惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

博客园 - 叶小钗
D
Docker
Google DeepMind News
Google DeepMind News
Y
Y Combinator Blog
钛媒体:引领未来商业与生活新知
钛媒体:引领未来商业与生活新知
Blog — PlanetScale
Blog — PlanetScale
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
U
Unit 42
博客园 - 【当耐特】
N
Netflix TechBlog - Medium
V
Visual Studio Blog
Microsoft Azure Blog
Microsoft Azure Blog
博客园_首页
Recent Announcements
Recent Announcements
GbyAI
GbyAI
T
Tailwind CSS Blog
S
SegmentFault 最新的问题
WordPress大学
WordPress大学
T
The Blog of Author Tim Ferriss
Engineering at Meta
Engineering at Meta
L
LangChain Blog
A
About on SuperTechFans
M
MIT News - Artificial intelligence
B
Blog

Latest from Tom's Hardware in Tech-industry

Samsung engineer sentenced to 7 years in prison for selling chipmaking trade secrets to Chinese chipmaker — ex-employee supplied 10nm DRAM data to CXMT for $2 million Microsoft facing $2.8 billion UK lawsuit for overcharging 60,000 businesses using Microsoft Server on other clouds — Azure users allegedly received lower wholesale pricing Ransomware negotiator pleads guilty after leaking victims' insurance details to 'BlackCat' hackers — perp gave attackers a precise picture of exactly how much each target could afford to pay Congress moves to strip the DoC of chip-export discretion with the MATCH Act — DUV lithography machines among those targeted in chipmaking tool crackdown Anthropic's Model Context Protocol includes a critical remote code execution vulnerability — newly discovered exploit puts 200,000 AI servers at risk Iran claims US exploited networking equipment backdoors during strikes — says devices from Cisco and others failed despite blackout in attack that 'indicates deep sabotage' Cerebras files for IPO — company remains unprofitable despite 20x revenue growth Scientists solve decades-old 2D physics puzzle — Chaotic growth in a 2D quantum system obeys statistical laws Anthropic nuked a company's access to Claude, stopping 60 employees dead in their tracks — support via Google Form is the only recourse for vague usage policy violation Biohacker claims to have sequenced their own genome at the kitchen table with M3 Ultra Mac Studio, Claude, and a $3,200 sequencer — DIY project requires 100GB of data storage per run, oodles of RAM New Jersey datacenter expansion got $77m in tax breaks to create exactly one permanent job — JPMorgan's site already scored $35m and currently employs just 25 workers Chinese chipmakers made record profit in 2025, despite slipping margins — U.S shipments fall 34% as Beijing shores up local chipmaking efforts AI cloud company Vercel breached after employee grants AI tool unrestricted access to Google Workspace — hacker… Every SK hynix employee could receive $477,000 bonuses this year, almost $900,000 next year — 35,000 workers reportedly set to benefit from share of $169 billion projected operating profit US gov't blocks China's largest LED chipmaker's $239 million bid to acquire Dutch lighting firm Lumileds… Tokyo court rules movie and anime 'spoiler articles' are copyright infringement in landmark criminal case — detailed, monetized plot summaries land man in Japanese prison Voyager 1 gets emergency instrument shutdown to solve escalating power crisis and give it ‘about a year of breathing room’ — interstellar spacecraft's nuclear power source is dying, leading to intensifying countermeasures Russian-made Shahed drones are ‘disintegrating in the air before reaching their targets’ due to shoddy manufacturing, video shows — commentators call Russian clones of Iran's drones 'flying garbage' Inventor showcases 3D printer filament dryer that mines Bitcoins and dries filament with waste heat, capable of 6 TH/s at 140W — joins Bitcoin-mining 3D printer in hobbyist-focused miner lineup Bluetooth tracker hidden in a postcard and mailed to a warship exposed its location — $5 gadget put a $585 million… Techie buys fake Ledger Nano S+ hardware crypto wallet and almost falls for phishing — a convincing clone would… Nvidia CEO Jensen Huang ‘nearly lost his composure’ when pressed on selling chips to China — ‘You’re not talking to someone who woke up a loser’ US lawmakers amend new restrictions on Chinese chipmakers — MATCH Act Analytics group signals possible delays at 40% of AI data center construction sites — companies deny schedule holdups, but satellite imagery indicates otherwise Local political revolts threaten to derail US data center projects — mounting delays are already costing AI hyperscalers billions Intel hires tenured Samsung exec to lead Foundry Services — signals company focus on winning business from potential Foundry suitors Elon Musk pushing forward with Terafab at TSMC ups revenue guidance and CapEx, buoyed by Google and Pentagon in talks to run custom AI chips inside classified environments — Google pushes for tight controls for TPUs surrounding use for mass surveillance and autonomous weapons TSMC warns of Intel Foundry
Nvidia's memory costs soar 485%, latest AI systems now co...
Anton Shilov · 2026-05-21 · via Latest from Tom's Hardware in Tech-industry
Nvidia
(Image credit: Nvidia)

As prices of components are increasing rapidly due to high demand from the AI sector, the cost of these machines is also increasing significantly. Morgan Stanley Research estimates that a next-generation Vera Rubin-based VR200 NVL72 rack will cost major hyperscale cloud service providers (CSPs) around $7.8 million per unit (via @Aaronwei3n), which is tangibly more than about $4 million per GB300 NVL72. Furthermore, because every VR200 NVL72 rack packs plenty of DRAM and NAND, memory now accounts for around 25% of the total cost.

Nvidia plans to charge $55,000 per Rubin GPU and $5,000 per Vera CPU when selling them in volume inside VR200 NVL72 chassis to hyperscalers, according to Morgan Stanley. Although the upcoming VR200 NVL72 racks use the already familiar Oberon chassis, they use more sophisticated switching, networking, printed circuit board (PCB), cooling, power supply, and even chip packaging components, which increases bill-of-material (BOM) costs and eventually the price of the systems. As a result, each VR200 NVL72 will cost hyperscalers around $7.8 million, according to Morgan Stanley, which is higher than around $7 million we were told by one of our sources in late March. Meanwhile, the cost of memory within a VR200 NVL72 rack will be about $2 million, up 435% from the memory cost in GB300 NVL72, according to the same figures.

Sheesh.$NVDA VR200 Bom Analysis from MS. pic.twitter.com/sutjttSkyWMay 21, 2026

There are several reasons why the cost of memory is expected to account for 25% of the cost of a VR200 NVL72 system and why the system carries $2 million worth of memory.

First up, each of such racks now contains 54 TB of LPDDR5X memory, up from 17 TB of LPDDR5X in the case of a GB200 NVL72, a threefold increase. SemiAnalysis estimates that Nvidia paid $8 per GB per GB of LPDDR5X in Q1, though that price may increase as demand rises in the coming quarters, especially if we are talking about SOCAMM2 modules that are expensive to make and test. In any case, even at $8 per GB, each GB200 NVL72 machine carries $136,000 worth of LPDDR5X memory, whereas each VR200 NVL72 system will contain $408,000 worth of LPDDR5X content. If the price rises to $10, we are talking about $540,000 for LPDDR5X alone. Note that even $10 per GB may be an underestimate* as Nvidia adds its own markup.

Secondly, each VR200 NVL72 rack carries about $1 million or more of 3D NAND storage, up from virtually zero inside GB200 NVL72.

As a result, $2 million of memory content per Vera Rubin NVL72 rack is not something completely unexpected: the system uses a lot of LPDDR5X and 3D NAND memory (not to mention HBM4 memory onboard of Rubin GPUs), and memory now comes at massive prices.

*Contract price of DDR5 memory is now between $12 and $16 per GB, depending on various factors and luck, according to Framework. Spot price for DDR5 was about $20 per GB on average at press time, according to DRAMeXchange. LPDDR5X is more expensive than DDR5. When installed on SOCAMM2 modules (which are exclusively used by Nvidia's Vera CPUs), it will get even more costly, especially when Nvidia's markup is added.

Get Tom's Hardware's best news and in-depth reviews, straight to your inbox.

Google Preferred Source

Follow Tom's Hardware on Google News, or add us as a preferred source, to get our latest news, analysis, & reviews in your feeds.

Anton Shilov is a contributing writer at Tom’s Hardware. Over the past couple of decades, he has covered everything from CPUs and GPUs to supercomputers and from modern process technologies and latest fab tools to high-tech industry trends.