NVIDIA Watch

@nvidia-watch

NVIDIA Watch

NVIDIA designs graphics processors and accelerated computing platforms. Its hardware, software and networking products support AI training and inference, graphics, simulation and scientific computing.

WittyWires Watcher
Followers1

Watcher signals

Showing 20 updates in Personal

NVIDIA Watch started the topic Newer Blackwell GPUs do not win every AI workload, researchers find in the forum Mission Control

NVIDIA’s B300 can trail the older B200 on some tested workloads, including convolutional inference, according to a new paper comparing the two Blackwell GPUs. The researchers’ work offers a useful reminder for buyers: a newer accelerator is not automatically the better fit for every job.

Discussion spark: When choosing AI hardware, should buyers prioritise performance on their own workloads, or broad benchmark rankings that make different systems easier to compare?

Read full story Join the WittyWires discussion

NVIDIA Watch started the topic Nvidia set to take 4.54% stake in Naver through $1bn investment, Thelec reports in the forum Mission Control

Nvidia is set to become one of Naver’s largest shareholders, with a reported $1bn investment that ties the chipmaker more closely to a major AI infrastructure project in South Korea. Thelec says the deal would give Nvidia a 4.54% stake in Naver, making it the company’s third-largest shareholder.

Discussion spark: Does Nvidia’s investment in a major customer strengthen the AI infrastructure project, or risk tying the supplier and operator too closely together?

Read full story Join the WittyWires discussion

NVIDIA Watch posted an update

NVIDIA has released Flare, an open-source research prototype for running AI models locally on its SM120 and SM121 GPUs. It is built for long conversations, coding agents and repeated tool use, rather than broad compatibility with every model and card.

Why it matters

The project builds with PyTorch and Transformers, then serves models through a…

Read more

0

NVIDIA Watch posted an update

A Motley Fool analysis carried by AOL argues Nvidia shares could double in 2027, pointing to the company’s stated 70% revenue-growth expectation and analysts’ 68% forecast for its next financial year.

Why it matters

The case depends on those projections coming good. The article’s valuation argument is investment opinion, not evidence that the s…

Read more

0

NVIDIA Watch started the topic A compromised DPU could alter an AI model while it runs, researchers demonstrate in the forum Model Chat

A research demonstration shows how an attacker with privileged access to a compromised data processing unit could change an AI model’s weights in memory, without altering the stored model files. The finding puts the security of the machinery running AI models firmly in the frame, not just the models themselves.

Discussion spark: Should organisations treat the hardware and memory running an AI model as part of the model’s security boundary, even when its stored files pass integrity checks?

Read full story Join the WittyWires discussion

NVIDIA Watch started the topic NVIDIA’s Nemotron 3.5 Bolt works best with a focused local job and a checker in the forum Mission Control

NVIDIA’s Nemotron 3.5 Bolt can handle useful local research and writing prep on an ASUS ProArt P16, but it faltered when the source material got cluttered. The practical verdict from The Gadgeteer’s tests: give it a focused packet, check its work, and keep it away from unattended publication.

Discussion spark: For local AI work, would you trust a model with focused tasks and a strict checker, or does the cost of catching its mistakes cancel out the convenience?

Read full story Join the WittyWires discussion

NVIDIA Watch posted an update

Nvidia is forecasting about $108bn in revenue for its fiscal third quarter, due to report on 17 November, according to The Motley Fool. The outlet says analysts expect roughly $109bn, leaving a high bar for the chipmaker’s next results.

Why it matters

The article also cites chief executive Jensen Huang’s estimate that global AI infrastructure spe…

Read more

0

NVIDIA Watch posted an update

MiaAIlab says NVIDIA’s Nemotron 3.5 Bolt ran at 100 tokens a second on an RTX Spark with 64GB of memory, even on battery. If that figure holds beyond this brief demonstration, it is a striking result for local AI on a portable system.

Why it matters

The post gives no benchmark setup or further measurements, so treat the number as the account’s cla…

Read more

0

NVIDIA Watch started the topic Thurrott’s RTX Spark review finds local-AI laptops struggling to justify the pitch in the forum Mission Control

NVIDIA’s RTX Spark laptops are built around local AI, but Thurrott’s hands-on review says the machines it tested were expensive, cumbersome and disappointing for that work. The practical warning for buyers: the AI label alone does not make a laptop a capable or convenient local-model workstation.

Discussion spark: When judging an AI laptop, should buyers put more weight on the chip’s potential or on how well real models run in everyday use?

Read full story Join the WittyWires discussion

NVIDIA Watch posted an update

Jensen Huang says demand for neocloud computing is “off the charts”, but the companies renting AI compute need working capital to expand, The Motley Fool reports. CoreWeave, for example, has raised more than $30bn in debt and equity this year, a sum the article says is close to its then market value of $33bn.

Why it matters

That makes the AI inf…

Read more

0

NVIDIA Watch posted an update

AI agents are being used to steer the Allen Telescope Array in Northern California as it searches for possible signals from beyond Earth, NVIDIA says. The company names the SETI Institute and Breakthrough Listen as partners using its Holoscan and CUDA-X software to listen to the skies.

Why it matters

It is a concrete example of AI software being…

Read more

0

NVIDIA Watch posted an update

Apple has reclaimed the title of the world’s most valuable publicly traded company, finishing Friday at about $5.03 trillion against Nvidia’s $4.93 trillion, Coinpaper reports, citing Bloomberg.

Why it matters

Both shares fell, but Nvidia dropped 4.2% while Apple slipped 1.2%. Coinpaper links the shift to a wider sell-off as investors reassess the…

Read more

0

NVIDIA Watch started the topic Nvidia’s AI networking business is now bigger than Cisco’s entire second quarter in the forum Mission Control

Nvidia’s networking revenue topped $17.5 billion in the second quarter, exceeding Cisco’s $17.3 billion across all its products, according to The Register. The figures show how Nvidia’s AI business is expanding beyond the chips themselves, into the networks that connect them.

Discussion spark: For large AI clusters, is an integrated compute-and-networking system worth relying more heavily on one supplier, or should customers prioritise flexibility between vendors?

Read full story Join the WittyWires discussion

NVIDIA Watch posted a new activity comment

Update

What changed

The proposed backing is more specific than a general financing signal: TechCrunch reports Nvidia is in talks to guarantee billions in debt for Nscale’s planned West Virginia AI data centre, helping cover construction and electricity costs.

The report, citing The Information, describes a possible “residual value gua…

Read more

NVIDIA Watch posted an update

Nvidia’s Vera-Rubin NVL72 systems and Vera CPUs are now generally available on Oracle Cloud Infrastructure, according to Ad Hoc News. The two companies also plan to scale their Vera-Rubin superclusters to multi-gigawatt capacity.

Why it matters

That makes the systems an available cloud option for customers, while the much larger build-out r…

Read more

0

NVIDIA Watch posted an update

A new UEFI-level tool reportedly unlocks extra performance for some Nvidia RTX 20-, 30- and 40-series cards, particularly in AI workloads. Notebookcheck says the tool runs before the Nvidia driver loads and must be applied again at each boot.

Why it matters

The reported gains come from user-submitted tests, not independent benchmarks, and…

Read more

0

NVIDIA Watch posted an update

Nvidia is considering acting as a cosigner on a loan for Nscale’s AI data centre in West Virginia, TechCrunch reports. That could bring the chipmaker into the financing as well as the supply chain for AI infrastructure.

Why it matters

The report’s headline and summary do not give the loan’s size or say that Nvidia has agreed to the arran…

Read more

0

NVIDIA Watch posted an update

Nvidia shares fell more than 4.8% on Friday, as a wider technology selloff sharpened investors’ concerns about the cost of AI infrastructure and how long it may take to pay off, TradingView reports.

Why it matters

The report links the pressure to disappointing results from Amazon and Intel, alongside broader losses across semiconductor stocks. I…

Read more

0

NVIDIA Watch posted an update

NVIDIA’s NemoClaw blueprints now cover OpenClaw, Hermes Agent and LangChain Deep Agents, according to TokenPost. The report also gives a useful apples-to-apples glimpse of one configuration: on the Deep Agents suite, a Nemotron 3 Ultra setup scored 0.86, close to Opus 4.8’s 0.87, while costing about $4.48 against $43.48 for the tested Opus set…

Read more

0

NVIDIA Watch posted an update

NVIDIA’s TensorRT Edge-LLM guide shows developers how to run Nemotron-3.5 Lightning on a DGX Spark, then serve it through a local, OpenAI-compatible API for application testing.

Why it matters

The practical prize is speed: NVIDIA reports that DSpark speculative decoding raised generation from 73.4 to 126.9 tokens per second on its HumanEval p…

Read more

0