THE APEX TIMES
Nvidia Kyber NVL144 rack delivery pushed into 2028 as manufacturing issue hits key circuit board, SemiAnalysis says
A manufacturing problem tied to a key circuit-board component is delaying shipment of Nvidia’s Kyber NVL144 rack by more than 12 months, according to SemiAnalysis, raising new questions about near-term supply tightness in AI data center hardware and the knock-on impact for Asian component suppliers.
Nvidia’s Kyber NVL144 AI rack is facing a further delivery delay, with a research firm saying the system has been pushed back to 2028 after a manufacturing issue tied to a key circuit board.
The report, carried by Quartz, attributed the setback to problems producing a critical circuit-board component used in the rack’s configuration. SemiAnalysis said the delay amounts to more than 12 months, a timeline shift that could compress how quickly finished systems move from factory lines into customer deployments.
The Kyber NVL144 is a “rack” product, meaning a standardized, pre-integrated cabinet of compute and supporting hardware designed to be installed into data centers as a unit rather than assembling the infrastructure from individual parts on site. For AI builders, rack-level delivery timing can matter because it affects when data center capacity can be commissioned and when customers can scale training and inference workloads.
Quartz also said the delay is weighing on supplier stocks, pointing to broader market sensitivity to disruptions in AI hardware supply chains. When a component bottleneck emerges, investors often reprice expectations for revenue timing among makers of electronics and subassemblies that feed the final system integrators.
Nvidia, for its part, did not disclose details of the reported Kyber-specific manufacturing issue in the Quartz account, and the report did not provide names of affected suppliers or quantify the magnitude of the stock moves.
For the broader technology sector, the episode underscores how AI infrastructure is still dependent on complex manufacturing yield and component availability, even as demand for accelerated compute remains strong. Racks and other system-level products can be held up by a single constrained component, especially where the design requires specialized boards and tight quality control.
What remains unclear from the published report is whether the delay reflects a temporary yield shortfall, a redesign, or a sourcing substitution for the affected circuit board. The Quartz write-up also did not spell out whether the change impacts all customer configurations of the NVL144 or only specific production runs.
Investors and customers will likely watch for follow-up guidance from Nvidia and its manufacturing partners on revised delivery schedules, any changes to bill of materials for the circuit board, and whether supplier-level constraints begin to ease in coming quarters.
Why It Matters
- A rack-level delay can affect when customers can deploy AI compute capacity, since data centers often commission infrastructure in standardized units.
- Component bottlenecks, such as circuit-board manufacturing constraints, can propagate through the supply chain and change revenue timing expectations across multiple tiers of suppliers.
- Market reaction in supplier stocks suggests investors are tracking manufacturing readiness closely for AI system shipments.
- If the bottleneck is not resolved quickly, system integrators may need to adjust build plans, potentially reshaping order pacing into later quarters.
Key Facts
- A Quartz report citing SemiAnalysis says Nvidia’s Kyber NVL144 rack is being delayed to 2028.
- SemiAnalysis attributed the delay to a manufacturing problem involving a key circuit-board component.
- The delay was described as more than 12 months.
- Quartz said the report is contributing to declines in supplier stocks tied to the supply chain.
- The Quartz account did not name specific affected suppliers or provide detailed delivery figures.
Technology Related
Meta and Tesla earnings refocus investors on the AI trade, with index futures watching momentum
As Meta Platforms and Tesla reported results, traders weighed how much of the market’s artificial-intelligence enthusiasm was already priced in, and whether earnings would sustain the broader risk-on tone reflected in U.S. index futures.
Meta Faces $25 Billion New Mexico Privacy Verdict Over Alleged Facebook Data Violations
A New Mexico jury ordered Meta Platforms to pay $25 billion in damages tied to alleged Facebook privacy violations connected to the Cambridge Analytica era.
Bill Ackman Says Netflix Rejected Ads, Then Rolled Out an Ad Model Weeks Later
The Pershing Square founder described being blindsided by Netflix’s messaging on advertising and said the change in view cost him about $400 million.
Meta investors focus on AI efficiency as engagement grows, after a review highlights margin momentum
A recent takeaways piece points to lower unit costs from AI-powered content processing and a scaling effect across Meta’s social platforms, a mix market participants often treat as a lever for sustained profitability.
TIER IV and Technical University of Munich expand work with NVIDIA Hyperion toward Level 4 autonomous driving in Germany
The robotics and autonomous-driving software firm TIER IV and the Technical University of Munich said they are expanding a collaboration focused on NVIDIA’s Hyperion platform, targeting Level 4 driving capabilities in Germany.
Amazon’s cash engine looks stronger than Target’s, but valuation multiples complicate the comparison
A market analysis comparing Amazon and Target argues that Amazon generates substantially more operating cash flow, while differences in how investors value each company may be the deciding factor for long-term holders.
Netflix shares fall after Q3 results reignite valuation worries
A post-earnings drop highlights how quickly investor expectations can shift for Netflix, even as the streaming leader has continued to emphasize profitability.
Netflix sets premiere date for new Korean series Tantara, starring Song Hye-kyo and Gong Yoo
The streamer says Tantara will debut December 11 exclusively on Netflix, pairing marquee Korean talent with writer Noh Hee-gyoung and director Lee Yoon-jung.
Intel joins Elon Musk’s proposed Terafab AI chip effort, raising questions about how a foundry vendor would fit into Musk’s AI hardware stack
Intel said it is partnering with Tesla and SpaceX on Elon Musk’s newly announced Terafab AI chip project, according to a report cited by Yahoo Finance. The companies have not, in the reporting, laid out technical specifications, timelines, or production plans.
Meta Says Muse Surpassed 10 Million Users, With Plans Aimed at Reaching 1 Billion People
CEO Mark Zuckerberg told investors that Muse adoption has passed 10 million users and that the company is positioning its AI efforts for a much larger audience over the next few years.