On October 23rd, Sichuan Xinjinlu Group announced that its revenue in the first three quarters was 1,590,168,868.36 yuan, a decrease of 7.69% year-on-year, and the net income attributable to shareholders of the listed company was 33,098,768.63 yuan, a decrease of 28.99% year-on-year.$NVIDIA (NVDA.US)$ CEO Jensen Huang launched a new supercomputer named Gefion in Denmark, equipped with 1528 GPUs, planned for use in drug research and development, disease diagnosis, treatment, complex life science challenges, among other areas. Huang believes that "the era of computer-aided drug research will definitely arrive within ten years, this will be the decade of digital biology."

The day before, on October 22 local time, Huang Renxun stated at the Gartner IT expo in Orlando, USA, that Nvidia is developing an "AI brain", which is the foundational technology powering artificial intelligence. Huang also mentioned that the next major trend is agent-based artificial intelligence, models capable of making decisions with minimal human intervention.

In the capital markets, on Monday, Nvidia's stock price hit a new all-time high, with a total market value exceeding $3.5 trillion at one point; on Wednesday, the semiconductor sector of the US stock market collectively adjusted in the early trading session, with Nvidia dropping nearly 2% by the time of publication.

Next, investors are most concerned about whether this semiconductor rally led by Nvidia can continue?

Nvidia makes big moves

Nvidia CEO Huang Renxun revealed during his visit to Denmark on Wednesday that$Taiwan Semiconductor (TSM.US)$With the help of tianfeng, the latest ai chip - the Blackwell chip, has had a design flaw affecting production fixed. As of the time of publication, taiwan semiconductor is up nearly 2%.

In March, Nvidia launched the Blackwell chip, which achieved a significant performance improvement, with speeds 7 to 30 times faster than the existing H100 chip, while reducing power consumption by 25 times. Nvidia had previously stated that the Blackwell chip would ship in the second quarter but was delayed, potentially impacting$Meta Platforms (META.US)$, Alphabet, Google, and$Microsoft (MSFT.US)$customers. Huang Renxun stated on Wednesday that the company's latest Blackwell chip had a design flaw, although its functionality was good, the design flaw resulted in a very low yield rate." He stated that this flaw was "100% Nvidia's fault", and it was only with the help of Taiwan Semiconductor that they were able to recover from this setback and "recover work at an amazing speed."

Morgan Stanley expects that NVIDIA's revenue from the Blackwell chip in the fourth quarter of 2024 will reach $10 billion, with a very high profit margin, and is very bullish on the revenue potential NVIDIA will gain from the Blackwell chip.

It is worth mentioning that Huang Renxun has launched a new supercomputer named Gefion in Denmark, which was developed in collaboration with NVIDIA, Novo Nordisk Foundation, Danish Export and Investment Fund. It has 1528 GPUs and is planned to be used in areas such as drug development, disease diagnosis, treatment, and complex life science challenges. Huang Renxun believes, 'The era of computer-aided drug development will definitely arrive within ten years, which will be the decade of digital biology.'

Huang Renxun also mentioned that the EU lags far behind the US and China in AI investment. Although Europe has only a few artificial intelligence companies, such as Mistral in France and Aleph Alpha in Germany, the EU has passed the world's first comprehensive rules governing AI, which took effect in August. Huang Renxun stated during his visit to Copenhagen, 'The EU must accelerate the development of artificial intelligence, with every country realizing that data is a national resource.'

Huang Renxun: Building the 'AI Brain'

On Tuesday, NVIDIA CEO Huang Renxun stated, 'We are witnessing the dawn of a new industrial revolution, except the driving force of this revolution is not steam or electricity, but artificial intelligence (AI).'

On that day, Huang Renxun told thousands of enterprise technology leaders at the Gartner IT Symposium in Orlando, USA, 'Your company must transform into an AI-driven organization and prepare for this 'new industrial revolution'.' Huang Renxun also mentioned that NVIDIA has put this concept into practice, applying AI to chip design, software writing, and supply chain management.

Huang Renxun stated that in the long term, NVIDIA is building the 'AI Brain.' The concept is to gather knowledge about how the company operates, its business processes, and customer interactions, and hand it over to AI. The ultimate goal is to transform this information into an AI that Chief Information Officers and Chief Executives 'can interact directly with.' He added, 'You still need programming, but you will communicate with AI in a structured way.'

It is understood that to achieve this goal, NVIDIA announced the launch of a tool to make it easier for artificial intelligence to access and understand PDF files - Huang Renxun mentioned that for traditional AI, most of the time, 'unstructured data' such as emails and PDFs are difficult to capture. He also stated that in the future, companies will have 'a large number of digital employees' or AI agents working alongside human employees in roles such as marketing, sales, engineering, and supply chain. According to Huang Renxun, to achieve this goal, everyone, including business and technology leaders, needs to learn how to guide AI.

Huang Renxun also stated that the next major trend is agent artificial intelligence, which is a model capable of making decisions with very little human intervention. Nvidia has been collaborating with companies like $SAP SE (SAP.US)$N/A.$ServiceNow (NOW.US)$to incorporate such agents into their systems.

Can the semiconductor market driven by AI continue?

In the capital markets, since mid-September, Nvidia's stock price has started a new round of upward trend. On Monday this week, Nvidia's stock price surged over 4%, reaching a new all-time high, with a total market value exceeding $3.5 trillion, approaching Apple. However, today the overall sentiment in the semiconductor sector is poor, and Nvidia is now down nearly 2%.

Nvidia's recent gains reflect $Meta Platforms (META.US)$N/A.$Amazon (AMZN.US)$Investors are becoming more optimistic about investing in artificial intelligence before the release of financial reports at the end of this month. The next question is whether these companies will commit to further increasing spending on AI.

CFRA analyst Angelo Zino stated that Meta may increase its capital expenditure by about 20% to around $46 billion in 2025, mainly to support AI projects. Blackrock's Global Chief Investment Strategist Li Wei wrote in a research report: "We see enough room for the artificial intelligence theme: it is still in the early stages of development. As technology companies continue to surpass high profit expectations, valuations of AI beneficiaries are supported. We believe optimism can expand."

DBS Bank stated that after the challenging year of 2023, driven by strong expansion in demand for AI-related semiconductors, the semiconductor industry is expected to show a strong rebound. In 2024, the logic and memory markets are expected to experience double-digit growth; the market expects the total size of the semiconductor market to grow by 16% this year, reaching $611 billion. With strong demand from cloud hyperscalers and data center providers, the demand for graphics processing units (GPUs) that support AI is more intense than ever. The memory market is crucial for data center information storage and retrieval, and the market expects a 77% growth in this year's AI-driven data center boom.

DBS Group remains bullish on the semiconductor sector, expecting the semiconductor industry to continue benefiting from the transformative impact of AI in the long term. Although the AI frenzy will eventually fade, considering Amara's Law (which states that the market often overestimates the short-term impact of a technology while underestimating its long-term impact), this is a key factor to consider when evaluating investment value in semiconductor companies and AI beneficiaries. While short-term market fluctuations may tempt people to react, the true attractiveness of semiconductor and AI investments lies in their sustained long-term growth potential.

Tianfeng Securities stated that the semiconductor industry cycle is currently at a relatively low level in the long cycle, with the second half of the year expected to enter the traditional peak season. Influenced by factors such as the release of new flagship phones, singles' day sales, and other consumer festivals, it is expected that the industry's end sales revenue will continue to grow month-on-month. Attention should be paid to the demand-side innovation, with priority given to AI terminals accepted by consumers, which are likely to become new popular applications. In the long term, semiconductor blue chip companies covered by Tianfeng's electronic team are currently at relatively low valuations. Companies that continuously optimize and iterate their operations are expected to gain better market share and profit levels in the next cycle peak. In terms of innovation, it is expected that artificial intelligence/satellite communications/MR will be significant industry trends, and individual stocks in the industry chain are expected to continue to reflect thematic opportunities as technological innovation progresses.

Tianfeng Securities pointed out that semiconductor equipment materials, as core elements of semiconductor production, currently have a low domestic substitution rate, with supply chain security being greatly affected by international politics, leading to an urgent need for domestic substitution. With continuous policy support and the drive from top domestic storage companies/wafer foundries, domestic equipment materials are expected to accelerate line validation. The sector is expected to benefit from local capacity expansion and the increase in domestic substitution rate, making it a favorable sector for investment.

Editor/rice