Blackwell May Be Nvidia's Hottest Ticket in 2025, but This Other Opportunity Could Be Even Bigger in the Long Run
Nvidia is set to launch its next-generation GPU architecture, Blackwell, in 2025, which is expected to significantly enhance capabilities for generative AI applications. Demand for Blackwell is high, with reports of it being sold out for the next year. Additionally, a successor named Rubin may be released as early as 2026, potentially offering even greater advancements. Analysts predict that Nvidia could capture substantial market share in the AI infrastructure sector, which is projected to exceed $1 trillion in investment over the next three years, positioning the company for sustained growth despite rising competition.
Over the last two years, semiconductor designer Nvidia (NVDA -2.25%) has become a household name thanks to its lead position in the artificial intelligence (AI) revolution. The company's bread and butter is a unique piece of hardware called the graphics processing unit (GPU). To use a metaphor, GPUs are essentially the engine powering the AI vehicle overall.
Nvidia's deep roster of GPUs has been the biggest spark for the company's rapid ascent to being one of the world's most valuable businesses. This newfound growth helped Nvidia generate billions in profits and free cash flow on a consistent basis over the last several quarters. The company has been investing relentlessly into enhanced research and development (R&D) initiatives and is on the cusp of its next big opportunity.
Of course, I'm talking about Nvidia's long-awaited next-generation GPU architecture, known as Blackwell.
Let's look at why the Blackwell launch is such an important moment for Nvidia in 2025. I'll also be revealing one other asset that Nvidia has up its sleeve that is currently receiving little coverage. In the long run, this other product could be even bigger than Blackwell.
All eyes are on Blackwell, but...
Blackwell is expected to offer a new host of capabilities and features geared toward training and inferencing data workloads for more sophisticated generative AI applications.
One of the biggest products of the AI movement over the last couple of years is the rise of large language models (LLM). Meta Platforms, Amazon, Alphabet, and Microsoft have all invested or developed proprietary LLMs. Moreover, each of these "Magnificent Seven" members is a customer of Nvidia, heavily relying on the company's GPUs and data center services.
With this in mind, it should come as no surprise that big tech catapulted to the front of the line for Blackwell purchase orders. In fact, demand is so high that Blackwell is reportedly sold out for the next 12 months.
Research analysts from Morgan Stanley and Piper Sandler are modeling sales from Blackwell to fall anywhere between $5 billion to $8 billion in the fourth quarter. At the high end of this forecast, that would represent 45% of Nvidia's entire sales from compute and networking products during its fiscal fourth quarter last year -- and it's coming from just one product.
While it's clear Blackwell is a big deal, Nvidia has something else in store that I think deserves a bit more attention.
Image source: Getty Images.
... something bigger could be in store sooner than later
One of the more in-the-know technology media personalities is Beth Kindig, the CEO of I/O Fund. A few days ago, Kindig posted on X (formerly Twitter) that a Blackwell successor named Rubin "is rumored to be six months ahead of schedule, with a roll out as early as the first half of 2026."
That's right. Nvidia is already working on a successor product to Blackwell. If the demand themes explored above are any indication for what could be in store for Nvidia, I'd say there's a good chance that the launch of the Rubin architecture will be a smash.
What does this mean for Nvidia investors?
Wedbush Securities research analyst Dan Ives recently suggested that investment in AI infrastructure could eclipse $1 trillion in just the next three years. I think it's fair to say that Nvidia is well-positioned to capture incremental market share, thanks to Blackwell and now Rubin, too.
While anything could happen during the manufacturing or quality assurance processes for Rubin, I am encouraged that the timeline is ahead of schedule (for now). The one-two punch of Blackwell and Rubin could help further separate Nvidia from an intense competitive landscape while also reaccelerating top-line growth and profit margins.
Although competition is rising from the likes of Advanced Micro Devices and even from some of Nvidia's own customers, I think Nvidia is laying the necessary groundwork for sustained growth. In the long run, I still see Nvidia as one of the foundations supporting the overall AI narrative, and I'm optimistic due to the company's continued investments in strategic innovation.