AMD va HUMILIER la RTX 5090… et pour 40 % moins cher !

AMD va HUMILIER la RTX 5090… et pour 40 % moins cher !

AMD is going to HUMILIATE the RTX 5090… and for 40% less!

Title history
  • AMD va HUMILIER la RTX 5090… et pour 40 % moins cher ! — observed on 20261101-0050 UTC (YouTube)

by MAKERTRONIC TECH · View the channel on PSI

Applied Sciences & Engineering Computing & Cybersecurity UComputing and Information TechnologyUKComputer hardware
🎙 MAKERTRONIC TECH 👥 6K 📅 August 23, 2026 ⏱ 14 min 👁 14K 📄 news review 🧭 2026-11-01
Available in: English (current) Français

Keywords

AMDRadeonUDNARTX 5090GPU

Summary

The video discusses rumors about AMD’s upcoming high-end GPU, possibly the Radeon RX 10800 XT or 10090 XT, which is alleged to compete with NVIDIA’s RTX 5090 at a 40% lower price. The creator explains that AMD is shifting to a unified architecture (UDNA) that merges its gaming RDNA and compute CDNA lineups. This move aims to simplify software development and compete with NVIDIA’s CUDA ecosystem. The specs reportedly include TSMC’s 3nm N3P node, doubled execution units per compute unit (128 shader processors per CU), and a die with up to 192 compute units. Compared to the RTX 5090, it would feature 96 CUs, 12,288 shader processors, and a boost clock around 3 GHz, but lower raw TFLOPS (73 vs 105). The video also mentions improvements in ray tracing and AI via new ‘Radeon Cores’ and ‘Neural Ray’ units, as well as advanced compression to reduce VRAM usage. The expected price is between $1000 and $1200, with a launch possibly delayed to late 2027 due to AIB constraints and coordination with next-gen consoles. The creator concludes that while AMD won’t necessarily ‘humiliate’ NVIDIA, it could offer a credible alternative, especially if priced aggressively.

193 words

Critical Evaluation

Value of the Information & Strength of the Argument

The video provides a fairly detailed technical analysis of the alleged architecture, including specific numbers for compute units, shader processors, process node, and expected performance. The presenter argues that AMD’s shift to a unified architecture is a strategic move to regain competitiveness, particularly in ray tracing and AI features where NVIDIA has held an advantage. The argumentation is structured around the premise that TSMC’s N3P node and the increased execution density will close the gap with NVIDIA’s newer GPUs. However, the reasoning relies heavily on leaks and unverified estimates, and the presenter occasionally acknowledges uncertainty (e.g., ‘probably’, ‘maybe’). The value of the information lies in its synthesis of many rumors into a coherent picture, but the lack of official confirmation limits its reliability.

Scientific Rigor, Source Quality, Title Accuracy

The video cites a single blog post from the creator’s site as a reference for more details. The information is largely based on leaks and speculation, with no direct quotes from official sources. The title is sensationalist, claiming AMD will ‘humiliate’ the RTX 5090, but the content is more moderate, suggesting only that AMD might compete closely. This overstates the case, though the video does caution that these are estimates. Regarding public comments, the 30 analyzed reveal skepticism about AMD’s ability to truly beat NVIDIA, concerns about driver quality, and a mix of hope and cynicism about the potential price and performance.

240 words

Title / Content Match

The title is sensationalist and claims AMD will 'humiliate' the RTX 5090, but the video itself is more nuanced, presenting it as a potential competitor. The content does discuss the possibility of performance parity at a lower price, but the title overstates certainty.

Quality & Reliability

5/10

The video is based on unconfirmed leaks and speculation, with the presenter acknowledging uncertainty. While the technical details are presented with some expertise, the reliability is limited by the lack of official sources.

Key Moments

Markers derived by PSI from the transcript: the creator did not define chapters.

Cited Sources

Contribution & Novelties

The video’s original contribution is to compile and analyze the latest rumors about AMD’s next-generation high-end GPU, placing them in the context of AMD’s strategic shift to a unified UDNA architecture. It offers a comparative table against the RTX 5090 and discusses practical implications for gamers and AI users. The presentation is straightforward, though it lacks in-depth evidence.

Pour aller plus loin :

  • AMD RDNA microarchitecture — Overview of AMD’s gaming GPU architecture, explaining its evolution towards UDNA.
  • ROCm — AMD’s software stack for computation and AI, which UDNA aims to unify across all products.
  • TSMC — The semiconductor foundry’s process nodes, including N3P, which are critical to the performance and efficiency gains discussed.

114 words

Radar Profile

The radar shows high scores for information quantity and technical level, but lower for reliability and quality, reflecting a video that is dense with speculative but technically detailed content. The overall profile is typical of tech YouTube channels covering leaks and rumors.

Reliability 4/10

💬 Among the 30 comments analyzed, the tone is overall skeptical, with many doubting AMD's ability to truly compete with the RTX 5090, but some express hope for increased competition. Main concerns involve drivers and the reliability of rumors.