Entering 2024, artificial intelligence is still the "ammunition" that dominates the earnings reports of technology companies. Last week, six of the "Big Seven" in the U.S. stock market reported earnings, and the stock prices of other tech companies except Apple and Tesla were boosted by the AI capital boom.
Among them, the gap between the market value of Microsoft and Apple has further widened, and the stock prices of Meta and Nvidia have soared, but whether Nvidia's special chip sales for the Chinese market can remain strong in the future has become the biggest concern of investors.
The "Big Seven" share price divergence.
Over the past month, the AI boom has continued to drive related tech stocks higher. As of last Friday**, Nvidia's stock price rose nearly 5%, the company's stock price has risen by nearly 40% in the past month, and its market value has soared by nearly $300 billion, becoming the company with the largest monthly market value growth in the history of the U.S. stock market, and the current market value has exceeded 1$5 trillion.
Although Nvidia's earnings report has not yet been released, Nvidia's "big customers", including Microsoft, Google, Amazon and Meta, all announced their earnings last week, with Meta's stock price soaring nearly 20% after the earnings report, and Zuckerberg's personal wealth soaring by $30 billion overnight.
At the same time, Microsoft, the world's most valuable company, continues to widen the gap with Apple. After the release of Microsoft's earnings report, its market capitalization stood firmly above $3 trillion, and Apple's market capitalization stood at 2$85 trillion. Driven by demand for artificial intelligence, Microsoft's market capitalization has increased by nearly $160 billion in the past month.
After the announcement of Tesla's earnings report, the company has fallen out of the "Big Seven", and its market value has been overtaken by Warren Buffett's Berkshire Hathaway. In January this year, Tesla's stock price became the worst performing company among the world's top 20 companies by market capitalization, with nearly a quarter of its market value evaporated, and its market value has now fallen below $600 billion.
After the release of the "Big Six" earnings report, the market will pay close attention to Nvidia's earnings report, and the company's latest results will be released on February 21.
Just last week, it was reported that Nvidia has begun "taking orders" to sell new Chinese versions of custom AI chips to Chinese customers, which are expected to be priced similarly to current Chinese competitors.
According to the information circulating on the Internet before, the highest-end of NVIDIA's Chinese version of the customized AI chip is H20, but the computing power of the chip is lower than NVIDIA's flagship AI chip H100, and it is also lower than the H800 chip previously sold for the Chinese market.
Nvidia did not respond to Yicai on the latest pricing of AI chips for the Chinese market. The market expects that the ** of this chip will be about 120,000 yuan. If a server is composed of 8 chips, the price of the server is expected to be more than 1 million yuan. In comparison, the previously launched server equipped with 8 NVIDIA H800 chips is priced at about 2 million yuan.
The first financial reporter inquired about Nvidia's official **, and has not found relevant pricing information and shipping date about H20, but according to previous industry expectations, the first batch of small batch shipments of H20 may be in the first quarter of this year, and large-scale production will begin in the second quarter.
However, industry insiders are also worried that Nvidia's AI chips may not be able to attract Chinese customers after the slowdown. A veteran in the field of chips told the first financial reporter: "The parameters seen at present do not represent the actual performance of the product, although Nvidia's chips may not be as good as competitors in some indicators, but it is still necessary to focus on actual use testing, looking at the cost performance of hundreds of GPUs together in clusters, after all, NVIDIA's ecology is more mature, and it is effortless to use." ”
AI investment will increase significantly.
The earnings reports of these tech giants will reveal an important message: to win the battle of artificial intelligence, you must spend money before you can make money. The continuous investment plan of enterprises in artificial intelligence has also become one of the important factors to boost market confidence.
Microsoft CEO Satya Nadella told investors that Microsoft will continue to prioritize AI as a corporate investment and is committed to expanding AI investment and cloud computing efforts.
Amazon CEO Andy Jassy said during the earnings conference that generative AI will eventually generate tens of billions of dollars in revenue for Amazon in the coming years. He pledged to invest more in generative AI projects and the necessary infrastructure, including research and development in AI chips.
Sundar Pichai, Google's chief executive, said investing in infrastructure such as data centers was "key to achieving ambitious AI goals," saying the company had cut back on non-priority projects and invested in automating certain processes. "We will continue to invest responsibly in data centers and cloud computing to support the next wave of growth in AI services for us and our customers. Pichai said.
On the other hand, tech companies have identified jobs that can be replaced with AI, allowing them to invest in AI while continuously optimizing the company's costs.
Within Google Cloud, the company will cut expenses by reallocating resources to the most important projects, slowing hiring, improving technology infrastructure and using artificial intelligence to streamline processes, Pichai said. Google expects capital expenditures of $11 billion in the fourth quarter, primarily from investments in infrastructure, servers and data centers. The company expects full-year 2024 capital expenditure to be "significantly higher than in 2023".
Meta CEO Mark Zuckerberg stressed that the company will continue to work on artificial intelligence while cutting costs more broadly. Meta said it plans to raise capital spending to $30 billion to $37 billion this year to strengthen its research and product development in the field of artificial intelligence.
Last month, Meta said it was replenishing its AI infrastructure "ammunition bank" and planned to buy 350,000 H100 GPU chips from Nvidia by the end of the year, bringing the company's total GPU to about 600,000, making it one of the most powerful AI systems in the world.
Even Apple CEO Tim Cook, who rarely mentions AI, highlighted the company's opportunities in generative AI during his earnings conference. Cook said AI is an important area of investment for the company, and Apple will announce its progress on AI later this year.
What are the areas of focus?
The financial reports of technology companies mainly talked about two key directions, one is AI model-as-a-service, that is, large AI models that customers can use and customize according to their needs; The other is the artificial intelligence "AI agent", which includes chatbots as well as other writing assistants.
This means that the AI model will no longer stay in the lab, but will truly become a tool to solve productivity. "You're going to see people using these tools to actually be more productive," Nadella said. ”
He also pointed out that Microsoft Azure is a major "model-as-a-service" offering, emphasizing that customers can access a range of language models of different sizes, including some from Cohere, Meta, and Mistral, as well as open-source models, without having to manage the underlying infrastructure.
Amazon's Jassy said enterprise customers are looking to customize the use and build of existing models, noting that Amazon's bedrock is a key focus area. "What we're seeing is that customers want choice. "They don't want to have just one pattern to build, they need to build different models for different applications, and they want to try all the different scales." ”
Tech companies are also bullish on artificial intelligence, including Microsoft's Copilot, Amazon's Rufus, and more. Nadella noted that Microsoft's Copilot assistant is an example of an AI application that has "evolved" in terms of productivity benefits and successful business models.
Amazon's Rufus is an AI-powered generative shopping assistant trained on company catalogs, customer reviews, user Q&A pages, and more. "Every business we have contains multiple generative AI applications that we're building, and they're in different stages of development, many of which are already in development and others are in development," Jassy said. ”
According to Google executives, the company's artificial intelligence**duet AI is designed to increase productivity and accomplish simple tasks. In Google Cloud, Duet AI helps software developers at companies like Wayfair and GE, as well as cybersecurity analysts at Spotify and Pfizer, Pichai said. He also said that Duet AI will soon integrate Google's large model Gemini to power Google's Bard chatbot.
Google also wants to provide an artificial intelligence that can complete more tasks on behalf of the user, including performing tasks in Google Search. Zuckerberg also said at the company's conference that Meta will also focus on building useful AI.