Amazing Nvidia Has Fans & Critics In Awe

International | May 27 2024

Nvidia didn't save the markets, but it did silence the naysayers, for now. What may possibly lay ahead?

-First quarter earnings hitting it out of the park (again!)
-Will the good times keep rolling?
-Analysts scrambling to raise forecasts and price targets
-'Momentum machine' could well become the world's second largest next year

By Danielle Ecuyer

First quarter 2025 results a standout, again!

“Their chips are a critical “cog in the wheel” in the advancement of AI across a broad range of industries. Other companies are nipping at its heels, but Nvidia is like a tornado tearing through the town – it’s going to [be] hard for even the tech giants to catch up.” {Cameron McCormack, VanEck Portfolio Manager}

Dan Ives, Wedbush Securities Managing Director, has been beating the Jensen Huang fan club drum and to his credit, the Godfather of AI as described by Ives has delivered again with the 1Q25 earnings.

Ives described the results as “a Masterpiece quarter”.

Hatem Dhiab from Gerber Kawasaki (NYSE:GK) said “One of the best earnings reports I’ve seen. I’ve said that last quarter too. They just keep astonishing”.

Nvidia reported 1Q25 revenue growth of 262% year-on-year (yoy) and 18% sequential quarter-on-quarter (qoq) growth, with data centre revenues expanding at a staggering 427% yoy and 23% qoq.

Morgan Stanley called it the “shock and awe” from the upside quarter this time a year ago when the company guided to US$10bn in revenue against the US$28bn guided this time around for 2Q25.

Alex Pollak, Chief Investment Officer of the top performing Loftus Peak Global Disruption Fund ((LPGD)) highlighted the US$2bn beat on revenue forecasts to US$26bn this quarter equaled the entire annual revenues for the company in FY22.

Pollak is on the record as a supporter and GenAi investor, but the dizzying heights of the Nvidia share price appreciation, over 200% in the last 12-months, have led the fund to trim exposure by over -50%.

Turning to the latest results, Pollak observes the scale of the growth in earnings, and highlights “what is the quality of the visibility (of earnings) from June 2025 to January 2026?” 

He explains investors will have a much better understanding of tangible (earnings) evidence towards the end of FY25 and into FY26 and is reasonably confident Nvidia can generate mid-US$30 per share EPS in FY26. Pollak highlighted the stock is trading on a circa 40x prospective multiple, which seems expensive for a hardware company.

Morgan Stanley has also adopted a conservative approach to earnings and is confident Nvidia can grow EPS to a comparable US$32 EPS in FY26.

UBS by comparison is more bullish on the earnings outlook, forecasting FY26 EPS of just over US$40 per share.

Earnings growth, where to now?

“We’re producing something that most people at the moment don’t understand. There will be new factories created, and we’re going to produce intelligence at scale.” [CEO Jensen Huang April 2024]

The real questions for many investors include 'can Nvidia justify the valuation' and 'how much more runway is there in the hardware GenAi investment cycle?’

This is where the rubber hits the road and views start to bifurcate around demand resilience in the face of new product introductions; the success of Nvidia customers in monetising the hardware, and decreasing the company’s reliance on Big Tech (hyperscalers) customers.

Nvidia CEO Jensen Huang stuck with the bullish narrative and allayed concerns around demand factors in FY25 and into FY26.

On the earnings call, Huang set the scene by stating, “The next Industrial Revolution has begun. Companies and countries are partnering with Nvidia to shift to the trillion-dollar installed base of traditional data centres to accelerated computing and build a new type of data centre, Ai factories to produce a new commodity, artificial intelligence.”

The transformation enables productivity gains to almost every industry from cost and energy efficiency savings while expanding revenue opportunities.

 “The world of computing will change from information retrieval to intention seeking and problem solving”, Dhiab explained.

Huang dismissed analyst and critic concerns around the transition from the Hopper GPU to Blackwell, with no evidence of Hopper purchases being passed over for Blackwell, which is expected to contribute in 4Q25. Morgan Stanley’s research checks also confirmed demand remains strong through the transition.

While Pollak highlighted question marks over the potential monetisation of the Ai infrastructure investments, Michael Frazis from Frazis Capital Partners pointed to Huang’s commentary on the financial multiplier effects for Nvidia customers. 

Huang articulated every US$1 of spend on a GPU returns US$5 in earnings to the hyperscalers over a four-year period, equating to a one-year payback. Every US$1 spend on the H100 (Hopped) servers equates to US$7 in revenue over a 4-year period. 

On both counts Frazis confirmed his own research pointed to the same conclusion.

Michael Dell of Dell Technologies recently told Reuters the Dell-Nvidia based servers were the fastest ramping product in history.

UBS and Morgan Stanley were encouraged by the broadening of the data centre customer base as the hyperscaler share declined to circa 40% from over 50% of total revenue, offset by Consumer Tech (the likes of META), Enterprise (Tesla for full-self driving) and a faster ramp up in sovereign states, than previously communicated.

Sovereign is now expected to contribute high single digit billions of dollars in revenue for the current financial year, ahead of previous expectations of 2025, with opportunities cited by management in Japan, France, Italy, and Singapore. China and the Middle East remain dormant due to export controls.

Morgan Stanley also sees the macro backdrop as shifting to more potential upside spending from the hyperscalers as they announce higher 2025 investment plans.

The transition from artificial intelligence training to inference is crucial to the evolution of the technology and the monetisation of the chip investment and Ai factory build out.

Hunter Wolf Research emphasised “inference performance will be key to the future success of the AI world.”

But what is inference? 

To understand the jargon, I applied Microsoft Co-pilot to answer what are “inference tokens from Hopper (Nvidia GPUs)?” which Huang discussed in the conference call.

“Inference is a term used in AI to describe the process of using a trained model to make predications. Nvidia’s Hopper architecture, which includes the H100 and H200 Tensor Care GPUs is designed to optimise this process. The term ‘inference tokens’ likely refers to the units of computation used by these GPUs to perform inference tasks.”

UBS summarised the inference tokens from Hopper are driving significant Ai service/software revenue.

Both Hunter Wolf Research and Beth Kindig of the IO Fund subscribe to the view Nvidia offers key competitive advantages from its CUDA algorithm, the software layer that assists engineers to accelerate applications and leverage the GPU’s power.

CUDA is considered the secret sauce or moat against competition. It secures a natural barrier as applications built on it make it harder for customers to switch to another platform.

Kindig, a long-term advocate for investing in Nvidia, remains firmly of the belief Nvidia will one day be the most highly valued software stack company in the world via the transition to the “edge” and away from data centers. Kindig sees upside growth over 2025/26 from taking GenAi computing to the data sources (edge); smart devices even autonomous vehicles which Tesla and Chinese EV makers are working on.

Momentum driven machine 

NYU’s Aswath Damodaran, also known as the Dean of Valuation, commented on CNBC, “It is testimonial to the power of momentum, everything they touch turns to gold”

Nvidia has taken up the momentum mantle where other momentum trades left off, think Meta, Google, Tesla in the past.

US$400bn in market cap was added post the March quarter result, which might have been boosted by the announced 10-for-1 stock split. The split has no bearing on the financial outlook, rather a tilt of the goodwill hat to retail investors to increase share purchase availability.

With a US$2.6trn market cap at US$1065 per share, Nvidia could reach US$3trn (15% upside) according to Ives in 2025, thereby superseding Apple at its current valuation and making it the second largest listed US company, after Microsoft.

When it comes to price targets post earnings, analysts implemented an across-the-board upgrade cycle.

Morgan Stanley sits at the more conservative end at US$1160 with Goldmans Sachs and UBS at US$1200, stretching to Jeffries at US$1350 and Cantor Fitzgerald at US$1400, 30% above the current share price.

It is Kindig who remains by far the most optimistic, with the caveat of extraneous factors of macro headwinds and the like. Under Huang's stewardship, Kindig believes Nvidia has the potential to attain the dizzy heights of a US$10trn valuation in the medium term.

In the shorter term, the debate centres around how much has been discounted in the price already, with the rolling out of generative Ai hardware and software determining both the success of applications and of monetising the billions of dollars of investment.

Note: Gerber Kawasaki manages the Adviser Shares Gerber Kawasaki ETF (NYSE:GK).

The author owns Nvidia shares.

More on Nvidia and investing in artificial intelligence and ASX-listed companies with exposure: 

-Generative AI, Investing In The 21st Century Megatrend, Part One:

-Generative AI, Investing In The 21st Century Megatrend, Part Two:

-Generative AI, Investing In The 21st Century Megatrend, Part Three:

Find out why FNArena subscribers like the service so much: "Your Feedback (Thank You)" – Warning this story contains unashamedly positive feedback on the service provided.

FNArena is proud about its track record and past achievements: Ten Years On

Share on FacebookTweet about this on TwitterShare on LinkedIn

Click to view our Glossary of Financial Terms