the same misconduct “undermines any suggestion that this is the type of frivolous suit that the PSLRA was meant to screen out.”

They’ve disputed Nvidia’s arguments that they’ve relied solely on a hired expert to support their claims, arguing that each fact was corroborated by employee witnesses and third-party reports.

If Nvidia wins, investors warned, the SCOTUS decision would risk harming a wide range of private securities litigation that Congress has found “‘is an indispensable tool’ for ‘defrauded investors’ to ‘recover their losses without having to rely upon government action.'”

Photo of Ashley Belanger

Ashley is a senior policy reporter for Ars Technica, dedicated to tracking social impacts of emerging policies and new technologies. She is a Chicago-based journalist with 20 years of experience.

Facebook, Nvidia push SCOTUS to limit “nuisance” investor suits after scandals Read More »

us-suspects-tsmc-helped-huawei-skirt-export-controls,-report-says

US suspects TSMC helped Huawei skirt export controls, report says

In April, TSMC was provided with $6.6 billion in direct CHIPS Act funding to “support TSMC’s investment of more than $65 billion in three greenfield leading-edge fabs in Phoenix, Arizona, which will manufacture the world’s most advanced semiconductors,” the Department of Commerce said.

These investments are key to the Biden-Harris administration’s mission of strengthening “economic and national security by providing a reliable domestic supply of the chips that will underpin the future economy, powering the AI boom and other fast-growing industries like consumer electronics, automotive, Internet of Things, and high-performance computing,” the department noted. And in particular, the funding will help America “maintain our competitive edge” in artificial intelligence, the department said.

It likely wouldn’t make sense to prop TSMC up to help the US “onshore the critical hardware manufacturing capabilities that underpin AI’s deep language learning algorithms and inferencing techniques,” to then limit access to US-made tech. TSMC’s Arizona fabs are supposed to support companies like Apple, Nvidia, and Qualcomm and enable them to “compete effectively,” the Department of Commerce said.

Currently, it’s unclear where the US probe into TSMC will go or whether a damaging finding could potentially impact TSMC’s CHIPS funding.

Last fall, the Department of Commerce published a final rule, though, designed to “prevent CHIPS funds from being used to directly or indirectly benefit foreign countries of concern,” such as China.

If the US suspected that TSMC was aiding Huawei’s AI chip manufacturing, the company could be perceived as avoiding CHIPS guardrails prohibiting TSMC from “knowingly engaging in any joint research or technology licensing effort with a foreign entity of concern that relates to a technology or product that raises national security concerns.”

Violating this “technology clawback” provision of the final rule risks “the full amount” of CHIPS Act funding being “recovered” by the Department of Commerce. That outcome seems unlikely, though, given that TSMC has been awarded more funding than any other recipient apart from Intel.

The Department of Commerce declined Ars’ request to comment on whether TSMC’s CHIPS Act funding could be impacted by their reported probe.

US suspects TSMC helped Huawei skirt export controls, report says Read More »

amd-unveils-powerful-new-ai-chip-to-challenge-nvidia

AMD unveils powerful new AI chip to challenge Nvidia

On Thursday, AMD announced its new MI325X AI accelerator chip, which is set to roll out to data center customers in the fourth quarter of this year. At an event hosted in San Francisco, the company claimed the new chip offers “industry-leading” performance compared to Nvidia’s current H200 GPUs, which are widely used in data centers to power AI applications such as ChatGPT.

With its new chip, AMD hopes to narrow the performance gap with Nvidia in the AI processor market. The Santa Clara-based company also revealed plans for its next-generation MI350 chip, which is positioned as a head-to-head competitor of Nvidia’s new Blackwell system, with an expected shipping date in the second half of 2025.

In an interview with the Financial Times, AMD CEO Lisa Su expressed her ambition for AMD to become the “end-to-end” AI leader over the next decade. “This is the beginning, not the end of the AI race,” she told the publication.

The AMD Instinct MI325X Accelerator.

The AMD Instinct MI325X Accelerator.

The AMD Instinct MI325X Accelerator. Credit: AMD

According to AMD’s website, the announced MI325X accelerator contains 153 billion transistors and is built on the CDNA3 GPU architecture using TSMC’s 5 nm and 6 nm FinFET lithography processes. The chip includes 19,456 stream processors and 1,216 matrix cores spread across 304 compute units. With a peak engine clock of 2100 MHz, the MI325X delivers up to 2.61 PFLOPs of peak eight-bit precision (FP8) performance. For half-precision (FP16) operations, it reaches 1.3 PFLOPs.

AMD unveils powerful new AI chip to challenge Nvidia Read More »

ars-technica-system-guide:-falling-prices-are-more-exciting-than-new-parts

Ars Technica system guide: Falling prices are more exciting than new parts

AMD's Ryzen 7700X makes enough sense to feature in our higher-end gaming build.

Enlarge / AMD’s Ryzen 7700X makes enough sense to feature in our higher-end gaming build.

Andrew Cunningham

It’s been a while since our last system guide, and a few new products—most notably AMD’s Ryzen 9000 series CPUs—have been released since then. But there haven’t been many notable graphics card launches, and new ones are still rumored to be a few months off as both Nvidia and AMD prioritize their money-printing AI accelerators.

But that doesn’t make it a bad time to buy a PC, especially if you’re looking for some cost-efficient builds. Prices of CPUs and GPUs have both fallen a fair bit since we did our last build guide a year or so ago, which means all of our builds are either cheaper than they were before or we can squeeze out a little more performance than before at similar prices.

We have six builds across four broad tiers—a budget office desktop, a budget 1080p gaming PC, a mainstream 1440p-to-4K gaming PC, and a price-conscious workstation build with a powerful CPU and lots of room for future expandability.

You won’t find a high-end “god box” this time around, though; for a money-is-no-object high-end build, it’s probably worth waiting for Intel’s upcoming Arrow Lake desktop processors, AMD’s expected Ryzen 9000X3D series, and whatever Nvidia’s next-generation GPU launch is. All three of those things are expected either later this year or early next.

We have a couple of different iterations of the more expensive builds, and we also suggest multiple alternate components that can make more sense for certain types of builds based on your needs. The fun of PC building is how flexible and customizable it is—whether you want to buy what we recommend and put it together or want to treat these configurations as starting points, hopefully, they give you some idea of what your money can get you right now.

Notes on component selection

Part of the fun of building a PC is making it look the way you want. We’ve selected cases that will physically fit the motherboards and other parts we’re recommending and which we think will be good stylistic fits for each system. But there are many cases out there, and our picks aren’t the only options available.

As for power supplies, we’re looking for 80 Plus certified power supplies from established brands with positive user reviews on retail sites (or positive professional reviews, though these can be somewhat hard to come by for any given PSU these days). If you have a preferred brand, by all means, go with what works for you. The same goes for RAM—we’ll recommend capacities and speeds, and we’ll link to kits from brands that have worked well for us in the past, but that doesn’t mean they’re better than the many other RAM kits with equivalent specs.

For SSDs, we mostly stick to drives from known brands like Samsung, Crucial, or Western Digital, though going with a lesser-known brand can save you a bit of money. All of our builds also include built-in Bluetooth and Wi-Fi, so you don’t need to worry about running Ethernet wires and can easily connect to Bluetooth gamepads, keyboards, mice, headsets, and other accessories.

We also haven’t priced in peripherals, like webcams, monitors, keyboards, or mice, as we’re assuming most people will re-use what they already have or buy those components separately. If you’re feeling adventurous, you could even make your own DIY keyboard! If you need more guidance, Kimber Streams’ Wirecutter keyboard guides are exhaustive and educational.

Finally, we won’t be including the cost of a Windows license in our cost estimates. You can pay a lot of different prices for Windows—$139 for an official retail license from Microsoft, $120 for an “OEM” license for system builders, or anywhere between $15 and $40 for a product key from shady gray market product key resale sites. Windows 10 keys will also work to activate Windows 11, though Microsoft stopped letting old Windows 7 and Windows 8 keys activate new Windows 10 and 11 installs relatively recently. You could even install Linux, given recent advancements to game compatibility layers!

Ars Technica system guide: Falling prices are more exciting than new parts Read More »

nvidia’s-ai-chips-are-cheaper-to-rent-in-china-than-us

Nvidia’s AI chips are cheaper to rent in China than US

secondhand channels —

Supply of processors helps Chinese startups advance AI technology despite US restrictions.

Nvidia’s AI chips are cheaper to rent in China than US

VGG | Getty Images

The cost of renting cloud services using Nvidia’s leading artificial intelligence chips is lower in China than in the US, a sign that the advanced processors are easily reaching the Chinese market despite Washington’s export restrictions.

Four small-scale Chinese cloud providers charge local tech groups roughly $6 an hour to use a server with eight Nvidia A100 processors in a base configuration, companies and customers told the Financial Times. Small cloud vendors in the US charge about $10 an hour for the same setup.

The low prices, according to people in the AI and cloud industry, are an indication of plentiful supply of Nvidia chips in China and the circumvention of US measures designed to prevent access to cutting-edge technologies.

The A100 and H100, which is also readily available, are among Nvidia’s most powerful AI accelerators and are used to train the large language models that power AI applications. The Silicon Valley company has been banned from shipping the A100 to China since autumn 2022 and has never been allowed to sell the H100 in the country.

Chip resellers and tech startups said the products were relatively easy to procure. Inventories of the A100 and H100 are openly advertised for sale on Chinese social media and ecommerce sites such as Xiaohongshu and Alibaba’s Taobao, as well as in electronics markets, at slight markups to pricing abroad.

China’s larger cloud operators such as Alibaba and ByteDance, known for their reliability and security, charge double to quadruple the price of smaller local vendors for similar Nvidia A100 servers, according to pricing from the two operators and customers.

After discounts, both Chinese tech giants offer packages for prices comparable to Amazon Web Services, which charges $15 to $32 an hour. Alibaba and ByteDance did not respond to requests for comment.

“The big players have to think about compliance, so they are at a disadvantage. They don’t want to use smuggled chips,” said a Chinese startup founder. “Smaller vendors are less concerned.”

He estimated there were more than 100,000 Nvidia H100 processors in the country based on their widespread availability in the market. The Nvidia chips are each roughly the size of a book, making them relatively easy for smugglers to ferry across borders, undermining Washington’s efforts to limit China’s AI progress.

“We bought our H100s from a company that smuggled them in from Japan,” said a startup founder in the automation field who paid about 500,000 yuan ($70,000) for two cards this year. “They etched off the serial numbers.”

Nvidia said it sold its processors “primarily to well-known partners … who work with us to ensure that all sales comply with US export control rules”.

“Our pre-owned products are available through many second-hand channels,” the company added. “Although we cannot track products after they are sold, if we determine that any customer is violating US export controls, we will take appropriate action.”

The head of a small Chinese cloud vendor said low domestic costs helped offset the higher prices that providers paid for smuggled Nvidia processors. “Engineers are cheap, power is cheap, and competition is fierce,” he said.

In Shenzhen’s Huaqiangbei electronics market, salespeople speaking to the FT quoted the equivalent of $23,000–$30,000 for Nvidia’s H100 plug-in cards. Online sellers quote the equivalent of $31,000–$33,000.

Nvidia charges customers $20,000–$23,000 for H100 chips after recently cutting prices, according to Dylan Patel of SemiAnalysis.

One data center vendor in China said servers made by Silicon Valley’s Supermicro and fitted with eight H100 chips hit a peak selling price of 3.2 million yuan after the Biden administration tightened export restrictions in October. He said prices had since fallen to 2.5 million yuan as supply constraints eased.

Several people involved in the trade said merchants in Malaysia, Japan, and Indonesia often shipped Supermicro servers or Nvidia processors to Hong Kong before bringing them across the border to Shenzhen.

The black market trade depends on difficult-to-counter workarounds to Washington’s export regulations, experts said.

For example, while subsidiaries of Chinese companies are banned from buying advanced AI chips outside the country, their executives could establish new companies in countries such as Japan or Malaysia to make the purchases.

“It’s hard to completely enforce export controls beyond the US border,” said an American sanctions expert. “That’s why the regulations create obligations for the shipper to look into end users and [the] commerce [department] adds companies believed to be flouting the rules to the [banned] entity list.”

Additional reporting by Michael Acton in San Francisco.

© 2024 The Financial Times Ltd. All rights reserved. Please do not copy and paste FT articles and redistribute by email or post to the web.

Nvidia’s AI chips are cheaper to rent in China than US Read More »

doj-subpoenas-nvidia-in-deepening-ai-antitrust-probe,-report-says

DOJ subpoenas Nvidia in deepening AI antitrust probe, report says

DOJ subpoenas Nvidia in deepening AI antitrust probe, report says

The Department of Justice is reportedly deepening its probe into Nvidia. Officials have moved on from merely questioning competitors to subpoenaing Nvidia and other tech companies for evidence that could substantiate allegations that Nvidia is abusing its “dominant position in AI computing,” Bloomberg reported.

When news of the DOJ’s probe into the trillion-dollar company was first reported in June, Fast Company reported that scrutiny was intensifying merely because Nvidia was estimated to control “as much as 90 percent of the market for chips” capable of powering AI models. Experts told Fast Company that the DOJ probe might even be good for Nvidia’s business, noting that the market barely moved when the probe was first announced.

But the market’s confidence seemed to be shaken a little more on Tuesday, when Nvidia lost a “record-setting $279 billion” in market value following Bloomberg’s report. Nvidia’s losses became “the biggest single-day market-cap decline on record,” TheStreet reported.

People close to the DOJ’s investigation told Bloomberg that the DOJ’s “legally binding requests” require competitors “to provide information” on Nvidia’s suspected anticompetitive behaviors as a “dominant provider of AI processors.”

One concern is that Nvidia may be giving “preferential supply and pricing to customers who use its technology exclusively or buy its complete systems,” sources told Bloomberg. The DOJ is also reportedly probing Nvidia’s acquisition of RunAI—suspecting the deal may lock RunAI customers into using Nvidia chips.

Bloomberg’s report builds on a report last month from The Information that said that Advanced Micro Devices Inc. (AMD) and other Nvidia rivals were questioned by the DOJ—as well as third parties who could shed light on whether Nvidia potentially abused its market dominance in AI chips to pressure customers into buying more products.

According to Bloomberg’s sources, the DOJ is worried that “Nvidia is making it harder to switch to other suppliers and penalizes buyers that don’t exclusively use its artificial intelligence chips.”

In a statement to Bloomberg, Nvidia insisted that “Nvidia wins on merit, as reflected in our benchmark results and value to customers, who can choose whatever solution is best for them.” Additionally, Bloomberg noted that following a chip shortage in 2022, Nvidia CEO Jensen Huang has said that his company strives to prevent stockpiling of Nvidia’s coveted AI chips by prioritizing customers “who can make use of his products in ready-to-go data centers.”

Potential threats to Nvidia’s dominance

Despite the slump in shares, Nvidia’s market dominance seems unlikely to wane any time soon after its stock more than doubled this year. In an SEC filing this year, Nvidia bragged that its “accelerated computing ecosystem is bringing AI to every enterprise” with an “ecosystem” spanning “nearly 5 million developers and 40,000 companies.” Nvidia specifically highlighted that “more than 1,600 generative AI companies are building on Nvidia,” and according to Bloomberg, Nvidia will close out 2024 with more profits than the total sales of its closest competitor, AMD.

After the DOJ’s most recent big win, which successfully proved that Google has a monopoly on search, the DOJ appears intent on getting ahead of any tech companies’ ambitions to seize monopoly power and essentially become the Google of the AI industry. In June, DOJ antitrust chief Jonathan Kanter confirmed to the Financial Times that the DOJ is examining “monopoly choke points and the competitive landscape” in AI beyond just scrutinizing Nvidia.

According to Kanter, the DOJ is scrutinizing all aspects of the AI industry—”everything from computing power and the data used to train large language models, to cloud service providers, engineering talent and access to essential hardware such as graphics processing unit chips.” But in particular, the DOJ appears concerned that GPUs like Nvidia’s advanced AI chips remain a “scarce resource.” Kanter told the Financial Times that an “intervention” in “real time” to block a potential monopoly could be “the most meaningful intervention” and the least “invasive” as the AI industry grows.

DOJ subpoenas Nvidia in deepening AI antitrust probe, report says Read More »

nvidia-is-ditching-dedicated-g-sync-modules-to-push-back-against-freesync’s-ubiquity

Nvidia is ditching dedicated G-Sync modules to push back against FreeSync’s ubiquity

sync or swim —

But G-Sync will still require specific G-Sync-capable MediaTek scaler chips.

Nvidia is ditching dedicated G-Sync modules to push back against FreeSync’s ubiquity

Nvidia

Back in 2013, Nvidia introduced a new technology called G-Sync to eliminate screen tearing and stuttering effects and reduce input lag when playing PC games. The company accomplished this by tying your display’s refresh rate to the actual frame rate of the game you were playing, and similar variable refresh-rate (VRR) technology has become a mainstay even in budget monitors and TVs today.

The issue for Nvidia is that G-Sync isn’t what has been driving most of that adoption. G-Sync has always required extra dedicated hardware inside of displays, increasing the costs for both users and monitor manufacturers. The VRR technology in most low-end to mid-range screens these days is usually some version of the royalty-free AMD FreeSync or the similar VESA Adaptive-Sync standard, both of which provide G-Sync’s most important features without requiring extra hardware. Nvidia more or less acknowledged that the free-to-use, cheap-to-implement VRR technologies had won in 2019 when it announced its “G-Sync Compatible” certification tier for FreeSync monitors. The list of G-Sync Compatible screens now vastly outnumbers the list of G-Sync and G-Sync Ultimate screens.

Today, Nvidia is announcing a change that’s meant to keep G-Sync alive as its own separate technology while eliminating the requirement for expensive additional hardware. Nvidia says it’s partnering with chipmaker MediaTek to build G-Sync capabilities directly into scaler chips that MediaTek is creating for upcoming monitors. G-Sync modules ordinarily replace these scaler chips, but they’re entirely separate boards with expensive FPGA chips and dedicated RAM.

These new MediaTek scalers will support all the same features that current dedicated G-Sync modules do. Nvidia says that three G-Sync monitors with MediaTek scaler chips inside will launch “later this year”: the Asus ROG Swift PG27AQNR, the Acer Predator XB273U F5, and the AOC AGON PRO AG276QSG2. These are all 27-inch 1440p displays with maximum refresh rates of 360 Hz.

As of this writing, none of these companies has announced pricing for these displays—the current Asus PG27AQN has a traditional G-Sync module and a 360 Hz refresh rate and currently goes for around $800, so we’d hope for the new version to be significantly cheaper to make good on Nvidia’s claim that the MediaTek chips will reduce costs (or, if they do reduce costs, whether monitor makers are willing to pass those savings on to consumers).

For most people most of the time, there won’t be an appreciable difference between a “true” G-Sync monitor and one that uses FreeSync or Adaptive-Sync, but there are still a few fringe benefits. G-Sync monitors support a refresh rate between 1 and the maximum refresh rate of the monitor, whereas FreeSync and Adaptive-Sync stop working on most displays when the frame rate drops below 40 or 48 frames per second. All G-Sync monitors also support “variable overdrive” technology to help eliminate display ghosting, and the new MediaTek-powered displays will support the recent “G-Sync Pulsar” feature to reduce blur.

Nvidia is ditching dedicated G-Sync modules to push back against FreeSync’s ubiquity Read More »

amd-signs-$4.9-billion-deal-to-challenge-nvidia’s-ai-infrastructure-lead

AMD signs $4.9 billion deal to challenge Nvidia’s AI infrastructure lead

chip wars —

Company hopes acquisition of ZT Systems will accelerate adoption of its data center chips.

Visitors walk past the AMD booth at the 2024 Mobile World Congress

AMD has agreed to buy artificial intelligence infrastructure group ZT Systems in a $4.9 billion cash and stock transaction, extending a run of AI investments by the chip company as it seeks to challenge market leader Nvidia.

The California-based group said the acquisition would help accelerate the adoption of its Instinct line of AI data center chips, which compete with Nvidia’s popular graphics processing units (GPUs).

ZT Systems, a private company founded three decades ago, builds custom computing infrastructure for the biggest AI “hyperscalers.” While the company does not disclose its customers, the hyperscalers include the likes of Microsoft, Meta, and Amazon.

The deal marks AMD’s biggest acquisition since it bought Xilinx for $35 billion in 2022.

“It brings a thousand world-class design engineers into our team, it allows us to develop silicon and systems in parallel and, most importantly, get the newest AI infrastructure up and running in data centers as fast as possible,” AMD’s chief executive Lisa Su told the Financial Times.

“It really helps us deploy our technology much faster because this is what our customers are telling us [they need],” Su added.

The transaction is expected to close in the first half of 2025, subject to regulatory approval, after which New Jersey-based ZT Systems will be folded into AMD’s data center business group. The $4.9bn valuation includes up to $400mn contingent on “certain post-closing milestones.”

Citi and Latham & Watkins are advising AMD, while ZT Systems has retained Goldman Sachs and Paul, Weiss.

The move comes as AMD seeks to break Nvidia’s stranglehold on the AI data center chip market, which earlier this year saw Nvidia temporarily become the world’s most valuable company as big tech companies pour billions of dollars into its chips to train and deploy powerful new AI models.

Part of Nvidia’s success stems from its “systems” approach to the AI chip market, offering end-to-end computing infrastructure that includes pre-packaged server racks, networking equipment, and software tools to make it easier for developers to build AI applications on its chips.

AMD’s acquisition shows the chipmaker building out its own “systems” offering. The company rolled out its MI300 line of AI chips last year, and says it will launch its next-generation MI350 chip in 2025 to compete with Nvidia’s new Blackwell line of GPUs.

In May, Microsoft was one of the first AI hyperscalers to adopt the MI300, building it into its Azure cloud platform to run AI models such as OpenAI’s GPT-4. AMD’s quarterly revenue for the chips surpassed $1 billion for the first time in the three months to June 30.

But while AMD has feted the MI300 as its fastest-ever product ramp, its data center revenue still represented a fraction of the $22.6 billion that Nvidia’s data center business raked in for the quarter to the end of April.

In March, ZT Systems announced a partnership with Nvidia to build custom AI infrastructure using its Blackwell chips. “I think we certainly believe ZT as part of AMD will significantly accelerate the adoption of AMD AI solutions,” Su said, but “we have customer commitments and we are certainly going to honour those”.

Su added that she expected regulators’ review of the deal to focus on the US and Europe.

In addition to increasing its research and development spending, AMD says it has invested more than $1 billion over the past year to expand its AI hardware and software ecosystem.

In July the company announced it was acquiring Finnish AI start-up Silo AI for $665 million, the largest acquisition of a privately held AI startup in Europe in a decade.

© 2024 The Financial Times Ltd. All rights reserved. Please do not copy and paste FT articles and redistribute by email or post to the web.

AMD signs $4.9 billion deal to challenge Nvidia’s AI infrastructure lead Read More »