friendica (DFRN) - Collegamento all'originale

NVIDIA GPU Price Report, August 2026: the Premium Over MSRP Tracks VRAM, Not Performance

Two findings from the August 2026 NVIDIA GPU price data.

First: memory capacity rather than performance tier is setting the premium over MSRP. Nothing with 8GB or 12GB lists more than 13% above launch price. Nothing with 16GB or more lists less than 18% above it. The band between is empty — and the RTX 5060 Ti 16GB, the cheapest card in the upper group, carries a larger premium than the RTX 5080.

The cleanest evidence comes from the pairs where the same GPU shipped in two memory configurations. On the RTX 5060 Ti, the extra 8GB costs $191 in the market: $23.88 per gigabyte, against $6.25 at launch and roughly $10 for the GDDR7 modules themselves. The RTX 4060 Ti, launched two years earlier, lands at $24.38.

Second: across the RTX 50 line, used cards now list 0–14% below new, three of them at zero. The RTX 40 generation, measured the same day, still lists 11–36% below.

#NVIDIA #GPUPrices #RTX50 #RTX5090 #VRAM #GDDR7 #DRAM #GPU #PCHardware #ITAD #ITAssetDisposition #DataCenter #ResaleValue

NVIDIA GPU Price Update

buysellram.com/blog/nvidia-con…

NVIDIA GPU Price Report, August 2026: the Premium Over MSRP Tracks VRAM, Not Performance

NVIDIA GPU price report, August 2026: memory capacity, not performance tier, sets the premium over MSRP. The same GPU with 16GB costs $191 more. NVIDIA charged $50 for it.
BSR Admin (BuySellRam)
#DataCenter #nvidia #DRAM #gpu #gddr7 #ITAD #pchardware #rtx5090 #resalevalue #rtx50 #Vram #gpuprices #itassetdisposition

friendica (DFRN) - Collegamento all'originale

I Token AI Sono Appena Diventati Ancora Più Economici. Risolveranno la Carenza di Chip per l'AI… o la Alimenteranno?

I token dell'AI diventano sempre più economici, ma la memoria diventa sempre più costosa. I costi dell'inferenza sono diminuiti di 280 volte in soli due anni; nonostante ciò, la spesa delle imprese per l'AI è triplicata e i prezzi contrattuali della DRAM hanno appena registrato un aumento record del 90-95% in un solo trimestre. William Stanley Jevons aveva già osservato questo fenomeno nel 1865 con il carbone. Uno sguardo a ciò che riequilibrerà davvero il mercato dei chip e al motivo per cui la risposta sono le nuove fabbriche di semiconduttori previste per la fine del 2027, non modelli più intelligenti o ASIC.

buysellram.com/blog/cheaper-to…

#AI #Semiconduttori #DRAM #CarenzaDiMemoria #DataCenter #HBM #GPU #SSD #ITAD #BuySellRam #ParadossoDiJevons #Hardware #Tecnologia

#tecnologia #ai #DataCenter #hardware #semiconduttori #DRAM #ssd #gpu #ITAD #hbm #buysellram #CarenzaDiMemoria #ParadossoDiJevons

friendica (DFRN) - Collegamento all'originale

Optical Interconnects Become Critical to AI Factory Expansion

As AI clusters become larger and GPUs become faster, networking and optical communication are becoming just as important as the compute hardware itself. Copper connections consume too much power and suffer from signal loss at very high speeds, so the industry is gradually moving toward optical interconnects. TrendForce forecasts the combined CPO/NPO market could grow from roughly $100 million in 2025 to over $39 billion by 2030, illustrating how strategic this technology is expected to become.

AI isn't just becoming a compute problem anymore—it's becoming a networking problem. Faster GPUs don't help much if data can't move between them efficiently, so it's not surprising to see optical interconnects becoming a strategic technology for future AI data centers.

trendforce.com/presscenter/new…

#AI #AIInfrastructure #DataCenter #GPU #OpticalNetworking #SiliconPhotonics #CPO #HighSpeedNetworking #HPC #Semiconductors #TechNews #FutureOfAI

#ai #DataCenter #semiconductors #hpc #technews #gpu #CPO #futureofai #aiinfrastructure #siliconphotonics #opticalnetworking #highspeednetworking

friendica (DFRN) - Collegamento all'originale

How HBF could make AI inference cheaper

Running AI in production? A lot of the bill is memory. Every token a model generates is served from fast memory next to the GPU — and that memory, HBM, is scarce and expensive.
High Bandwidth Flash (HBF) is built to attack that cost: much of HBM's bandwidth at a fraction of the price per GB. A new explainer covers what HBF is, how it could lower inference costs, and how close it actually is.

How HBF could make AI inference cheaper


#HBF #HighBandwidthFlash #AIinference #HBM #AImemory #GPU #DataCenter #Semiconductors #NAND #SKhynix #SanDisk #MemoryWall

#DataCenter #semiconductors #gpu #hbf #sandisk #skhynix #hbm #nand #AIInference #aimemory #HighBandwidthFlash #MemoryWall
⇧