Author: Huang Yu

The weather is good today The weather is good today.

Nvidia is once again at the center of the AI ​​war.

The GPU leader, Nvidia, had a roller coaster week this week: it fell 7% on Tuesday, rose 13% on Wednesday, and then fell nearly 7% on Thursday. The volatility of its stock price even exceeded that of Bitcoin. Data shows that Nvidia's 30-day option implied volatility has soared to 71%, while the Bitcoin DVOL index (a gauge of 30-day option implied volatility) has fallen to 49%.

Behind the sharp volatility of Nvidia's stock price, the shadow of cost is looming over the technology giant, and the market's concern about the return on enormous AI investments is increasing.

The short sellers, represented by Wall Street investment banks and hedge funds, believe that technology giants spend huge sums of money on the AI field, but fail to generate sufficient returns, and that the current application scenarios for AI are limited. They view such huge investments as unwise and dangerous.

The bulls, represented by firms like Mag 7, believe that increasing capital expenditures in the AI field is necessary now, otherwise, the AI era that is about to come will be missed.

Wall Street, where money talks, attaches more importance to the investment return ratio.

On the one hand, Wall Street, which is gradually losing patience, believes that the capital expenditures of technology giants in the AI field are so high, yet they have not brought corresponding returns and more efficient applications. In the past two years, there have only been two AI products that have become phenomenons, ChatGPT and Github Copilot.

Goldman Sachs pointed out in a report at the end of June that the development of AI technology may not be as rapid as expected, and its cost-effectiveness ratio may not be as attractive as imagined.

It is predicted that in the next ten years, AI can only increase US productivity by 0.5%, and its contribution to GDP growth will only be 0.9%.

Barclays also pointed out in a report in August that the frenzy of technology giants pouring money into AI is due to FOMO, fearing to miss out on the development opportunities of AI.

Barclays' latest research report pointed out that although AI technology is still in its infancy, the capital expenditure of large factories on AI has shown an irrational prosperity, and FOMO sentiment dominates. As this sentiment fades, major factories will gradually cut back on AI investment next year.

Analysts predict that capital expenditures in the AI field will reach $167 billion from 2023 to 2026, and this number is based on optimistic expectations of demand for AI products.

However, in sharp contrast to this, the incremental revenue of AI cloud services is only expected to be $20 billion by 2026.

And according to media reports on Friday, Elliott Management, a top hedge fund on Wall Street, was even more radical, pointing out that large technology giants, especially Nvidia, were in a bubble, and the AI technology that pushed its stock price skyrocketing was overhyped.

Elliott believes that current AI cannot really work efficiently and consumes a lot of energy:

So far, artificial intelligence has failed to deliver on its promise of significant productivity gains. There are almost no practical use cases beyond summarizing meeting notes, generating reports, and helping computers code.

Artificial intelligence is actually a software that has so far not delivered the value commensurate with the hype.

Jim Tierney, strategy director at AllianceBernstein, said: "Technology giants are increasing investment in AI and creating an atmosphere of 'trust us'. But investors are still unsure about the corresponding business models and returns. But from the perspective of total expenditures and returns, investors will not be reassured."

Silicon Valley's tech giants believe that AI will change the future and choose to continue pouring money into it.

On the other hand, there are technology giants who persist in seeing the future of AI, choosing to continue pouring money into this field, and resisting pressure from doubt.

Including Microsoft, Alphabet, Amazon, and Meta, in the latest quarterly financial reports, they revealed that capital expenditures in the first six months of 2024 will increase significantly, and the total amount will reach $106 billion, and will further increase investment in the next 18 months.

Alphabet, Google's parent company, saw its capital expenditures skyrocket by 90% to $25 billion in the first half of 2024.

In just the fourth quarter, Microsoft's capital expenditures increased by 78%, reaching $33 billion.

Amazon's real estate and equipment investment (including e-commerce and logistics network expenditures) surged by 27% to reach $32.5 billion in the first half of this year, and is expected to see a sharp increase in capital expenditures by 2024.

Even though the stocks of Google, Microsoft, and Amazon were immediately sold off after the financial reports were released, the executives of these technology giants still insist on expanding their investment in the AI field.

Google CEO Sundar Pichai said that AI products need time to mature and become more useful.

AI is expensive, but the risk of underinvestment is greater. Google may have invested too much in AI infrastructure, including purchasing NVIDIA GPUs. Even if the AI boom slows down, the data centers and computer chips the company has purchased can be used for other purposes. For us, the risk of underinvestment is far greater than the risk of overinvestment.

Amazon's CFO Brian Olsavsky said that these expenditures will mainly be used for new cloud computing infrastructure, and generative AI is now a key business for Amazon.

Meta's CEO Zuckerberg also stated last week that in order to maintain Meta's leading position in the AI field, the company has spent billions of dollars to purchase Nvidia's GPUs to develop and train advanced AI models. Nevertheless, he also admitted that hype around AI could lead to excessive investment.

Rather than waiting and risking falling behind, Meta is increasing its investment, as being left behind would put them in a disadvantageous position for the most important technology in the next 10 to 15 years. It is expected that Meta's capital expenditures for this year could reach $40 billion.

Zuckerberg estimates that the computing power needed to train the next generation of large-scale language models will be nearly 10 times that of the previous version, and he also admits that it will take 'years' for Meta's AI chatbot to make money on its own.

Analysts at Dell'Oro Group predict that as much as $1 trillion in funds may be invested in AI data centers and other infrastructure in the next five years.

What is the key to this 'bull and bear' war about AI?

Regardless of how Wall Street views the continued investment by tech giants in AI, the key point of this tug-of-war ultimately revolves around whether or not the momentum of these companies' investment can be sustained, and when AI technology will achieve greater breakthroughs.

Firstly, the confidence of these tech giants to continue investing in AI comes from the fact that their core businesses remain relatively stable and profitable.

For example, Meta's share of the core business digital advertising market continued to grow in the second quarter, with growth in this segment of the business coming mainly from Facebook and Instagram, both of which brought in a year-on-year increase of 22% in advertising revenue, resulting in a year-on-year increase of 73% in Meta's net income to $13.47 billion in the second quarter.

Zuckerberg also emphasized during the earnings call how AI is helping to drive growth in advertising:

AI improves recommendation features and helps people find better content, and makes advertising more effective. These products are now brought to scale.

In contrast, Google's advertising revenue in the second quarter only increased by 11% to $64.6 billion, but its core business remains healthy enough to support expanding expenditures.

Although the tech giants have deep pockets and are willing to invest in AI, analysts at Barclays predict that by 2025 or later, some major players will retreat and reduce their (AI) capital expenditure plans.

As technology advances, breakthroughs in the smaller base model fields recently, and many AI products and search functions will be able to move from the cloud to local PCs or smartphones in the next few years. The cost of AI inference may be greatly reduced, alleviating the need for such large capital expenditures.

As for when AI technology will achieve greater breakthroughs, it is more like an unknown number. After investing heavily in AI, currently only two phenomenal AI products, ChatGPT and Github Copilot, have been developed. There is still a long way to go in the future.

Barclays predicts that by 2026, AI capital expenditures (about $167 billion) will be sufficient to support more than 12,000 ChatGPT-scale products, with about $70 billion invested in training foundational models and the remaining approximately $95 billion used for inference.

Many new AI-based services are expected to emerge at that time, and these services will drive positive market and industry development.

But what can be determined is that in the AI competition of tech giants, Nvidia is still the biggest winner. Due to the investment in AI capital expenditures by Amazon, Google, Microsoft, and Meta, it has mostly flowed into building datacenters and purchasing Nvidia's GPU products, and this momentum may continue until next year.

Although Nvidia will not release its latest quarterly earnings report until the end of August, analysts generally expect that Nvidia's datacenter revenue for the new quarter will reach nearly $25 billion, equivalent to the datacenter revenue for the entire year of 2023.

It will soon be time to verify the "AI faith".