Investors have been buying shares of Nvidia (NVDA -1.99%) hand over fist in 2023 thanks to its artificial intelligence (AI)-fueled growth. This is not surprising, since the chipmaker's revenue and earnings are getting a serious boost from the booming demand for graphics processing units (GPUs), which are needed for training AI models and for inferencing.
From reporting a 13% year-over-year decline in revenue to $7.2 billion in the previous quarter, Nvidia is now on track to deliver $11 billion in revenue in the current quarter. That would translate into a 64% jump over the prior year. The biggest reason why Nvidia is anticipating such outstanding growth despite the woeful situation in the personal computer (PC) market -- which has wrecked the company's gaming business -- is because of its dominant position in the data center GPU market.
Nvidia reportedly controls 95% of the machine learning GPU market, according to New Street Research. As these graphics cards reportedly start from $10,000 and could go up to more than $40,000, it is not surprising to see why the company's growth is about to accelerate big time. Its revenue is expected to increase close to 60% in fiscal year 2024, while earnings are estimated to jump 133%.
However, investors will have to pay a rich premium to buy Nvidia's potential growth. It is trading at a whopping 41 times sales and sports a forward price-to-earnings (P/E) ratio of 57. While that premium may seem justified considering Nvidia's terrific grip over the AI GPU market, there may be a cheaper alternative for investors looking to capitalize on this lucrative market -- Advanced Micro Devices (AMD 1.14%). Let's see why this Nvidia rival could take advantage of the AI market despite being late to the game.
The market for AI chips is big enough to supercharge AMD
According to AMD CEO Lisa Su, AI is "the largest strategic growth opportunity for AMD." She estimates that the market for data center accelerators such as GPUs, data processing units (DPUs), and other types of processors could jump from $30 billion in 2022 to $150 billion in 2027.
Nvidia is dominating this massive market, but there is space for another player, such as AMD. This is evident from reports that the waiting time for Nvidia's data center GPUs ranges between three and six months. The company is reportedly unable to keep up with the huge demand for its GPUs, which are deployed for training AI models. It is scrambling to place more orders with its foundry partner to ramp up output and satisfy customer demand.
However, the shortage of Nvidia's data center GPUs has led to sharp price hikes, as these chips are reportedly commanding a 40% premium over the MSRP (manufacturer's suggested retail price). AMD can fill this void and capitalize on the multibillion-dollar AI chip market. The company is on track to launch its MI300 family of data center GPUs for AI training and inference by the end of 2023.
These chips are already attracting customer interest, which isn't surprising as AMD is going all out to capture the AI chip opportunity.
For instance, AMD recently revealed the Instinct MI300X data center accelerator, which it claims is the "world's most advanced accelerator for generative AI." With 192 gigabytes (GB) of high-bandwidth memory (HBM), it is not surprising to see why AMD is making such a claim, as this spec exceeds Nvidia's H100 data center GPU's 120GB memory capacity.
AMD management points out that more memory will give data centers the "compute and memory efficiency needed for large language model training and inference for generative AI workloads." The chipmaker intends to start sampling these chips with customers in the third quarter. So AMD's growth could start accelerating once again from the end of 2023, assuming its AI-focused GPUs are a hit with customers.
We have already seen how fast the AI accelerator market is expected to grow and the amount of revenue it is expected to generate. Success in this market will be critical for a turnaround in AMD's fortunes, and it looks like Wall Street is expecting the company's move in this space to pay off.
Investors can expect better times after a woeful 2023
According to consensus estimates, AMD's revenue in 2023 could fall 2.5% to $23 billion. Its earnings are expected to drop to $2.86 per share from $3.50 per share in 2022. However, the next two years are looking relatively better for AMD as per estimates, with the company expected to deliver solid double-digit growth.
AMD EPS Estimates for Next Fiscal Year data by YCharts
However, AMD may be able to outpace Wall Street's expectations if it can hit gold in data center accelerators. Nvidia has already used this opportunity to execute a major turnaround, and AMD could replicate its bigger rival's feat. That's why investors looking to buy a potential AI winner trading at a cheaper multiple should take a closer look at AMD.
The company sports a price-to-sales ratio of 8, which is significantly lower than Nvidia's, while AMD's forward P/E ratio of 39 indicates that investors are getting a good deal on the stock right now given the massive AI opportunity that lies ahead.
