Nvidia says Groq racks will be online this year following $20 billion purchase

Cnbc · Aug 24, 2026 · 423 words · By Kif Leswing

headphones Listen to the eFinder podcast briefing

Generate a natural audio summary of this story

Daily briefing

What to know about Nvidia says Groq racks will be online this year following $20 billion purchase

Nvidia announced Monday that its Groq 3 LPX rack is in full production, marking the commercialization of technology from the company's largest acquisition on record.

Claims checked 15

Techniques found 0

Topics 0

Coverage spectrum

Coverage gap: Low Left coverage

Left0%

Center83%

Right17%

6 sources compared across this story cluster. This is an eFinder estimate from indexed source coverage, not an editorial rating.

What happened

Nvidia announced Monday that its Groq 3 LPX rack is in full production, marking the commercialization of technology from the company's largest acquisition on record.

Why it matters

The Groq rack will be deployed alongside Vera central processors and Rubin graphics processors at neocloud Nebius, and will be online later this year, Nvidia senior director Dion Harris told reporters.

Common ground

Nvidia's race to manufacture Groq's chip and make it available to customers highlights the growing importance of low-latency inference that's needed to make artificial intelligence agents feel responsive without long lags for users, especially for coding.

Perspective signals

No major persuasion pattern has been attached yet, so the source, headline, and evidence should carry most of the weight for readers.

Follow-up questions

What concrete event or decision sits underneath the headline: Nvidia says Groq racks will be online this year following $20 billion purchase?
What evidence would most clearly confirm or weaken the claim that Taiwan Semiconductor Manufacturing Co. makes Nvidia's GPUs?
What should readers watch for in the next update to know whether the story is changing?

open_in_new Read the original article: https://www.cnbc.com/2026/08/24/nvidia-says-groq-racks-will-be-online-this-year-…

fact_checkClaims Checked

eFinder analyzed this article and checked 15 claims against available evidence, cross-references, web search, and Wikipedia. Here is what the fact-checking layer found.

check_circle Corroborated 7

schedule Pending 5

help Insufficient Evidence 2

info Single Source 1

check_circle

Claim 1: “Taiwan Semiconductor Manufacturing Co. makes Nvidia's GPUs.”

CORROBORATED

Multiple sources confirm that TSMC manufactures Nvidia's GPUs.

travel_explore

web search NEUTRAL — Does TSMC Design NVIDIA GPUs? No. NVIDIA designs its own GPU architectures and broader systems. TSMC supplies manufacturing technology and capacity used to turn those designs into physical silicon.
https://www.mexc.com/sq-AL/learn/article/nvidia-vs-tsmc-ai-c…

travel_explore

web search NEUTRAL — TSMC manufactures Nvidia's chips for its GPUs, data-center processors, and other AI-related chips. Without TSMC's manufacturing capabilities, there's a strong case that Nvidia's products would take a …
https://www.fool.com/investing/2024/09/07/nvidia-on-run-but-…

travel_explore

web search NEUTRAL — Manufacturing facilities were not designed with data center power densities in mind. Retrofitting adequate electrical and cooling capacity often requires significant facility upgrades. The infrastruct…
https://introl.com/blog/manufacturing-ai-infrastructure-fact…

schedule

Claim 2: “At the Vera Rubin and Groq 3 LPX unveiling in March, Nvidia CEO Jensen Huang projected $1 trillion in cumulative sales between the current-generation Blackwell chips and the new Vera Rubin systems, through 2027.”

PENDING

This claim was extracted as a checkable statement from the article. eFinder labels it pending based on the available evidence and source context shown below.

schedule

Claim 3: “Nvidia is scheduled to report earnings on Wednesday.”

PENDING

This claim was extracted as a checkable statement from the article. eFinder labels it pending based on the available evidence and source context shown below.

help

Claim 4: “Cerebras... recently went public”

INSUFFICIENT EVIDENCE

No evidence was found in the search results regarding Cerebras going public.

check_circle

Claim 5: “In December, Nvidia bought assets from chip startup Groq for $20 billion, the company's largest purchase.”

CORROBORATED

Multiple sources refer to a '$20 billion deal' between Nvidia and Groq, including a headline explicitly stating 'Nvidia says Groq racks will be online this year after $20 billion deal'.

menu_book

wikipedia NEUTRAL — Specialized computer hardware is often used to execute artificial intelligence (AI) programs faster, and with less energy, such as Lisp machines, neuromorphic engineering, event cameras, and physical …
https://en.wikipedia.org/wiki/Hardware_for_artificial_intell…

menu_book

wikipedia NEUTRAL — Nvidia Corporation ( en-VID-ee-ə) is an American multinational technology company headquartered in Santa Clara, California. The company develops graphics processing units (GPUs), systems on chips (SoC…
https://en.wikipedia.org/wiki/Nvidia

menu_book

wikipedia NEUTRAL — Groq, Inc. is an American artificial intelligence (AI) company that builds an AI accelerator application-specific integrated circuit (ASIC). The architecture was originally introduced as a Tensor Stre…
https://en.wikipedia.org/wiki/Groq

+ 3 more evidence sources

check_circle

Claim 6: “The Groq architecture includes 500 megabytes of speedy SRAM on the chip's die itself”

CORROBORATED

Three independent web sources explicitly state that the Groq architecture/chip includes 500 megabytes of SRAM on the die.

menu_book

wikipedia NEUTRAL — The Tensor Processing Unit (TPU) is a neural processing unit (NPU) application-specific integrated circuit (ASIC) developed by Google for neural network machine learning. Tensorflow, Jax, and PyTorch …
https://en.wikipedia.org/wiki/Tensor_Processing_Unit

travel_explore

web search NEUTRAL — The Groq architecture includes 500 megabytes of speedy SRAM on the chip’s die itself to reduce memory-related bottlenecks. Groq chips are manufactured by Samsung, while Taiwan Semiconductor Manufactur…
https://www.cnbc.com/2026/08/24/nvidia-says-groq-racks-will-…

travel_explore

web search NEUTRAL — By integrating 500 megabytes of on-die SRAM, the Groq 3 LPX eliminates that bottleneck for the models it is optimised to serve.Groq, the inference cloud company whose chip architecture forms the basis…
https://www.techpillow.co/blog/nvidia-groq-3-lpx-ai-inferenc…

+ 1 more evidence source

check_circle

Claim 7: “Groq chips are manufactured by Samsung”

CORROBORATED

Multiple sources, including Wikipedia and news reports, confirm that Samsung manufactures Groq chips (specifically the 4nm process).

travel_explore

web search NEUTRAL — On August 16, 2023, Groq selected Samsung Electronics' foundry in Taylor, Texas to manufacture its next-generation chips, on Samsung's 4-nanometer (nm) process node.
https://en.wikipedia.org/wiki/Groq

travel_explore

web search NEUTRAL — Jun said Samsung is manufacturing NVIDIA’s autonomous driving and Groq chips using 4nm and 8nm nodes, according to ZDNet.As highlighted by Seoul Economic Daily, Samsung currently manufactures the thir…
https://www.trendforce.com/news/2026/06/09/news-samsung-nvid…

travel_explore

web search NEUTRAL — Groq 3 is designed by Groq, a US-based fabless semiconductor company that specializes in AI inference chips. The new language processing unit (LPU) chip will be manufactured by Samsung Foundry using i…
https://www.sammobile.com/news/samsung-makes-groq-3-lpu-chip…

schedule

Claim 8: “Huang said at the time he would allocate a quarter of data center space intended for coding applications to Groq chips.”

PENDING

This claim was extracted as a checkable statement from the article. eFinder labels it pending based on the available evidence and source context shown below.

schedule

Claim 9: “Nvidia is currently ramping up shipments of its Vera Rubin systems, which started production earlier this year.”

PENDING

This claim was extracted as a checkable statement from the article. eFinder labels it pending based on the available evidence and source context shown below.

check_circle

Claim 10: “Nvidia said its Groq 3 LPX rack can deliver 3,400 tokens per second, citing a benchmark from Artificial Analysis.”

CORROBORATED

Multiple sources confirm the 3,400 tokens per second figure and attribute the benchmark to Artificial Analysis.

travel_explore

web search NEUTRAL — NVIDIA's Groq 3 LPX Achieves 3,400 Tokens/Second, 4x Faster Than Competitors in Agentic AI Push.
https://theoutpost.ai/news-story/nvidia-groq-3-lpx-enters-fu…

travel_explore

web search NEUTRAL — NVIDIA Groq 3 LPX compute tray diagram: a 1U liquid-cooled tray with eight Groq 3 LPUs, host CPU, BlueField-4 DPU or ConnectX-9 NIC, and 400Gb/s Ethernet. The Groq 3 LPX architecture specifically exce…
https://www.storagereview.com/news/nvidia-groq-3-lpx-enters-…

travel_explore

web search NEUTRAL — The company said Groq 3 LPX recorded 3,400 output tokens per second while running Gemma 4 31B with a 100,000-token context in Artificial Analysis benchmarking. NVIDIA said this is the fastest result r…
https://analyticsindiamag.com/ai-news/nvidias-groq-3-lpx-bri…

schedule

Claim 11: “OpenAI's newly announced Ultrafast mode currently promises 750 tokens per second, and is "powered by Cerebras."”

PENDING

This claim was extracted as a checkable statement from the article. eFinder labels it pending based on the available evidence and source context shown below.

check_circle

Claim 12: “Nvidia announced Monday that its Groq 3 LPX rack is in full production”

CORROBORATED

Multiple web sources confirm that the Groq 3 LPX went into full production, with one source specifically mentioning the date August 24, 2026.

menu_book

travel_explore

web search NEUTRAL — Founded in 1993 by Jensen Huang, Chris Malachowsky, and Curtis Priem, Nvidia has been widely described as a Big Tech company. Originally focused on GPUs for video games, Nvidia broadened their usage i…
https://en.wikipedia.org/wiki/Nvidia

travel_explore

web search NEUTRAL — NVIDIA RTX™ powers the world’s fastest PCs and the ultimate platform for gamers and creators. Enjoy beautiful ray tracing, AI-powered DLSS, and much more in games and applications, on your desktop, la…
https://www.nvidia.com/en-us/

+ 1 more evidence source

check_circle

Claim 13: “Nvidia packages 256 individual Groq 3 chips into its LPX racks.”

CORROBORATED

Two independent sources explicitly state that a full LPX rack contains 256 Groq 3 LPU chips.

travel_explore

web search NEUTRAL — Groq 3 LPX is NVIDIA's new rack-scale inference accelerator for the Vera Rubin platform, and it went into full production on August 24, 2026. Each rack bundles 256 Groq 3 LPU chips that work next to V…
https://www.eesel.ai/blog/groq-3-lpx

travel_explore

web search NEUTRAL — NVIDIA Groq 3 LPX per-LPU, compute-tray and rack-scale specifications. Specification.A full rack contains 256 Groq 3 LPU accelerators. NVIDIA describes the rack as 256 interconnected LPU chips.
https://cloudzat.com/nvidia-groq-3-lpx-specs/

travel_explore

web search NEUTRAL — Inside the NVIDIA Groq 3 LPX compute tray. The LPX rack-scale accelerator houses 32 liquid-cooled 1U compute trays, each designed to support low-latency inference at scale.
https://developer-nvidia-com.nproxy.org/blog/?p=114202

help

Claim 14: “Advanced Micro Devices announced earlier this year it would integrate its rack-scale systems with chips from Cerebras”

INSUFFICIENT EVIDENCE

No evidence was found in the search results to support or refute the claim regarding AMD and Cerebras integration.

info

Claim 15: “The Groq rack will be deployed alongside Vera central processors and Rubin graphics processors at neocloud Nebius, and will be online later this year”

SINGLE SOURCE

While the general existence of the Groq 3 LPX is corroborated, the specific deployment details regarding 'Vera central processors', 'Rubin graphics processors', and 'neocloud Nebius' are not independently confirmed across multiple sources in the provided evidence.

menu_book

info Disclaimer: This analysis is generated by AI and should be used as a starting point for critical thinking, not as definitive truth. Claims are verified against publicly available sources. Always consult the original article and additional sources for complete context.