Free RTX 5080 After Cancellation, GPU Swap Scam, NVIDIA's 22.2 TB/s Bandwidth Upgrade

A Reddit user received a free RTX 5080 after cancelling an Amazon order due to US consumer protection laws, while another buyer fell victim to an RTX 5060 Ti swapped as an RTX 5080, and NVIDIA raised its Vera Rubin memory bandwidth to 22.2 TB/s to surpass AMD's MI455X.

macrozheng
macrozheng
macrozheng
Free RTX 5080 After Cancellation, GPU Swap Scam, NVIDIA's 22.2 TB/s Bandwidth Upgrade

Amazon Sends Free RTX 5080 After Order Cancellation

Reddit user spaceman329 ordered an ASUS ROG RTX 5080 Night God Edition for $1,850 (≈¥12,878) on Amazon, then cancelled the order after finding stock at a local Micro Center. Despite the cancellation and refund, the GPU was delivered to his door. When he contacted Amazon customer service offering to pay, they verified the situation and told him to keep the card at no cost.

Reddit user spaceman329's post about free RTX 5080
Reddit user spaceman329's post about free RTX 5080

This decision stems from US consumer protection laws : unsolicited merchandise delivered to a recipient cannot be charged for, and the recipient may treat it as a gift. Violators face heavy penalties, so Amazon opted to let the customer keep the GPU.

GPU Swap Scam: RTX 5060 Ti Disguised as RTX 5080

Another buyer purchased a supposedly new ASUS Prime GeForce RTX 5080 OC 16GB on Amazon. Upon arrival, the card had only a single 8‑pin power connector, whereas the genuine Prime RTX 5080 requires the new 16‑pin (12V‑2x6) connector and typically includes an 8‑pin adapter.

RTX 5080 with only single 8-pin power connector
RTX 5080 with only single 8-pin power connector

Videocardz cross‑referenced its database and confirmed that the cooler design matches the Prime series, but the PCB layout and power specifications do not match the RTX 5080. Detailed comparison reveals the card is actually an RTX 5060 Ti Prime with the original label removed and a counterfeit RTX 5080 label applied.

Comparison showing RTX 5060 Ti PCB vs RTX 5080 specifications
Comparison showing RTX 5060 Ti PCB vs RTX 5080 specifications

The scam exploits a gap in Amazon's return inspection: a fraudster buys a genuine RTX 5080, swaps in a cheaper RTX 5060 Ti, returns the package, and Amazon restocks it as "new" without verifying the hardware. The victim noticed lifted and re‑applied label edges — clear evidence of tampering.

Counterfeit label with lifted edges on swapped GPU
Counterfeit label with lifted edges on swapped GPU

Advice: Always record a complete unboxing video when purchasing high‑value hardware to protect against such fraud.

NVIDIA Boosts Vera Rubin Memory Bandwidth to Counter AMD MI455X

According to SemiAnalysis , at CES 2026 NVIDIA raised the memory bandwidth specification of its Vera Rubin NVL72 platform to 22.2 TB/s , surpassing AMD's Instinct MI455X (19.6 TB/s) by about 10%.

NVIDIA Vera Rubin memory bandwidth progression chart
NVIDIA Vera Rubin memory bandwidth progression chart

The bandwidth targets have escalated rapidly: NVIDIA initially announced 13 TB/s in March 2025, then 20.5 TB/s in September 2025, and now 22.2 TB/s in January 2026. This reflects the critical role of memory bandwidth as generative AI systems become dominant in 2026.

AMD's MI455X achieves 19.6 TB/s using 12‑Hi HBM4 stacking . To leapfrog, NVIDIA is pushing an aggressive approach: 8‑layer HBM4 stacks with pin speeds of 11 Gbps , well above the JEDEC standard rating. This overclocking strategy aims to secure NVIDIA's absolute dominance among hyperscale cloud providers.

AMD MI455X vs NVIDIA Vera Rubin memory bandwidth comparison
AMD MI455X vs NVIDIA Vera Rubin memory bandwidth comparison
Original Source

Signed-in readers can open the original source through BestHub's protected redirect.

Sign in to view source
Republication Notice

This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactadmin@besthub.devand we will review it promptly.

NVIDIAAMDAmazonmemory bandwidthVera RubinGPU scamHBM4RTX 5080
macrozheng
Written by

macrozheng

Dedicated to Java tech sharing and dissecting top open-source projects. Topics include Spring Boot, Spring Cloud, Docker, Kubernetes and more. Author’s GitHub project “mall” has 50K+ stars.

0 followers
Reader feedback

How this landed with the community

Sign in to like

Rate this article

Was this worth your time?

Sign in to rate
Discussion

0 Comments

Thoughtful readers leave field notes, pushback, and hard-won operational detail here.