DiscoverGold
17 hours ago
$NVDA a bearish sign has emerged; a Wyckoff Distribution pattern near all-time highs suggests institutions may be quietly unloading positions ahead of the report
By: TrendSpider | February 2, 2025
• NVDA has declined over 12% in 2025, erasing more than $500 billion in market value. Unlike its tech peers, it has yet to report Q4 earnings, set for February 25. Analysts expect revenue of $38.04 billion (up 84% YoY) and EPS of $0.85 (up 83% YoY), maintaining its impressive growth streak. However, a bearish sign has emerged; a Wyckoff Distribution pattern near all-time highs suggests institutions may be quietly unloading positions ahead of the report.
DiscoverGold
doc2016
2 days ago
deepseeks seems to have used other nvidia ip besides the gpu hardware. "did deepseek use risc-v gpu for it's llm? distillation might be considered a natural extension/evolution of a llm, but they need to make sure they maintain control of the model data/libraries and get paid appropriately?
gemini ai says, "
DeepSeek's recent advancements in large language models (LLMs) do not appear to directly involve RISC-V GPUs. Their primary innovation lies in their efficient training methods and model architecture, not the underlying hardware.
Here's a breakdown of what we know about DeepSeek's technology:
Mixture-of-Experts (MoE) Model: DeepSeek utilizes a Mixture-of-Experts (MoE) architecture for its LLMs. This involves breaking down the model into smaller, specialized "expert" networks that handle different aspects of the task. This approach allows for greater efficiency and scalability compared to traditional monolithic models.
Efficient Training: DeepSeek has demonstrated impressive efficiency in training its LLMs, achieving comparable performance to leading models like GPT-4 while using significantly less computational resources and training time. This is attributed to their innovative training methods and optimizations.
PTX Programming: DeepSeek's models are programmed using PTX, a low-level parallel thread execution language developed by NVIDIA. This allows for direct control over GPU operations and further optimization of performance.
No Mention of RISC-V: There is no public information or announcements suggesting that DeepSeek is using RISC-V GPUs for its LLMs. Their focus seems to be on software and algorithmic optimizations rather than specific hardware platforms.
Potential Future for RISC-V in LLMs:
Open and Customizable: RISC-V's open and customizable nature makes it an attractive option for future AI hardware development. It could potentially enable the creation of specialized AI chips optimized for specific LLM architectures or tasks.
Growing Ecosystem: The RISC-V ecosystem is rapidly growing, with more companies and research institutions exploring its potential for AI applications.
Future Collaboration: It's possible that DeepSeek might explore using RISC-V GPUs or collaborating with RISC-V hardware developers in the future to further enhance the performance and efficiency of its LLMs.
Key takeaway:
DeepSeek's current advancements in LLMs are primarily driven by software and algorithmic innovations, not the use of RISC-V GPUs. However, the potential for RISC-V to play a role in future AI hardware development, including for LLMs, is significant, and we might see closer ties between DeepSeek and the RISC-V ecosystem in the years to come.
Sources and related content"
might be they transformed nvidia to opensource as well? for function or for cover of their activities?
ShawnP123
2 days ago
Another article which is interesting to read:
https://finance.yahoo.com/news/semtech-corporation-smtc-powering-nvidia-015948475.html
Regarding the claims regarding the insufficient funding, Trump said that he doesn’t know but he hopes that they do.
“I don’t know if they do, but you know, they’re putting up the money. The government’s not putting up anything, they’re putting up money. They’re very rich people, so I hope they do”.
That's a strange comment from Trump considering his support for the program.
JJ8
2 days ago
Obviously something happened since the morning.
Chart is based on what happened already, in the past. But life goes on and the future unfolds as we see, as things change. Guesses base on what happened in the recent past, in case you didn't know that.
May be it was the meeting of Jensen with Trump that may have caused it?
The share price was based on thing that happened before and they are always subject to what happens after. We shall know more about it tomorrow, hopefully.
The future has a mind of its own. Chart projections are guesses based on the past. All and everything is subject of what unfolds as things change and depends how impactful a change it is.
I was surprised as well. But, I listen what the market is doing and trying to convey to us.
We shall see what comes and unfolds ahead and in due time.
What is your understanding?
Did you think events follow what the chart/methodology guesses is always a certainty?
Take care, cheers & GLTY
ShawnP123
3 days ago
From ABC News
https://abcnews.go.com/Technology/wireStory/president-trump-meet-nvidia-ceo-jensen-huang-white-118320967
The Republican president is banking on AI to foster economic growth and draw hundreds of billions of dollars in investments, but he also saw the performance of China’s DeepSeek AI technology as a sign that the technology can be developed more cheaply.
Speaking Monday to House Republicans in Miami, Trump called the DeepSeek news “positive” if it’s accurate because “you won’t be spending as much and you’ll get the same result.”
Trump signed an order on his first day in office last week that said his administration would “identify and eliminate loopholes in existing export controls,” signaling that he might continue and harden Biden’s approach.
So, from what I've read, it seems like a large part of the meeting could be about further limiting Nvidia's product to China which would have a definite impact on sales. This would mean another cut in Nvidia's revs. Of course you won't hear anything from Jensen.
So people can come up with all kinds of hype but the SP tells the story and not some wishful thinking.