Share prices of Nvidia (NVDA -1.99%) hit a record high last week, driven by updated reports showing a booming demand for its graphics cards, which are being sucked up by companies in the artificial intelligence (AI) arms race.
The semiconductor stock is up a massive 232% in 2023 so far, which also explains why it is trading at an expensive valuation. With a price-to-sales ratio of 37 and trailing earnings multiple of 117, there is no doubt that Nvidia is valued at a huge premium to the S&P 500's sales multiple of 2.5 and price-to-earnings ratio of 20.
But despite these rich multiples, certain Wall Street analysts still consider Nvidia stock to be a bargain. Let's see why that's the case.
Nvidia's potential earnings growth makes the stock look cheap
Independent investment research and analytics firm Melius Research recently wrote in a note to investors that Nvidia stock is trading at just 28 times its earnings estimates for 2024. Melius analyst Ben Reitzes points out that Nvidia's forward earnings multiple is cheaper than those of other potential AI winners such as Amazon, Adobe, and Microsoft.
Melius' argument seems to be on point if we take a look at the following valuation chart based on consensus estimates.
NVDA PE Ratio (Forward 1y) data by YCharts
Nvidia is indeed cheaper than the likes of Amazon and Adobe on a forward earnings basis, and it is slightly more expensive than Microsoft. Given that Nvidia is expected to deliver way bigger earnings growth than Adobe and Microsoft, while also outperforming Amazon in the next fiscal year, it is easy to see why Reitzes calls the semiconductor specialist's stock cheap.
| Current FY | Next Year FY | 2-Year Ahead FY | |
|---|---|---|---|
| Nvidia EPS estimate | $10.82 | $16.01 | $19.92 |
| Nvidia EPS projected growth (YOY) | 224% | 48% | 24% |
| Amazon EPS estimate | $2.16 | $3 | $4.28 |
| Amazon EPS projected growth (YOY) | NA | 39% | 43% |
| Microsoft EPS estimate | $11.01 | $12.67 | $14.66 |
| Microsoft EPS projected growth (YOY) | 12% | 15% | 16% |
| Adobe EPS estimate | $15.73 | $17.75 | $20.51 |
| Adobe EPS projected growth (YOY) | 15% | 13% | 16% |
Source: YCharts and company earnings reports. YOY = Year over year. FY=Fiscal Year. Note: Amazon reported a loss of $0.27 per share in the previous fiscal year, so the year-over-year comparison isn't available.
At the same time, Melius Research adds that Nvidia's earnings growth estimates are likely "conservative," suggesting that the company's bottom line could increase at a much faster pace than what consensus estimates suggest. It is not surprising to see that that could indeed be the case.
Two simple reasons why Nvidia can deliver stronger earnings growth
It is well known that AI has been central to Nvidia's terrific growth of late. The company's data center revenue surged 171% year over year in the second quarter of fiscal 2024 to $10.3 billion and produced 76% of its top line. As a result, Nvidia's overall revenue jumped a whopping 101% over the prior-year period to $13.5 billion. More importantly, Nvidia's net income increased a whopping 422% year over year to $6.74 billion last quarter. Its non-GAAP earnings increased more than five-fold to $2.70 per share.
The first reason why Nvidia's earnings grew so strongly is, of course, because of its healthy revenue jump. The company is shipping data center graphics cards in large volumes, which is not surprising, as numerous companies are lining up to buy its chips to train AI models. According to the Financial Times (via Tom's Hardware), Nvidia could ship 550,000 of its H100 data center graphics processing units (GPUs) for AI workloads in 2023.
Given that the cheapest version of this GPU is sold for around $30,000, Nvidia could generate at least $16.5 billion in revenue from sales of the H100 processor. The number could be much higher than that, as the H100 GPU is also available in more powerful configurations, and the chip is as expensive as $70,000 in markets such as China.
Looking ahead, Nvidia could see a major bump in its AI-specific GPU shipments next year. That's because its foundry partner Taiwan Semiconductor Manufacturing (popularly known as TSMC), is reportedly going to substantially ramp up its capacity to manufacture advanced AI chips. More specifically, TSMC could increase its chip-on-wafer-on-substrate (CoWoS) packaging capacity to 16,000 wafers a month by the end of 2024, compared to an estimated 8,000 a month at present.
It is estimated that Nvidia could get access to two-thirds of that supply, which points toward a substantial increase in the number of data center GPUs that it may be able to manufacture. So Nvidia could continue to witness stronger sales volume of its data center graphics cards going forward.
This brings us to the second reason why Nvidia's earnings could increase at a faster-than-expected pace: pricing power. We have already seen that the company's latest H100 processors command a price of more than $30,000 depending on the configuration. It is anticipated that the successor to the H100 could be priced at a 40% premium, according to UBS analyst Timothy Arcuri.
So there is further room for margin expansion at Nvidia, which should translate into faster earnings growth for the company. All this indicates that investors who haven't bought this AI stock yet can still consider doing so, as the stock's outstanding rally seems here to stay.
