Nvidia: Stand Back In Awe – 6 Recent Insights To Help Clarify The Medium Term

Summary:

• Nvidia Corporation’s CEO Jensen Huang has confirmed strong medium- to long-term revenue growth, with a potential to double or triple current revenues, reinforcing Nvidia’s market leadership.
• Nvidia’s strategic partnerships in India and other countries highlight its expanding AI infrastructure, driving sustained demand for its chips globally.
• TSM’s commitment to doubling chip-production capacity by 2025 and Foxconn’s new factory in Mexico ensure robust supply to meet Nvidia’s high demand.
• Nvidia’s continuous innovation in algorithms enhances existing chip performance, maintaining partial value and driving ongoing software subscription revenue.

In my very first article on Seeking Alpha a couple of months ago, I recommended Nvidia Corporation (NASDAQ:NVDA) in Nvidia’s Market, Market Share, Margins and Multiples. I wanted to help investors think big about Nvidia’s potential growth, at least into the medium term.

In this article, I want to review six developments and tidbits that have creeped out into the public sphere that investors might have missed. I will demonstrate that sales are expected to be strong for years to come—not just next quarter. Finally, I will discuss certain risks and explain my reconfirmed Strong Buy rating.

1. CEO Jensen Huang’s Recent Admission

In an October talk with hedge fund manager Brad Gerstner of Altimeter Capital, Nvidia’s bottom threshold for medium-term growth prospects was made astoundingly clear, compared to everything else I’ve seen and read on the matter. In the course of the discussion, Brad asked CEO Jensen a key question.

Brad Gerstner: As you sit here today, is there any reason… if you’re only $125 billion out of a multi-trillion-dollar TAM, that you’re not going to have 2x the revenue, 3x the revenue, in the future, that you have today?

Jensen Huang: No.

Jensen goes on to explain that market makers like Nvidia can do well and become “quite large” if they do invent a new market or create a new pond for the fish, while fish remaining in the smaller pond are restricted to a maximum size limited by the smaller pond.

I view this segment of the interview as critical for any investors or detractors of Nvidia’s prospects to understand. Jensen has been more reserved in his forecasts on earnings calls, discussing potential next quarter and into next year, but not often so blatantly solidifying the medium-term and even long-term revenue outlook for the company. Perhaps this was due to his remarks not being prepared beforehand. Thankfully, he exposed his real views on revenue potential in the coming several years.

2. Nvidia Partners with India

Nvidia’s recent partnership in India demonstrates the company’s prospects for building AI infrastructure outside the US. This is a trend that investors can expect will continue as the company’s chips reach second- and third-tier companies and countries outside the Magnificent 7 and other large US companies.

Indian companies such as Reliance Industries (RLNIY), Tata Communications, Infosys (INFY), TCS, Tech Mahindra, Wipro (WIT), and others are signing up to be part of the effort to bring AI data centers to India. The goal is for smaller players without the required capital for their data centers can pay-as-you-go for AI from the servers, in the form of cloud computing, from the large players’ AI models, like Reliance Jio.

India is just one of over a hundred countries in which large companies will be looking to develop their AI infrastructure. With several examples already, investors should factor this secondary demand into their forecasts.

3. Blackwell Sold out for 2025

In case you missed it, Nvidia’s Blackwell chip is officially sold out for 2025 (Nvidia’s fiscal 2026). While this was largely discussed and digested a few weeks ago, it is worth exploring a bit further for a few reasons.

As many investors know, Nvidia’s sales have not been dependent on demand. They have been dependent on actual chipmakers’ fabrication output and yield maximums. In fact, gaming chips have had to take a back seat at TSMC (TSM) so that more H100s and Blackwell series chips could be created (at higher sale prices). This was warned about as far back as June, with this development confirmed on the most recent earnings call.

Further, TSM has pledged to increase capacity at massive volumes both this year and next year. What that means is that 1) current capacity is sold out; potential capacity is not. 2) As such, 2025 will see increases in sales beyond the sold-out level as further capacity at TSM and its other fabricators increases.

4. Nvidia Asks SK Hynix to Speed Up Memory Chip Production

In early November, Jensen reportedly asked SK Hynix Chair Chey Tae-won if it was possible to move its production (supply) of high bandwidth memory (HBM) chips forward by six months, presumably to help Nvidia meet its “insane” demand for AI chips.

On the technical side, I’ll quote from press coverage of the announcement.

HBM is a type of dynamic random access memory, known as DRAM, where chips are vertically stacked to save space and reduce power consumption. Adding more layers to a HBM will, in theory, give it more capacity to handle complex AI applications.

Keen investors will realize that all inputs to Nvidia’s chips, coming from more than a dozen companies, need to be fast-tracked to meet the monstrous demand.

Requests for parts supply ramping help indicate Nvidia’s expectations for sales in the future. Keen investors will watch for announcements from Nvidia, or its suppliers, on supply ramping efforts.

5. TSM Pledges to Double Chip-Production Capacity in 2025

Similar to the above, and as some investors might know, TSM’s pledge to double chip-production capacity in 2025 is another, and perhaps among the strongest, sign of health for the demand for Nvidia’s chips from customers. While this capacity is expected to serve other clients, such as Microsoft and Amazon, Nvidia is reported to occupy more than 50% of this new capacity.

Given Nvidia’s main chipmaker, TSM is spending billions (described below) of its capital to serve the increased demand for just a handful of customers, including Nvidia, this vote of confidence demonstrates a strong belief in continued demand not just for one year—which would be a waste of a factory—but for several.

6. Nvidia Invents New Algorithms, Increasing Old Infrastructure Efficiency

The common paradigm thinking is that once a chip is sold to a customer, that customer will need to buy new chips for better performance. In the same interview as the one cited above, Jensen clarified that Nvidia chips operate differently.

We also put a lot of energy into continuously re-inventing new algorithms, so that when the time comes, the Hopper architecture is 2, 3, 4 times better than when they bought it so that yield, that infrastructure continues to be effective.

This implies that the H100 series of Nvidia chips, despite now not being the best Nvidia has to sell, will retain its ownership. Not only will they be compatible with the new Blackwell chips, but they will also be improving in performance.

With retained ownership, the software subscriptions can continue. Another potential is that if the H100 series are updated with new algorithms, that would make them valuable. Some big tech companies may wish to sell them to smaller outfits and use the cash to buy more of the latest Blackwell or Rubin chips. This puts a hole in the “return on AI” or ROAI debate that has gone on for too long in the face of big tech companies literally attributing certain parts of their growth to AI. These include Meta Platforms (META) with increased user engagement via stronger algorithms, and Microsoft (MSFT) with greater Azure use.

6. Foxconn Building AI Factory in Mexico Specifically for Nvidia’s GB200 Series

For investors worried about chip fabrication concentration, you can now worry less. Foxconn announced recently that it would be building a factory in Mexico just for one level of Nvidia’s array of chips—the top-level GB200 series. This is the chip the hyperscalers have been buying. The implication is the bulk of Nvidia’s revenues will now have a production pipeline that could go unaffected during a potential China-Taiwan conflict, at least compared to being more dependent on TSM in Taiwan.

To add to this, TSM itself is in the process of building a $65B chip factory in Phoenix, Arizona, as part of its attempt to increase AI chip supply for Nvidia. Despite the factory facing initial hurdles, TSM has had recent wins, with yield production surpassing Taiwan’s in an October 24 announcement.

Valuation

When I rated Nvidia a Strong Buy in late September, the company was trading at $116.52. Since then, the next quarter’s results have come out. Revenues increased 16.78%. The current share price is hovering around $137, a 17.58% increase. Thus, little has changed in valuation since my first report on Nvidia. The Blackwell “masking” issue has been resolved, and overheating of installed chips has been explained as par for the course in chip installation, something engineers are on site to adjust.

One could view the recent quarter’s revenues of $35B as $35B in chip demand (and gaming, automotive, and other segments) already met, hence less demand forward. I choose to view it as $35B of the $2T in chip replacement and growth expected over the next four to six years, as Jensen has repeatedly stated. As such, there is still plenty of demand left for Nvidia to fill, especially with GPU total addressable market, or TAM, growth in the 20% CAGR range, according to Jensen.

Based on my proprietary Five Factor Model and a key factor update, which I hope to apply in a review for Nvidia in the future, the company trades at a very reasonable 4.8% forward (R&D+EBIT)/EV. This is well within an acceptable range given its outrageous growth rates, which appear on track. For those unaware, this figure entails that if Nvidia cut all research and development spending, and instead paid it and its operating income out as a 100% payout ratio dividend, that forward yield would be 4.8% on a pre-tax, pre-adjustment basis.

Conclusion

Steady as she goes. Nvidia is still in the early quarters of a years-long phenomenon. It is driven by hundreds or thousands of companies, countries, universities, militaries, and billions of citizens, finally gaining access to chips or models powerful enough to do the calculations required for AI and related large computational use cases. Until we start to develop a clear picture as to where this demand levels off, there is little sense exiting Nvidia. It remains the investor’s job to ride valuations of any type, even to an outrageous level, until such demand has been met.

My article focused on a few key developments and tidbits that help demonstrate that the overall demand trend is still incredibly strong. This is evidenced by new factories being purpose-built for GB200 series chips, overseas partnerships, requested output increases for suppliers, and other indicators cited above.

With CEO Jensen Huang literally telling us there is no way revenues won’t double or triple in the coming years, it’s difficult to imagine a clearer signal without Jensen giving specific figures and dates, which of course could put him in trouble. As such, I rate Nvidia a Strong Buy. Investors should know that my Strong Buy ratings indicate I own shares in the company I’m covering and that I don’t give Buy ratings. For Nvidia, the company remains my largest position well into the double-digit percentages, a concentration I feel equally comfortable with compared to my other holdings.

---

Analyst’s Disclosure: I/we have a beneficial long position in the shares of NVDA either through stock ownership, options, or other derivatives. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.

Seeking Alpha’s Disclosure: Past performance is no guarantee of future results. No recommendation or advice is being given as to whether any investment is suitable for a particular investor. Any views or opinions expressed above may not reflect those of Seeking Alpha as a whole. Seeking Alpha is not a licensed securities dealer, broker or US investment adviser or investment bank. Our analysts are third party authors that include both professional investors and individual investors who may not be licensed or certified by any institute or regulatory body.

---