Investing

Here’s How NVIDIA’s Share Price Could Plummet to $50 Per Share

Eric Bleeker

Published: June 29, 2024 2:12 pm

Last Updated: July 4, 2024 8:07 pm

NVIDIA’s (NASDAQ: NVDA) share price has experienced significant volatility across the past week, with shares dropping from an intraday high of $140.76 to $123.54 at Friday’s close. While this volatility has been driven by profit-taking in shares rather than any specific news, we evaluated the biggest risks facing NVIDIA investors today. 

The Biggest Risk Facing NVIDIA Today: Why it Could Send Shares Back to $50

Here are some highlights from the discussion between 24/7 Wall Street Analysts Austin Smith and Eric Bleeker.

• It’s been a shaky past week for NVIDIA. After hitting a high of $140.76 per share, their shares have seen huge levels of volatility and headed into Friday traded for $123.99 per share.
• So, we wanted to explore the downside scenarios that NVIDIA investors need to be aware of.
• First, we need to note these scenarios are NOT currently playing out. The recent drops in NVIDIA’s share price aren’t driven by news. Instead, most recent volatility has been the result of profit-taking after the stock soared nearly 50% in the month after announcing earnings on May 22. 
• Still, it’s worth noting that the main comparison to NVIDIA’s rise from “bears” has been that it’s the equivalent of Cisco in the Dot Com bubble.
• And Cisco only briefly passed Microsoft in March 2000 before its shares saw a steep crash it’s never recovered from. 
• And now in 2024, we have NVIDIA passing Microsoft (once again) for just a matter of hours before its shares began retreating. So, you can see why this situation has some investors concerned.
• In the near term, we’ve laid out the case why we’re confident NVIDIA will hit $150 per share by the end of summer, and that’s all based on continuing momentum in AI spend and NVIDIA reaching estimates of earnings per share close to $4 per share in 2025 and trading at similar multiples to its peers.
• But, there is a scenario its shares could drop, and drop substantially – which we can explore here.
• It’s worth noting that at $50 per share, NVIDIA would be worth about $1.25 trillion. It came into 2023 trading for about $350 billion. The company also came into 2024 trading for less than $50 per share.
• So while $50 per share sounds like an “impossibly” huge drop from NVIDIA’s recent highs, it is to levels investors valued NVIDIA at just six months ago!
• The biggest threat to NVIDIA is that hyperscalers – the biggest cloud companies – are planning $167 billion in cumulative CapEx right now, and a huge chunk of that is going to NVIDIA.
• For perspective, that level of spend is enough to support 12,000 ChatGPTs, according to estimates from Barclay’s.
• $70 billion of that spend is going to training models – while $95 billion is going toward inferencing.
• It’s important to break these two areas of AI spend out from each other as they have very different dynamics.
• The problem with data center companies continuing to increase investment at the rate they are in inferencing is that the cost of running AI models is coming down at incredible rates. In the past year, inferencing costs have decreased about 80%, for example.
• So, even if the companies in an arm’s race to train new models – companies like Meta, OpenAI, Anthropic, and xAI – keep increasing spending, a decrease in spending on inferencing infrastructure could suddenly crimp NVIDIA’s growth.
• And it appears that when you run the numbers – investment in inference is outpacing any possible usage in the near term absent breakthroughs like AI agents that use vastly more tokens.
• And we’ve seen in recent years that capital expenditures from large hyperscalers can be lumpy.
• It’s likely they’ll continue investing heavily with the release of NVIDIA’s new Blackwell architecture, but the timeframe to watch is the back half of 2025. If these hyperscalers pull back on spending because they’ve overinvested on the inferencing side this could rapidly decelerate NVIDIA’s revenue growth, potentially even sending them to negative growth.
• Long-term NVIDIA investors will know that they’ve seen waves in every market that led to significant booms. That spans video games, to crypto, to earlier AI waves like machine learning growth.
• So, this is a long way of saying, this is the number one situation we’re monitoring when it comes to threats to NVIDIA.
• If you’re an investor in the company and you see hyperscalers cutting back orders, we think the market is going to react fairly harshly.
• How harshly? Well, right now NVIDIA is trading at a slight premium on forward earnings relative to Magnificent 7 peers while on a trailing P/E basis it’s valued much more richly.
• But if spending is pulling back it’s entirely possible NVIDIA could trade at a discount to peers if the dominant fear is they’ve passed peak earnings.
• That could send NVIDIA shares back to levels closer to where it was trading at the beginning of the year, which was around $50 per share.
• As a final note: we are big believers in AI long-term.
• But, it’s important to note that a lot of current spending in AI is driven by FOMO. No one quite knows how close we are to AGI or how big this market can be, so there’s a lot of spending to avoid any risk of being left behind.
• But, spending on FOMO can lead to boom and bust cycles.
• Our long-term belief in AI is up and to the right, but it’s not going to be a “flat ride.” The Internet eventually fulfilled its promise but we know what a volatile ride it was.
• Regardless, I (24/7 Wall Street Analyst Eric Bleeker) continue to personally own NVIDIA, but this is the area I’m most concerned with and I spend a lot of time analyzing as a drop in hyperscaler spending is far and away the number one risk facing the company that I see today.

Transcript:

Eric, it’s been a shaky past week for NVIDIA investors.

After hitting a high of over $140 per share, let’s just pause for a moment, that is a breathtaking level.

But since then, shares in NVIDIA have seen a lot of volatility and they’re heading into Friday trading for $123, $124 a share.

So an idea that I really want to explore with you is what is the downside from NVIDIA today?

Plenty of investors have bought NVIDIA on the promise of continuing gains and the flywheel that is the AI industry and the arms race and gathering all these chips.

But as an investor, you always need to be aware of your downside scenario as well.

The upside is unlimited, but the downside certainly hurts.

And especially after a run like this, there’s a lot of downside below NVIDIA’s current share price.

So for NVIDIA today, what are the risks that investors need to be watching?

Yeah, it’s a great question, Austin, and you know, emotions can get really charged on some of these story stocks.

So I need to note right now that these scenarios are not currently playing out at this time.

The recent drops in NVIDIA’s share price isn’t really driven by any specific news.

It’s more the result of profit-taking and a lot of general volatility in the stock.

But I must note that the comparison that the bears, if you want to call them, have been comparing NVIDIA to has been Cisco in the dot-com bubble.

And what we saw with Cisco was they briefly passed Microsoft in March 2000 for just a very short period before shares retreated.

And that was the peak.

And now we have NVIDIA only briefly passing, wait for it, Microsoft for a matter of hours in 2024 before shares repeated.

So you can see history doesn’t repeat, but it rhymes.

It’s got some people concerned.

So I wanted to talk about, you know, what’s driving NVIDIA shares.

I should note too, in the near term, I’ve laid out the case why I believe NVIDIA will hit $150 per share by the end of summer.

That’s all based on continuing momentum in AI spend and NVIDIA reaching earnings per share kind of multiples that are similar to peers.

You could search us on 24-7 Wall Street.

You could see our full analysis there.

We’ve covered it in detail.

But there are scenarios its shares could drop substantially, which we will explore in this video.

First, I need to note, at $50 per share, NVIDIA would be worth about 1.25 trillion.

That’s in the neighborhood of where Meta’s at right now.

Now, you got to think about it came into 2023, which was last year, trading for about $350 billion.

And it came into this year, believe it or not, trading for less than $50 a share.

So we are just talking about a scenario which sounds like an incredibly huge drop back to levels where the stock started the year.

So from that perspective, it’s probably not as impossible as it might sound in terms of the large percentage drop from today.

So what’s the biggest threat to NVIDIA?

Well, it’s the hyperscalers.

That is the biggest cloud computing companies.

They are planning to invest a cumulative $167 billion, just that group, right now in AI infrastructure and a huge chunk of that is going to NVIDIA.

For perspective, and this is a note I couldn’t believe this when I read it, but it was from Wall Street research.

It was that this is enough spend to support 12,000 chat GPTs and that’s according to research from Barclays.

So where is the spending going?

Well, it’s important how you break down the categories.

70 billion of that spending is going towards training models, while 95 billion is going towards inferencing.

Now, here’s the problem.

The cost of inferencing, which is actually running AI models, if you will, it’s come down incredible levels, about 80% annually right now in terms of cost reduction.

So while training the models, that’s what we hear about a lot with these models moving to trillions of parameters and trying to reach artificial general intelligence.

Well, that can be a continuing arms race costing billions.

It’s less than 50% of the total spend outlay.

So the infrastructure around inference, it’s outpacing at the current time, any possible real usage absent some near-term breakthroughs like you could have AI agents, that are independently using more inferencing.

And what could cause kind of this downside scenario?

Well, there isn’t any near-term let up of these hyperscalers that have large cloud computing instances laying up on spending.

But we’ve seen time and time again, capital expenditures from these huge companies, they can be lumpy.

And we’ve seen NVIDIA itself go through waves in nearly every market it’s ever been in that we’re driving it.

That’s video games, that’s crypto, that’s past machine learning cycles.

So the number one issue I’m watching when it comes to NVIDIA, if you’re an investor in the company and you see big hyperscalers cutting back on orders, I think the market’s going to react pretty harshly because you can construct right now a forward year where data center spending is pulling back probably based on too much spending in this inference category I’ve broken down in this video.

And NVIDIA could quickly lose the premium it has right now on a Ford P basis.

And if it began trading at a slight discount to Magnificent Seven Piers because it was no longer growing as strongly on a trailing basis, well, that’s what gets it back to the levels it was at when it began this year.

And skepticism around the stock was still pretty high.

Now, just as a final note, Austin, you know as well as anyone, I’m a huge believer in artificial intelligence long term.

But if you do own a lot of NVIDIA, you need to look at the long-term chart and you need to see how these cycles have persisted throughout its history.

There are harsh cycles in the moment.

Over time, the company’s performed incredibly well.

So while I believe in the long-term potential of AI, let’s be honest right now, right now a lot of spend is company spending based on FOMO.

No one knows exactly where this technology is going and they don’t want to be, you know, risk being left behind, I should say.

But anytime you’re spending on FOMO, that’s going to create booms, it’s going to create busts.

The future for AI is up and to the right.

But if you’re thinking it’s going to have the same historical returns it had with a flat ride, we’ve never seen a market do that before.

Even the internet was extremely bumpy.

And I think if you’re an NVIDIA investor, you just need to be aware of what the risks are.

And I know it can be a little wonky breaking out training and inference.

But that’s where the risk is, especially on the inference side.

And that’s where the potential downside is.

If you own the stocks and you’re watching for any negative factors, I would put right there, watch for the spending from these data center companies, because that’s the one thing that could really start creating some short-term disruption for NVIDIA shares.

Yeah, a really great advice and really great reminder from investors out there, you know, as we have seen, you know, the semiconductor industry historically has been cyclical.

So we should expect some measure of cyclicality going forward.

And you talked a lot about the FOMO that’s going on with AI hyperscalers and some of the magnificent companies, which I know people are tired of talking about.

But really, let’s put some numbers to how important they are in this equation here.

We’ve seen some estimates from Google’s own executives that they’re going to spend, this is one company, $100 billion on AI.

And of course, if Google’s spending that much on AI, there’s very much a me too, let me catch up moment with Amazon and Meta, and they can fuel this higher.

But if we see any of these companies saying, hey, we’re going to pull back on this cycle, maybe we’re going to take that down to 50, 40, 25 billion, wait for another batch of chips.

That could be the sort of cyclicality that really does say, and it turns out it’s a lot lumpier.

But we need to remind investors that this is all part of a much longer journey as you’re investing in new technologies.

We should expect some cyclicality.

We should expect some investment booms and busts on the path to building out this technology.

Yep.

The same people who are chasing NVIDIA on the upside right now, if there’s a pause in spending, will be the same people who will exit extremely quickly.

So just be aware.