Jump to content

Nvidia: Difference between revisions

From Consumer Rights Wiki
← Older editNewer edit →
VisualWikitext
Azaydan (talk | contribs)
confirmed
5 edits
Nvidia limits FP64 compute on consumer GPUs via firmware despite sharing identical silicon with enterprise Tesla/Quadro cards.
Azaydan (talk | contribs)
confirmed
5 edits
Corrections about the Artificial FP64 Limitation on Consumer GPUs
Line 21: Line 21:
In Nvidia's terms of service regarding accessing their website, under "Informal Resolution" users are required to agree to resolve legal disputes with Nvidia by arbitration from Judicial Arbitration and Mediation Services (JAMS).<ref>{{Cite web |title=NVIDIA Legal Notices |url=https://www.nvidia.com/en-gb/about-nvidia/terms-of-service/ |access-date=2025-06-19 |website=NVIDIA |language=en-gb |archive-url=http://web.archive.org/web/20260221043438/https://www.nvidia.com/en-gb/about-nvidia/terms-of-service/ |archive-date=21 Feb 2026}}</ref>
In Nvidia's terms of service regarding accessing their website, under "Informal Resolution" users are required to agree to resolve legal disputes with Nvidia by arbitration from Judicial Arbitration and Mediation Services (JAMS).<ref>{{Cite web |title=NVIDIA Legal Notices |url=https://www.nvidia.com/en-gb/about-nvidia/terms-of-service/ |access-date=2025-06-19 |website=NVIDIA |language=en-gb |archive-url=http://web.archive.org/web/20260221043438/https://www.nvidia.com/en-gb/about-nvidia/terms-of-service/ |archive-date=21 Feb 2026}}</ref>


===Artificial FP64 Limitation on Consumer GPUs (''~2013 - Present'')===
===Artificial FP64 Limitation on Consumer GPUs (''2012 - 2014'')===
Nvidia manufactures consumer GPUs that share identical or near-identical silicon with their enterprise Tesla and Quadro product lines, yet artificially restricts double precision floating point (FP64) compute performance on consumer variants through firmware and driver configuration rather than hardware incapability. The GTX 780 and the Tesla K40, for instance, both use the GK110 die, yet the Tesla K40 delivers FP64 performance at 1/3 the rate of FP32, while the GTX 780 is limited to 1/24.<ref>{{Cite web |last=ArrayFire |date= |title=Explaining FP64 performance on GPUs |url=https://arrayfire.com/blog/explaining-fp64-performance-on-gpus/ |url-status=live |access-date=2026-03-02 |website=ArrayFire}}</ref> The original GeForce GTX Titan briefly shipped with near-full FP64 capability enabled, serving both as Nvidia's consumer flagship and entry-level FP64 compute card simultaneously. From Maxwell onward, Nvidia permanently dropped FP64 on consumer parts to 1/32 of FP32 performance, fully divorcing consumer and enterprise silicon feature sets despite continued die sharing.<ref>{{Cite web |last=Angelini |first=Chris |date=2015-03-17 |title=GM200 - All Graphics, Hold The Double Precision |url=https://www.anandtech.com/show/9059/the-nvidia-geforce-gtx-titan-x-review/2 |url-status=live |access-date=2026-03-02 |website=AnandTech}}</ref> This practice steers users requiring FP64 compute toward enterprise products that can cost several times more than equivalent consumer hardware built on the same die.
On Kepler-era GK110 graphics cards, Nvidia used a driver-imposed clock rate restriction to artificially limit double precision floating point (FP64) compute performance on consumer variants, despite the hardware being fully capable. The GTX 780 and GTX 780 Ti share the GK110 die with the Tesla K40, which delivers FP64 at 1/3 the rate of FP32. On the consumer cards, the driver deliberately clocked the FP64 units at 1/8 of the chip's base rate, yielding a 1/24 ratio.<ref>{{Cite web |last=Angelini |first=Chris |date=2013-11-07 |title=Nvidia GeForce GTX 780 Ti Review: GK110, Fully Unlocked |url=https://www.tomshardware.com/reviews/geforce-gtx-780-ti-review-benchmarks,3663.html |url-status=live |access-date=2026-03-02 |website=Tom's Hardware}}</ref> The GTX Titan, also based on GK110, shipped with a driver control panel toggle that could unlock the full 1/3 FP64 rate at the cost of disabling GPU Boost, confirming the limitation was a software restriction rather than a hardware one.<ref>{{Cite web |last=Angelini |first=Chris |date=2013-02-19 |title=Compute Performance And Striking A Balance - Nvidia GeForce GTX Titan 6 GB: GK110 On A Gaming Card |url=https://www.tomshardware.com/reviews/geforce-gtx-titan-gk110-review,3438-3.html |url-status=live |access-date=2026-03-02 |website=Tom's Hardware}}</ref> From the Maxwell architecture onward, Nvidia switched to physically reducing the number of FP64 cores on consumer dies, ending the driver restriction but continuing hardware-level segmentation between consumer and enterprise products.


===Stagnation of Consumer GPU Offerings===
===Stagnation of Consumer GPU Offerings===

Revision as of 05:06, 2 March 2026

Nvidia
Basic information
Founded 1993-04-05
Legal Structure Public
Industry Semiconductors, Software
Also known as
Official website https://www.nvidia.com/

Nvidia Corporation is an American technology company that designs and sells computer components such as graphics processing units (GPUs) for both commercial and enterprise use. It was founded on April 5, 1993, by current CEO (as of 2026) Jensen Huang, Chris Malachowsky, and Curtis Priem. The company is the largest providers of GPUs for both consumer and enterprise.

Consumer-impact summary

As Nvidia's market share in consumer GPUs rose to >90%, hence becoming the market leader,[1] and their enterprise wing started outperforming the consumer wing, their perception of consumers waned alongside. This has lead to the prioritization of enterprise products while the consumer products become less appleaing. Nvidia on multiple accounts has employed deceptive practices such as to mislead the everyday buyer and even enthusiasts. This has led to many people developing a sour taste towards the brand and ultimately branding them as a monopoly.

Nvidia is notable for heavy proprietization and vendor-locking practices, manifesting in technologies such as the CUDA compute platform, PhysX physics simulation and DLSS upscaling, which are exclusively available on their devices.

Incidents

This is a list of all consumer-protection incidents this company is involved in. Any incidents not mentioned here can be found in the Nvidia category.

Forced Arbitration

In Nvidia's terms of service regarding accessing their website, under "Informal Resolution" users are required to agree to resolve legal disputes with Nvidia by arbitration from Judicial Arbitration and Mediation Services (JAMS).[2]

Artificial FP64 Limitation on Consumer GPUs (2012 - 2014)

On Kepler-era GK110 graphics cards, Nvidia used a driver-imposed clock rate restriction to artificially limit double precision floating point (FP64) compute performance on consumer variants, despite the hardware being fully capable. The GTX 780 and GTX 780 Ti share the GK110 die with the Tesla K40, which delivers FP64 at 1/3 the rate of FP32. On the consumer cards, the driver deliberately clocked the FP64 units at 1/8 of the chip's base rate, yielding a 1/24 ratio.[3] The GTX Titan, also based on GK110, shipped with a driver control panel toggle that could unlock the full 1/3 FP64 rate at the cost of disabling GPU Boost, confirming the limitation was a software restriction rather than a hardware one.[4] From the Maxwell architecture onward, Nvidia switched to physically reducing the number of FP64 cores on consumer dies, ending the driver restriction but continuing hardware-level segmentation between consumer and enterprise products.

Stagnation of Consumer GPU Offerings

Since assuming its dominant position at >90% of the consumer GPU market, Nvidia has been credibly accused of shrinkflation across its entire modern GeForce product line by multiple renouned yet credible reviewers.[5][6][7]

GeForce Partner Program (2018.03.01 - 2018.05.04)

Main article: [[GeForce Partner Program]]

On March 1, 2018, Nvidia in a blog post first announced the marketing program[8] which aimed to provide partners with benefits such PR support, video game bundling and market development funds. Kyle Bennett from HardOCP spoke with seven companies off the record and they more or less had the same to say:

  • The terms of the GPP agreement are potentially illegal
  • The GPP will hurt consumer choices
  • The GPP will hurt a partner's ability to do business with other companies like AMD and Intel[9]

Nvidia cancelled the program on May 4, 2018, citing "The rumors, conjecture and mistruths go far beyond its intent. Rather than battling misinformation, we have decided to cancel the program." [10][11]

12VHPWR/12V-2x6 (H++) Connector Failures (2022 - Present)

Main article: PCIe power connector design linked to melting GPU cables

12VHPWR and its revision 12V-2x6 (H++) are modern PCIe power connector standards published by the PCI-SIG that aim to deliver more power to PCIe devices such as GPUs but in a smaller foorprint than the older PCIe 6- and 8-pin connectors.[citation needed] Since 12VHPWR's debut on Nvidia's RTX 40-series GPUs,[12] the connector's been infamous for its design flaws that make it quite prone to ignition.[12][13] The minor revision 12V-2x6 aimed to mitigate these issues but ultimately succumbed to its predecessor's shortcomings.[12] As of July 2025, there are still reports of the connectors catching on fire even when using accessories that advertise less risk of catching ablaze.[citation needed]

Data scraping without permission for AI training (2024 - Present)

Nvidia is heavily involved in training AI models though plagiarism.[14][15] In January 2026, a court filing revealed that Nvidia developers contacted Anna's Archive to evaluate the use of pirated content for training, which management gave green-light for despite legality concerns.[16]

RTX 50-series GPUs Missing ROPs (2025 - Present)

Following the release of the RTX 50 series GPUs, several RTX 5090 and 5080 cards were noticed to be missing Raster Operation Units (ROPs), which are fundamental part for rendering and caused affected cards to suffer from lower performance.

Threatening Hardware Unboxed (2020.12.10)

On 10 December, 2020, Hardware Unboxed, a pro consumer tech YouTube channel, tweeted on Twitter (now X), "Nvidia have officially decided to ban us from receiving GeForce Founders Edition GPU review samples. Their reasoning is that we are focusing on rasterization instead of ray tracing. They have said they will revisit this 'should your editorial direction change.' "

In emails that were disclosed by Walton from Nvidia Senior PR Manager Bryan Del Rizzo, Nvidia had said:

...your GPU reviews and recommendations have continued to focus singularly on rasterization performance, and you have largely discounted all of the other technologies we offer gamers. It is very clear from your community commentary that you do not see things the same way that we, gamers, and the rest of the industry do.

On 12 December, 2020, as a result of immense consumer and prominent creator backlash, Nvidia reversed their stance. In an email, Nvidia apologized to Hardware Unboxed and on Twitter a second apology was issued by Nvidia's Del Rizzo that said "to withhold samples because I didn't agree with your commentary is simply inexcusable and crossed the line."

On 14 December, 2020, Hardware Unboxed uploaded a video on YouTube explaining the controversy from their viewpoint.

Disingenious RTX 5070 marketing (2025.03.01 - Present)

Nvidia on stage when introducing the RTX 5070 on stage, branded it as having the same performance as a RTX 4090 but for half the price.[17] This disingenious marketing was disproven by multiple credible reviewers who showed that the 5070, even when using MFG 4x Frame Generation could not achieve parity with the 4090.[18]

GeForce previews (2025.03.01 - Present)

Starting 2025, Nvidia started the practice of pre allocating review units to select review media platforms for "preview". Previews when compared to reviews are highly restricted on which benchmarks can be presented. Most often involving the presentation of cherry-picked and biased benchmarks towards the latest model of GPU. Hence previews end up touting the same message as Nvidia's marketing. This practise was first implemented with the RTX 5070. [19][20][21][22]

Threatening Gamers Nexus (2025.05.18)

Main article: Nvidia threatens GamersNexus

On May 18, 2025, GamersNexus, a pro consumer tech publication and YouTube channel, uploaded a video to their YouTube channel exposing Nvidia's scheme of threatening them. Nvidia strongly suggested to GN that access to their engineers for future interviews would only be possible if GN underwent an editorial change according to Nvidia's specification. The company's ransom was for more postive and longer emphasis on their new MFG (Multi-frame Generation) technology in GN's future reviews of their GPUs.[23]

Acquisition of stake in Intel (2025.12.29)

Main article: Nvidia buys stake in Intel

On 18 September, 2025, Nvidia and Intel jointly announced in a press release, Nvidia's $5,000,000,000 USD investment into Intel stock.[24] Finalized on December 29, 2025, Nvidia purchased 214.8 million Intel shares at $23.28 per share.[25] Both Nvidia and Intel develop GPUs for consumer and enterprise applications, making them competitors. In both sectors Nvidia holds the highest market share, dwarfing both AMD and Intel by magnitudes. As a result many have developed concerns over the persistence of Intel's GPU offerings and have branded the investment as Nvidia's attempt at furthering their monopoly by investing in their competitor.

The U.S. Federal Trade Commission (FTC) announced in a notice that U.S. antitrust agencies cleared the investment,[26] despite this being anti-consumer and making Nvidia a monopoly.

Partnership with Palantir (2025.10.28)

During their GTC Washington keynote, Nvidia announced a strategic partnership with Palantir.[27]

Products

GeForce:

RTX 30 Series:

RTX 40 Series:

RTX 50 Series:

See also

References

  1. ↑ Peddie, Jon (2025-06-04). "Q1'25 PC GPU shipments decreased by -12% from last quarter; quarter following seasonality but still down from historical average". Jon Peddie Research. Archived from the original on 2025-07-11. Retrieved 2025-07-11.
  2. ↑ "NVIDIA Legal Notices". NVIDIA. Archived from the original on 21 Feb 2026. Retrieved 2025-06-19.
  3. ↑ Angelini, Chris (2013-11-07). "Nvidia GeForce GTX 780 Ti Review: GK110, Fully Unlocked". Tom's Hardware. Retrieved 2026-03-02.{{cite web}}: CS1 maint: url-status (link)
  4. ↑ Angelini, Chris (2013-02-19). "Compute Performance And Striking A Balance - Nvidia GeForce GTX Titan 6 GB: GK110 On A Gaming Card". Tom's Hardware. Retrieved 2026-03-02.{{cite web}}: CS1 maint: url-status (link)
  5. ↑ Burke, Steve; Clayton, Jeremy; Makhnovets, Vitalii; Thang, Jimmy (2025-04-08). "The Great NVIDIA Switcheroo | GPU Shrinkflation". GamersNexus. Archived from the original on 2025-07-02. Retrieved 2025-07-02.
  6. ↑ Nexus, Gamers (2025-06-26). "NVIDIA's Exploitation | Waste of Sand RTX "5050" for $250". YouTube. Archived from the original on 2025-07-15. Retrieved 2025-07-15.
  7. ↑ Unboxed, Hardware (2025-07-10). "Real Stagnation: 6 Years Of GeForce RTX 60 Class GPUs". YouTube. Archived from the original on 2025-07-15. Retrieved 2025-07-15.
  8. ↑ Teeple, John (2018-03-01). "GeForce Partner Program Helps Gamers Know What They're Buying". NVIDIA Blog. Archived from the original on 2023-07-10. Retrieved 2023-07-10.
  9. ↑ Scheisser, Tim (2018-03-12). "Nvidia gets anti-competitive with unsavory GeForce Partner Program". TechSpot. Archived from the original on 2025-07-03.
  10. ↑ Teeple, John (4 May 2018). "Pulling the Plug on GPP, Leaning into GeForce". Nvidia Blog. Archived from the original on 4 May 2018. Retrieved 15 Jun 2025.
  11. ↑ Forrest, Derek published (4 May 2018). "Nvidia Ends Notorious GeForce Partner Program". Tom's Hardware. Archived from the original on 25 Mar 2025. Retrieved 15 Jun 2025.
  12. ↑ 12.0 12.1 12.2 Burke, Steve; Lathan, Patrick; Makhnovets, Vitalii; Coleman, Andrew; Thang, Jimmy (2024-10-07). "12VHPWR is a Dumpster Fire | Investigation into Contradicting Specs & Corner Cutting". GamersNexus. Archived from the original on 2025-07-15. Retrieved 2025-07-15.
  13. ↑ Wallossek, Igor (2022-10-27). "The horror has a face - NVIDIA's hot 12VHPWR adapter for the GeForce RTX 4090 with a built-in breaking point". igor´sLAB. Archived from the original on 2025-07-15. Retrieved 2025-07-15.
  14. ↑ Cole, Samantha (5 Aug 2024). "Leaked Documents Show Nvidia Scraping 'A Human Lifetime' of Videos Per Day to Train AI". 404 Media. Archived from the original on 14 Apr 2025. Retrieved 15 Jun 2025.
  15. ↑ Morales, Jowi (6 Aug 2024). "Nvidia accused of scraping 'A Human Lifetime' of videos per day to train AI". Tom's Hardware. Archived from the original on 11 Feb 2025. Retrieved 15 Jun 2025.
  16. ↑ "NVIDIA Contacted Anna's Archive to Secure Access to Millions of Pirated Books". torrentfreak. 22 Jan 2026. Archived from the original on 22 Feb 2026. Retrieved 22 Jan 2026.
  17. ↑ Leadbetter, Richard (2025-01-13). "DF Weekly: can the new RTX 5070 really match RTX 4090 performance?". Digital Foundry. Archived from the original on 21 Dec 2025. Retrieved 2026-01-15.
  18. ↑ Burke, Steve; Gaglione, Mike; Makhnovets, Vitalii; Phetdara, Tim; Thang, Jimmy (2025-03-13). "NVIDIA is Selling Lies | RTX 5070 Founders Edition Review & Benchmarks". GamersNexus. Retrieved 2026-01-15.{{cite web}}: CS1 maint: url-status (link) (Archived)
  19. ↑ James, Dave (2025-03-01). "Nvidia RTX 5070 hands-on preview: a sneak peek at my unreleased RTX Blackwell card hitting triple digit frame rates in Cyberpunk 2077 at 1440p and Avowed at 4K". PC Gamer. Archived from the original on 2025-08-03. Retrieved 2025-08-03.
  20. ↑ Guyton, Christian (2025-05-17). "I tested the RTX 5060 - is 8GB of VRAM really enough in 2025?". TechRadar. Archived from the original on 2025-08-03. Retrieved 2025-08-03.
  21. ↑ Hayton, Phil (2025-05-17). "I spent my first few hours with the RTX 5060 playing Doom: The Dark Ages, and it feels pretty slick for under $300". GamesRadar+. Archived from the original on 2025-08-03. Retrieved 2025-08-03.
  22. ↑ England, Jason (2025-05-17). "I played 5 games on the RTX 5060 — is this budget GPU actually enough?". Tom's Guide. Archived from the original on 2025-08-03. Retrieved 2025-08-03.
  23. ↑ Nexus, Gamers (2025-05-18). "NVIDIA's Dirty Manipulation of Reviews". YouTube. Archived from the original on 23 Feb 2026. Retrieved 2025-07-11.
  24. ↑ "NVIDIA and Intel to Develop AI Infrastructure and Personal Computing Products". NVIDIA Newsroom. 2025-09-18. Archived from the original on 2025-11-05. Retrieved 2025-11-05.
  25. ↑ Carroll, Shannon (Dec 29, 2025). "Nvidia's $5 billion Intel bet just became official". Quartz. Archived from the original on Jan 2, 2026. Retrieved Jan 7, 2026.
  26. ↑ Bajwa, Arsheeya (2025-12-29). D'Silva, Anil (ed.). ""Nvidia takes $5 billion stake in Intel under September agreement"". Reuters. Retrieved 2026-01-07.{{cite news}}: CS1 maint: url-status (link)
  27. ↑ "Palantir and NVIDIA Team Up to Operationalize AI — Turning Enterprise Data Into Dynamic Decision Intelligence". 2025-10-28. Archived from the original on 15 Jan 2026. Retrieved 2026-01-19.
Retrieved from "https://consumerrights.wiki/index.php?title=Nvidia&oldid=40517"