①NVIDIA will hold its annual GTC conference next week in San Jose, where, in addition to the highly anticipated computing hardware, it will also provide on-site "lobster" installation services; ②The capital market is also looking forward to key technological advancements in inference chips, optical communication, and custom CPUs; ③Analysts believe that if the inference chip is confirmed, it indicates that NVIDIA will accelerate its entry into the AI inference market, competing with self-developed chips from companies like Google and Microsoft.
Cailian Press, March 13 (Editor: Shi Zhengzhi) The craze for installing 'lobsters' is coming to San Jose, California. The computing giant $NVIDIA (NVDA.US)$ will hold its annual GTC conference next week. In addition to launching new products this year, there will be a special live interactive session where NVIDIA technical experts will help attendees install 'lobsters'.
According to NVIDIA's official blog, during the GTC conference from Monday to Thursday next week, except for the two hours of Jensen Huang’s speech on Monday, the 'Building Lobster' event will run continuously at the GTC Park, where attendees can deploy an always-on AI assistant on-site.
The computing giant subtly pointed out that participants are encouraged to bring NVIDIA chip-powered devices to the event, including the personal supercomputer DGX Spark, whose computing power (and price) far exceeds that of the Mac Mini. Hardware purchases will be available on-site, and 'Lobster' can also be deployed directly in the cloud.
Official data shows that more than 30,000 people from over 190 countries and regions worldwide will attend the event in person next week. Considering that many of them are professional developers, this year’s GTC may not only become the 'Lobster Conference' but could even turn into a 'Shrimp Battle'.
What are the other key points of interest?
GTC (GPU Technology Conference) is NVIDIA's flagship annual event, where the company typically launches new products and articulates its vision for industry development.
For the capital markets, the keynote speech by NVIDIA’s leader Jensen Huang at 11:00 AM local time on Monday (2:00 AM Beijing time on Tuesday) will be the highlight of the event.
Ask Futubull AI:In terms of hardware, the roadmap for computing chips, potential inference chips, and proprietary optical communication products have become key concepts targeted by capital.
Vivek Arya, a senior analyst at Bank of America Securities, outlined in a research report this week that there are three main areas of focus:
1) The latest product roadmap extending to Feynman architecture;
2) A series of new co-designed (customized) and modularly split products (e.g., CPX for pre-filling during the inference stage, and LPU for low-latency decoding);
3) Proprietary optical interconnect technology for large-scale system expansion (e.g., switches integrated with CPO).
Arya also anticipates that NVIDIA may discuss the 102.4T Spectrum-6 switch (compatible with the Rubin platform) and the 115T Quantum-X switch (using co-packaged optics CPO technology). NVIDIA might even jointly release a customized x86 CPU with Intel, further expanding its adoption in enterprise data centers and potentially extending to the consumer-grade CPU market.
Arya stated that he does not expect Jensen Huang to formally provide sales guidance for the next two years, but news regarding the mass production ramp-up of the Rubin architecture could boost the stock price.
As for the 'unprecedented chip' previewed by Jensen Huang, current speculation is focused on the LPU jointly released by NVIDIA and inference chip company Groq.
Inference refers to the process by which an AI model applies knowledge learned during training to make decisions and generate responses. Unlike the computing power required for training AI, faster and lower-cost inference is considered one of the final bottlenecks for widespread AI application deployment. If this chip is successfully launched as expected, it means that NVIDIA, which dominates the training market, will officially address competition from custom chips developed by companies like Google and Microsoft.
Kevin Cook, Senior Equity Strategist at Zacks Investment Research, also stated that investors are looking forward to Jensen Huang clarifying how NVIDIA will manage its relationship with Groq. It is reported that NVIDIA paid $20 billion at the end of last year to obtain a patent license from Groq. Additionally, Groq's founder Jonathan Ross, President Sunny Madra, and key team members have all joined NVIDIA, drawing widespread attention to this collaboration.
On the software front, reports this week revealed that NVIDIA will release an open-source platform for enterprise AI agents, named NemoClaw (which also references 'lobster'). This platform will provide enterprises with a structured approach to building and deploying AI agents, which are software capable of autonomously performing multi-step tasks.
The 'Jensen Huang Roundtable' Focuses on Open Models
As per recent years' convention at GTC, Jensen Huang will hold a roundtable discussion next week to showcase the key industry directions NVIDIA considers most important.
This year's theme is 'The Current State and Future of Open Models.' A lineup of emerging companies in the industry, including Cursor, a star in AI programming, Perplexity, an AI search engine, and OpenEvidence, a medical AI assistant, as well as Mira Murati, former CTO of OpenAI and founder of Thinking Machines Lab, will be in attendance.
According to NVIDIA's announcement, one of the most significant developments in the AI field over the past year has been the rapid progress of open frontier models. Open innovation is accelerating advancements across companies and industries, clearly demonstrating that AI will be ubiquitous. Jensen Huang will join these industry leaders in a candid discussion about 'the latest developments and future of open frontier models.'
Looking to pick stocks or analyze them? Want to know the opportunities and risks in your portfolio? For all your investment-related questions,just ask Futubull AI!
Editor/Rocky