Nvidia dominated 2024 to a large extent. Next year? Lots of challenges.

Nvidia (NVDA) has had the kind of year most companies can only dream of.

Its revenue and stock price soared thanks to predictable investments in artificial intelligence technologies that are paying off from the generative AI wave.

That’s not all. Changed places repeatedly with Apple (AAPL) as the world’s largest publicly traded company by market capitalization, exceeding $3 trillion. CEO Jensen Huang has become one of Silicon Valley’s most sought-after executives, meeting with everyone from fellow tech luminaries to world leaders and then some.

And there’s more to come. The company is ramping up production of its high-powered Blackwell chip for AI applications and expects to ship several billion dollars worth of hardware in the fourth quarter alone, with much more expected in the next year

“Nvidia really has the (hardware and software) for the era of AI computing,” Futurum Group CEO Daniel Newman told Yahoo Finance. “Everything is connected in the (server) rack, out of the (server) rack, and then the software is really good … it’s liked by the developer communities.”

But the competition does not stand idly by.

Companies like AMD (AMD) are looking to poach Nvidia’s customers and cut into its estimated 80% to 90% market share. Even Nvidia’s own customers are working on chips aimed at reducing their reliance on the graphics giant’s semiconductors.

And Wall Street gets on board.

Broadcom stock (AVGO), which works with companies like Google (GOOG, GOOGLE) to design AI chips, are up 113% year to date and soared 44% in the past month alone after CEO Hock Tan said AI could represent a $60 billion to $90 billion opportunity for the company only in 2027.

Still, taking on Nvidia will be a tough task for any company. And dethroning him as king of AI, at least by 2025, will be next to impossible.

Nvidia gained a first-mover advantage in the AI market from early investments in AI software that unlocked its graphics chips for use as high-powered processors. And it has managed to maintain that leadership in the space thanks to continued advancements in its hardware, as well as its Cuda software that allows developers to build applications for its chips.

For this reason, the so-called hyperscalers, massive cloud computing providers such as Microsoft (MSFT), Alphabet’s Google, Amazon (AMZN), Meta (TARGET), and others continue to pour money into buying as many Nvidia chips as possible. In its most recent quarter, Nvidia reported total revenue of $35.1 billion. Of that, $30.8 billion, or 87%, came from its data center business.

“Everybody wants to build and train these big models, and the most efficient way to do that is with CUDA software and Nvidia hardware,” TECHnalysis Research president and chief analyst Bob O’Donnell told Yahoo Finance .

Top of the stack: CEO Jensen Huang speaks during the Nvidia GTC keynote in San Jose, Calif., Monday, March 18, 2024. (AP Photo/Eric Risberg) · ASSOCIATED PRESS

Nvidia is expected to continue to power most of the AI industry in 2025 as well. The company’s Blackwell chip, the successor to its popular Hopper line of processors needed to power AI applications, is in production, and its customers, such as Amazon, are already adding new cooling capabilities to their data centers to manage the immense heat generated by the processors.

“I don’t know what the current backlog is (for Nvidia chips), but if it’s not a year, it’s close to a year,” O’Donnell said. “So they’re pretty much sold out for most everything they’ll probably do next year.”

With hyperscalers calling for an increase or at least the same level of capital expenditure in 2025 as in 2024, you can expect some of that to end up going towards buying Blackwell chips.

While Nvidia will retain control of the AI crown, there’s no shortage of challengers looking to take its throne. AMD and Intel (INTC) are the main competitors among chipmakers, and both have products on the market. AMD’s MI300X line of chips is designed to take on Nvidia’s H100 Hopper chips, while Intel has its Gaudi 3 processor.

However, AMD is better positioned to steal market share from Nvidia as Intel continues to struggle amid its transformation efforts and look for a new CEO. But even AMD is having a hard time breaking Nvidia’s lead.

“What AMD needs to do is make the software really usable, build the systems where there’s the most demand … with developers, and ultimately that could create more sales,” Newman said. “Because these cloud providers will sell what their customers demand.”

It’s not just AMD and Intel, though. Nvidia customers increasingly develop and drive their own AI chips. Google has its tensor processing unit (TPU) chips based on Broadcom, while Amazon (AMZN) has its Trainium 2 processor and Microsoft (MSFT) has its Maia 100 accelerator.

StockStory aims to help individual investors beat the market.

There are also concerns that the shift to “AI model inference” will reduce the need for high-powered Nvidia chips.

Tech companies develop AI models by training them on large amounts of data, also known as the training process. Training requires incredibly powerful chips and a lot of energy. Inference, or the operationalization of these AI models, requires less resources and energy. As inference becomes a bigger part of AI workloads, it’s thought that companies will move away from the need to buy so many Nvidia chips.

Huang has said he’s ready for it, explaining at various events that Nvidia’s chips are just as good at inferring as they are at training.

Even if Nvidia’s market share slips, it doesn’t necessarily mean its business will be worse than before.

“This is definitely a case of lifting all boats,” Newman said. “So even with much stronger competition, which I think they will certainly have, that doesn’t mean they’re going to fail. That’s people making a bigger pie.”