Nvidia chips are selling like crazy: revenue increased by 265%, profits increased by 769%

Latest update time：2024-02-22

Reads：

????If you want to see each other often, please mark it as a star???? and add it to your collection~

On Wednesday, the much-anticipated Nvidia reported fourth-quarter earnings. Data show that the company's revenue in the past quarter was US$22.1 billion, compared with the expected US$20.55 billion, and earnings per share were US$4.93, compared with the expected US$4.64, significantly exceeding analysts' expectations. Revenue increased 22% from the previous quarter and 265% from the same period last year.

Shares of the Santa Clara, California-based company rose late in the session, adding more than $80 billion to its market value to $1.75 trillion.

Already huge demand for the company's data center chips and graphics processing units (GPUs) continues to grow as companies scramble to expand their artificial intelligence offerings. Nvidia's chips dominate the global artificial intelligence chip market, including well-known companies such as Microsoft, META and Google as its customers.

According to Reuters, Nvidia's forecast exceeded expectations, but by a smaller margin than last year. In the first three quarters of 2023, Nvidia reported quarterly revenue that exceeded analyst expectations by 10% to 20%.

Insider Intelligence analyst Jacob Bourne said bluntly: "Nvidia's another sensational quarter raises the question of how long its soaring performance will last." "It has a huge lead in the growing global artificial intelligence chip field, but Can’t be satisfied with the status quo.”

Data centers surge, driving NVIDIA

In NVIDIA's revenue data, the data center is undoubtedly the most shining department. According to statistics, the fourth-quarter sales of this AI-focused division soared from US$3.6 billion last year to US$18.4 billion annually, a growth rate of more than 400%. It is worth mentioning that Nvidia’s data center revenue exceeded analysts’ expectations of $17.2 billion.

Nvidia's boom has coincided with huge earnings growth driven by AI - its quarterly profits are more than 10 times higher than four years ago. Analysis shows that Nvidia's net profit increased 770% from the same period last year to US$12.3 billion, which also exceeded analysts' expectations of US$10.4 billion. -- but the question this raises, according to Fortune, remains whether Nvidia can deliver the financial results needed to justify its valuation and place it among the most valuable companies. Wall Street analysts largely believe Nvidia can achieve this goal, as average forecasts predict that Nvidia's profits and sales will each increase 60% or more this fiscal year.

Such an outstanding performance led Carson Group strategist Ryan Detrick to even comment: "There is nothing more certain than death, taxes and Nvidia earnings exceeding expectations."

"Accelerated computing and generative artificial intelligence have reached a tipping point. Demand is surging across companies, industries and countries around the world," Nvidia founder and CEO Jensen Huang said in a press release. "Our data center platform is powered by an increasingly Diverse drivers are supporting demand for data processing, training and inference from large cloud service providers and GPU specialist providers, as well as enterprise software and consumer internet companies, across industry verticals led by automotive, financial services and healthcare. It’s reached billions of dollars,” Huang added.

While Nvidia's data center division is driving the ship, its gaming business remains an important part of the company. The segment's revenue peaked at $2.9 billion in the fourth quarter. Investors expected revenue of $2.7 billion, up from $1.8 billion last year. Full-year revenue increased by 15% to US$10.4 billion; in terms of professional visualization, Nvidia's fourth-quarter revenue was US$463 million, an increase of 11% from the previous quarter and a year-on-year increase of 105%. Full-year revenue increased by 1% to US$1.6 billion; in terms of automobiles, fourth-quarter revenue was US$281 million, an increase of 8% from the previous quarter and a decrease of 4% year-on-year. Full-year revenue grew 21% to $1.1 billion.

As you can see from the figure above, Nvidia's revenue in the fourth quarter of fiscal year 2024 increased by 265% year-on-year, and its profits increased by 769%. If we look at it from a fiscal year perspective, as shown in the figure below, Nvidia's revenue in fiscal year 2024 will reach $60.9 billion, an increase of 126% year-on-year. The company's gross profit margin will also increase from 56.9% to 73.7% in fiscal 2024.

If this figure is followed, Nvidia will become the world's second largest semiconductor manufacturer in the past fiscal year, second only to TSMC's US$69.3 billion in 2023. If wafer foundries are excluded, NVIDIA will become the world's number one chip company.

Nvidia shares are up more than 30% year to date, making it among the most valuable companies. As of February 20, in the past 30 trading days, Nvidia's stock changed hands worth about US$30 billion on average every day, ahead of electric car maker Tesla, which averaged US$22 billion per day during the same period.

It’s cold at the top: good news, bad news

For Nvidia, the past few years must have been a dream, but standing at this peak, they must be overwhelmed by the cold. So much so that Huang Renxun has always remained humble when talking about the status of the company itself. This also puts them facing unprecedented pressure and challenges.

On the bright side, Nvidia is solving several problems. The first of course is the product delivery issue that bothers everyone.

According to the latest report to customers from UBS, Nvidia has significantly shortened the delivery cycle of AI GPUs (that is, the time interval between customers placing an order and delivery), from 8-11 months at the end of last year to 3 now. -4 months. UBS said the shift suggested a possible peak in recent growth, but could also signal trouble for Nvidia's future growth prospects. However, UBS still raised its outlook for Nvidia, raising its profit forecast and target price to $850.

In Tomshardware's view, the shortened delivery time is a double-sided edge: on the one hand, it shows that Nvidia's production partner TSMC is expanding its CoWoS packaging capacity, which means better supply and faster fulfillment of orders. This efficiency may lead to short-term revenue and profit growth because the company can ship more products in less time.

On the other hand, as delivery times accelerate, Nvidia will switch from shipping based on backlog and new orders to only shipping based on new orders, which may lead to a decrease in revenue.

According to reports, TSMC said it will expand CoWoS production capacity by the end of 2023 and double its production capacity by the end of 2024 from mid-2023 levels. Perhaps TSMC or its partners' CoWoS capacity expansion is progressing faster than expected, resulting in shorter H100 delivery times. Meanwhile, Nvidia is now unable to ship fully functional H100s to China due to U.S. restrictions, which has reportedly led to a drop in demand there, but it's unclear whether this situation has helped shrink H100 delivery times.

Another good news for Nvidia is that the company will bring a new generation of B100 flagship this year, which will also give the company the confidence to resist competition from more potential competitors.

According to relevant reports, B100 will significantly improve the reasoning capabilities of the GPT-3 model, which has 178 billion parameters. This improvement could be double or more than what the existing H200 architecture already offers. However, to be clear, while the graph clearly highlights large language model (LLM) performance, this does not necessarily translate into the same level of raw compute power unless core and power efficiency improve significantly, which seems to be the case It's unlikely to be possible on that scale.

Additionally, NVIDIA confirmed that the B100 will see higher growth in memory bandwidth. The current version of Hopper GPUs already feature HBM3e technology, providing bandwidth up to 4.8 TB/s. Upcoming Blackwell chips are expected to continue using the technology, possibly adding stacks and increasing bandwidth. NVIDIA emphasizes its commitment to designing architectures with future memory technology support in mind, as evidenced by the integration of HBM3e with the GH200 GPU last year.

Despite the good news, Nvidia cannot avoid challenges.

First, Nvidia Chief Financial Officer Colette Kress warned at the earnings briefing that data center revenue from China fell "significantly" in the fourth quarter due to U.S. licensing requirements.

Meanwhile, the company's main rival AMD is investing heavily in its own AI chips, including its new MI300X. According to AMD, the chip outperforms Nvidia's H100 on certain workloads, but Nvidia refuted those claims, saying AMD's testing was incomplete.

Additionally, Intel has its own server chips and continues to enhance its capabilities.

Furthermore, chip companies that already exist or may exist, such as Sam Altman’s seven trillion chip plan, Masayoshi Son’s $100 billion challenge to NVIDIA, Jim Keller’s eagerness to bring chip products, and the recently popular Groq LPU, are working hard to Want to unseat Nvidia. This will undoubtedly bring them a threat.

Furthermore, Nvidia's customers are increasingly seeking to develop their own dedicated artificial intelligence chips to reduce dependence on Nvidia products. Amazon, Google, Meta, Microsoft, Tesla and several Chinese companies are among the well-known companies that currently offer or are actively developing their own in-house AI chips.

That could be particularly problematic because the company said more than half of its data center revenue in the fourth quarter came from large cloud providers for both internal workloads and external customers. But Nvidia isn't taking this threat lying down. According to Reuters, the company has met with Alphabet, Amazon, Meta, Microsoft and OpenAI to discuss building custom chips for them.

Because of all this good news and bad news, everyone is extremely interested in how far this chip giant can rise. For example, in the stock market, when do you think they will break through two trillion US dollars?

Reference link

https://www.forbes.com/sites/dereksaul/2024/02/21/nvidia-earnings-mark-another-historic-quarter-for-ai-titan/?sh=32a00f7e29c4

https://nvidianews.nvidia.com/news/nvidia-announces-financial-results-for-fourth-quarter-and-fiscal-2024

https://www.tomshardware.com/tech-industry/artificial-intelligence/wait-times-for-nvidias-ai-gpus-eases-to-three-to-four-months-suggesting-peak-in-near- term-growth-the-wait-list-for-an-h100-was-previously-eleven-months-ubs

https://videocardz.com/newz/nvidia-teases-next-gen-b100-blackwell-gpu-performance-in-gpt-3-175b-large-language-model

Click here to follow and lock in more original content

END

*Disclaimer: This article is original by the author. The content of the article is the personal opinion of the author. The reprinting by Semiconductor Industry Watch is only to convey a different point of view. It does not mean that Semiconductor Industry Watch agrees or supports the view. If you have any objections, please contact Semiconductor Industry Watch.

Today is the 3683rd issue of "Semiconductor Industry Observation" shared with you. Welcome to pay attention.