<div id="readability-page-1"><div id="ArticleBody-InlineImage-108330581"><div><div><p>Jensen Huang, chief executive officer of Nvidia Corp., speaks next to a Vera Rubin Ultra Kyber Compute Tray and a Vera Rubin Ultra Kyber NVLink MidPlane during a keynote address at the Nvidia GTC conference in San Jose, California, US, on Monday, March 16, 2026.</p><p>Bloomberg | Bloomberg | Getty Images</p></div></div></div><div><p>NVIDIA's next marquee product — the Kyber rack-scale architecture designed to house its 2027 Rubin Ultra chips — has been delayed by more than 12 months to 2028, according to research firm SemiAnalysis, the latest in a string of reported setbacks raising questions about the AI giant's product roadmap.</p><p>Kyber is a server cabinet that packs 144 of Nvidia's most powerful chips into a single unit so they can work together as one giant computer, providing the horsepower AI companies need to train and run their most advanced models. </p><p>The design mounts graphics processing units in compute trays that sit vertically instead of horizontally to boost density and reduce latency, and had been slated to debut with Vera Rubin Ultra, Nvidia's next-generation rack-scale system, in 2027.</p><p>The setback stems from difficulties manufacturing a key circuit board at the heart of the system, SemiAnalysis said in a post on Monday. </p><p>&#34;Kyber NVL144 rack architecture has been delayed to 2028 as the PCB midplane remains challenging from a manufacturability standpoint,&#34; the firm said, referring to a specialized, multi-layer printed circuit board that connects electronic modules within a system. </p><p>NVL576 — a larger system linking eight racks via optical connections — is also likely delayed or limited to small volumes, the research firm said. </p><p>Nvidia did not respond to CNBC's request for comment. </p><p>The reported delay adds to mounting strains across Nvidia's product lines, underscoring concerns that Nvidia's breakneck annual release cadence is colliding with manufacturing limits. </p><p>A backup plan — bolting two of Nvidia's current-generation racks together for similar power — has also been scrapped after cloud customers rejected the design as awkward and costly to operate. &#34;It has since been cancelled due to heavy pushback from CSPs [cloud service providers] and hyperscalers over its odd design and heavy operational burden,&#34; SemiAnalysis said. </p><p>That leaves Nvidia with &#34;no proven solution to expand the scale-up world size for Rubin Ultra,&#34; SemiAnalysis said, predicting that could give rivals <span id="RegularArticle-QuoteInBody-2">Advanced Micro Devices<span><span id="-WatchlistDropdown"></span></span></span> and <span id="RegularArticle-QuoteInBody-3">Google<span><span id="-WatchlistDropdown"></span></span></span>, whose in-house chips are already winning business from top AI labs, a rare technical opening at the high end of the market. </p><p>Nvidia's current-generation Rubin systems are in full production and begin shipping this fall to eight cloud partners, including Amazon Web Services, Microsoft Azure and Google Cloud. SemiAnalysis also projects Nvidia's data-center compute revenue will run 20% above Wall Street consensus in the second half of fiscal 2027.</p><p>Shares of Nvidia fluctuated in premarket trading, last down less than 0.1% at $194.79.   </p></div><div><p>Choose CNBC as your preferred source on Google and never miss a moment from the most trusted name in business news.</p></div></div>

NVDA

GOOG

<p>Nvidia's next-gen Kyber rack-scale architecture, designed for 2027 Rubin Ultra chips, is delayed to 2028 due to manufacturing challenges with its PCB midplane. This setback raises concerns about Nvidia's ability to maintain its rapid release cadence amid production limits. A backup plan was scrapped after cloud providers rejected it, leaving a potential opening for rivals AMD and Google in the high-end AI market.</p>

- Nvidia's new Kyber rack-scale architecture, designed for its 2027 Rubin Ultra chips, has been delayed to 2028 due to manufacturing challenges with a key circuit board.  
- This setback raises concerns regarding Nvidia's product roadmap and allows competitors like AMD and Google to potentially gain market share.  
- Despite these issues, Nvidia's current-generation Rubin systems will ship to major cloud partners this fall, and data-center compute revenue is expected to exceed Wall Street forecasts in the latter half of fiscal 2027.

CNBC

NVIDIA

Alphabet - C

Alphabet

AXS 1.5X NVDA Bear Daily ETF

ProShares Ultra NVDA

T-Rex 2X Inverse NVIDIA Daily Target ETF

YieldMax NVDA Option Income Strategy ETF

Direxion Daily NVDA Bull 2X Shares

GraniteShares 2x Long NVDA Daily ETF

XI2CSOPNVDA

XL2CSOPNVDA

T-Rex 2X Long NVIDIA Daily Target ETF

Direxion Daily NVDA Bear 1X ETF

GraniteShares 2x Short NVDA Daily ETF

Leverage Shares 2X Long NVDA Daily ETF

XI2CSOPNVDA-U

Direxion Daily GOOGL Bull 2X Shares

Roundhill AMD WeeklyPay ETF

Leverage Shares 2X Long AMD Daily ETF

Direxion Daily AMD Bull 2X Shares

Direxion Daily Googl Bear 1x Shares

Roundhill GOOGL WeeklyPay ETF

GraniteShares 2x Long GOOGL Daily ETF

GraniteShares 2x Long AMD Daily ETF

Leverage Shares 2x Cpd Acclrtd NVDAMnETF

NVDA WeeklyPay ETF - Roundhill

GraniteShares YieldBOOST NVDA ETF

xETFs NVDA Daily Income ETF

YieldMax Short NVDA Option Inc Strgy ETF

Direxion Daily AMD Bear 1X Shares

Roundhill Daily 2X Lon Mag Sev ETF

Defiance Daily Target 2X Short AMD ETF

YieldMax AMD Option Income Strategy ETF

YieldMax GOOGL Option Income Strategy ETF

T-Rex 2X Long Alphabet Daily Target ETF

Kurv Yield Premium Strategy Google GOOGL ETF

VanEck Semiconductor ETF

iShares Semiconductor ETF

Direxion Semicon Bull 3X

Invesco PHLX Semiconductor ETF

SPDR S&P Semicon

Invesco Semiconductors ETF

First Trust Nasdaq Food & Semicon

Amazon

Microsoft

NVIDIA Corporation

Nvidia's next-gen AI rack system delayed to 2028 on manufacturing snags, SemiAnalysis says