Words: 1,215 Time: 6 Minutes
- Nvidia’s good news was already baked in…
- But Jensen doesn’t see any sign of a bubble?!
- New barometer for credit risk… a stodgy database company
Earlier this week I said “Nvidia (NVDA) will handily beat all revenue and earnings estimates”.
And they did.
Analysts had expected AI chip leader to show more than 50% growth in both net income and revenue in its fiscal third quarter (with Microsoft, Amazon, Alphabet and Meta representing 40% of all NVDA sales)
NVDA easily surpassed the 50% expectation – reporting $57.01B in total revenue – good enough for 62% YoY
Profit rose 65% YoY to $31.9B.
Looking forward – they see Q4 revenue of about ~$65B – which is $4B higher than analyst expectations.
Whats not to like? Well I will answer that in a moment…
But first a short fun personal story:
Circa 2016 I met with NVDA’s executive team to discuss acquiring GPUs for a new device we were making (with the intent to scale it across all smartphones).
The device used Computer Vision with an infra-red (depth) sensor to create three dimensional ‘point clouds’ of the terrain. Smartphones did not offer the capability we needed (that was to come some ~7 years later)
There were many use-cases we could see – but initially we were looking at applications for (a) indoor mapping (in the absence of other positioning signals); and (b) virtual scene recreation.
Fast forward approx 10 years and Niantic Spatial is pursuing a similar (global) vision today (think of it as a map that machines can read to guage things like depth etc).
For the purpose of my story – at the time NVDA was the only company which manufactured the GPU’s we needed (which made it challenging to negotiate a good price!)
I also remember thinking their business (and revenue) was very much dependent on gaming.
And it was…
However, their team said to me “we are not just gaming – we’re very much an AI company”.
Our (potential) use case intrigued them – as no other phone manufacturer was doing anything like it.
But I found their language curious… as I didn’t see them as AI – but rather a chip company in what would be a heavily commoditized space.
Their share price was between 70c to 80c and had done nothing since 2008 (where it traded for around $1.00 ~10 years ago.
But as they say – the rest is history – turns out they were an AI company!
NVDA share price enjoyed an ~83% CAGR for the next ~10 years.
Very few company’s have done better.
NVDA is an outstanding business by just about every financial metric. Whether its profitability, operational efficiency or capital allocation – their numbers are unmatched.
However, investors are right to ask questions regarding valuation (circled in orange below):
But it’s the not profound impact AI will have on our every day lives that investors are questioning.
It’s a case of how much do you pay (and risk) to own a piece of this (inevitable) growth?
It was the same question many investors failed to ask circa 1997 to 1999 (I was in that boat!)
When asked about a potential bubble – Jensen Huang (CEO) doesn’t see one.
“From our vantage point, we see something very different. As a reminder, Nvidia is unlike any other accelerator. We excel at every phase of AI from pre-training to post-training to inference”
He is right about where NVDA plays… but Huang didn’t talk to valuations.
He talked about use cases and demand.
And in fairness – Huang is a product guy – he’s not an equity analyst!
But if Huang alluded to the possibility of (AI chip) malinvestment – imagine the carnage to his share price given:
- EV/EBIT ratio is ~48x;
- P/FCF ratio is ~64x; and
- Fwd PE of 51.5x
Pending your appetite for risk (yours is likely different to mine) – these are high multiples.
And the higher you go – the greater the risk.
That’s not to say taking high risks don’t work – they can.
But you should at least be aware of the risk you’re taking.
For example, for these valuation to makes sense, NVDA will need to post 55%+ revenue growth at similar (or greater) margins for at least the next 5+ years.
They might do that… but I know from experience that when you build a “honey pot”… expect a lot of competition.
And that competition is not only going to come from other chip manufacturers (e.g. AMD, INTC, QCOM, AVGO etc etc) – but also their hyperscaler customers (who are currently about 40% of all NVDA’s revenue)
That bet is not for me… but you might be ok with that.
Not Quite Enough
Investors initially cheered NVDA’s result – but that lasted all of 5 minutes.
Post their results – the S&P 500 continued its 2-month slide (failing to make any gains since the first week of Sept) – losing another 1.6%.
I shared this weekly chart the other day – suggesting there’s likely more downside ahead – as the market loses momentum.
The unease from investors is straight forward (from my lens):
As I wrote the other day – there are unanswered questions around whether AI is generating enough (3-5 year) revenue (and profits) to justify the massive spending on infrastructure.
My question for NVDA is whether this is:
- Cisco (CSCO) in 1999 – which saw the stock lose 90% of its value the next decade); or
- Apple (AAPL) in 2018 – a great opportunity to buy?
If only we knew?
Jensen would say it’s obviously “Apple in 2018″… but others might beg to differ.
For example, let’s turn back to our Net PPE growth table I calculated from their last 10 years quarterly reports:
| Ticker | Net PPE 2015 | Net PPE TTM 2025 |
10Y CAGR | Avg Dep / PPE 2015–2024 |
TTM Dep / PPE | Under-depreciation Assessment |
|---|---|---|---|---|---|---|
| AAPL | $22.5B | $61.0B | 10.5% | 3.4% | 5.1% | Slight, improving |
| GOOGL | $29.0B | $238.0B | 23.5% | 3.2% | 2.6% | High |
| MSFT | $14.7B | $229.0B | 31.1% | 3.6% | 4.8% | Moderate, improving |
| META | $5.7B | $177.0B | 37.3% | 4.1% | 2.8% | High |
Here we can see the explosion in Net PPE (Property Plant and Equipment) from 2015 to 2025 (TTM)
GOOGL – for example – has gone from $29.0B to $238.0B today. That’s a CAGR of 23.5%.
For MSFT and META those PPE CAGR’s are greater again.
What this tells me is these companies – with their fabulous free cash flows and above average ROICs and ROEs – are changing from “asset light” businesses to “asset heavy”
META – for example – has been an extremely successful platform play (i.e. asset light).
However, their business today is changing shape – as they effectively shift to owning highly expensive infrastructure.
That’s the question investors now need to answer… will the returns be the same as the past decade?
I don’t think so.
For example, if the (Capex) investment continues at the current pace and the profit is not there – this becomes “value negative”
And this is where our valuation multiple risk(s) come in.
Putting It All Together
In terms of risk – there is one stock worth watching.
And that company is Oracle (ORCL)
The once stodgy database turned “AI start up” has borrowed tens of billions of dollars to be a player.
It’s tethered its entire survival to the AI boom and the market is justifiably concerned.
Will the profits come to pay back the debt?
That’s the bet Larry Ellison is making.
Sink or swim.
Bloomberg reported this week that traders have piled into the company’s credit-default swaps in recent months as ORCL’s massive AI-related spending spree, its central role in a web of interrelated (circular) deals and its weaker credit grades compared with players such as Microsoft or Alphabet
Keep an eye on this… canary in the (AI) coal mine?
Meanwhile Goldman’s trading desk saw a pickup in shorting across macro products including exchange-traded funds, custom baskets and futures.
Bloomy reported the desk also flagged poor liquidity, with S&P 500 top-of-book depth slipping below $5 million versus a one-year average of $11.5 million.
Mmm… the correction chorus grows.
Stay patient.