Nvidia's (NVDA) market cap, now at $2.8 trillion, continues to grow as it gradually closes in on Apple (AAPL), which sits at $2.9 trillion. The flip would make the chipmaker the second-largest publicly traded company behind Microsoft (MSFT).

The stock's rise was quick, as it cashed in on growing demand for hardware that powers data centers.

"The company made a massive switch in its consumer mix," said Sandeep Rao, a senior researcher at Leverage Shares, an investment management company. "It basically went from being the darling of crypto miners and gamers to being the darling of corporates who were building out data centers and AI-driven computing; 87% of its revenue comes purely from data centers."

He added that Nvidia is a great company, but that still doesn't justify a trailing 12-month price-to-earnings ratio of 67 times when the industry average is drastically lowerNagpur Investment. Semiconductor manufacturers, on average, have a P/E near 34.

Like many of the blue-chip mega-cap companies, Rao believes the stock price is quite divorced from the company's fundamentals despite its solid performance. Nvidia's first-quarter earnings beat expectations, with quarterly revenue of $26 billion, up 18% from the previous quarter and up 262% from a year earlier. Its net income rose by a whopping 462% from last year.

"Over the course of the year, you could expect a 20% to 30% correction wherein the trailing 12-month P/E should be down to around 50," Rao said. "Which is, again, still a premium over the rest of the industry. And that is basically because Nvidia is strongly holding on to a very loyal customer base until something else comes along with it."

If investors want to understand Nvidia's future prospects, Rao suggests looking past the current numbers and paying attention to industry trends.

Investors are looking at peak performance and extrapolating that into the future. But even if Nvidia manages to fend off competitors and keep most of its market share, demand from its current customers will soon slow as data centers complete the initial phase of their construction.

"If you spent $10 billion building out a data center, you are not going to spend another $20 billion upgrading it the next year," Rao said. "You're going to rationalize the cost of the $10 billion you've spent over the course of, let's say, six or seven yearsKolkata Stocks. Then they'll optimize performance through various software-driven methods rather than hardware-driven methods."

Rao expects Nvidia to continue to see relatively strong revenue growth but at a slowing pace. Growth will gradually go from quarterly increases to annual ones or over two years as they pick up repeat business, he noted.

As for competition, even if Nvidia's contenders appear weak, more entrants will grab market share. As Nvidia prepares to release the next phase of its AI chips, Blackwell and Rubin, AMD'sRadeon could bring better costs to consumers. Beth Kindig, the CEO and lead tech analyst for the I/O Fund, estimates that AMD's market share could reach 20%.

Nvidia and AMD have similar production costs since neither makes their GPUs but uses the same supplier, Taiwan Semiconductor (TSMC). The proprietary element of their chips is in the design architecture, and each prioritizes different performance efficiencies such as data or graphics, Rao said. Since Nvidia has a larger customer base, it's at an advantage because it allows them to clarify which efficiencies to optimize based on demand. AMD isn't there yet, he noted.

Nvidia's 12-month earnings per share is near 17.12, while AMD is at a mere 0.69 because it's still burning through cash, Rao said. But based on Nvidia's evolution of its product improvements over a one-year timeline, AMD is on a similar track and is likely two or three quarters away from producing GPUs that are closer in performance to Nvidia's Blackwell, he added.Bangalore Wealth Management

"So AMD is a little bit behind in terms of performance," Rao said. "But really, Nvidia success lies in its sales teams being able to market their products more effectively than AMD. That's about it."

Other major competitors include Intel, which recently revealed its Xeon 6 data center processors, which are set to compete with Nvidia's chips.

On the longer horizon, Rao points to threats from startups such as Graphcore and SambaNova System that have been working on prototypes that could beat Nvidia's performance in terms of computing speed.

Once a data center is locked into its hardware setup, it will probably be in place for seven to 10 years, Rao noted. But it's not unusual for centers to integrate chips from other makers as they expand or upgrade in the future. But it would happen gradually, which means Nvidia could counter the threat by releasing even more revolutionary products, he added.

Finally, investors must watch out for geopolitical risks, particularly the tensions between the US and China. The US administration has bipartisan support in restricting the export of AI-relevant technology to China, where 17% of Nvidia's customer base is, he said. On the flip side, Chinese officials are pushing tech giants like Alibaba, Baidu, ByteDance, and Tencent to use domestically-made AI chips to replace Nvidia's.