• One of my favorite stocks is way too expensive
  • Why this time is different for nuclear fusion
  • “The ‘voice’ quickly spoke back my order”

Dear Reader,

Welcome to our weekly mailbag edition of The Bleeding Edge. All week, you submitted your questions about the biggest trends in tech and biotech. Today, I’ll do my best to answer them.

If you have a question you’d like answered next week, be sure you submit it right here. I always enjoy hearing from you.

NVDA is headed for a fall…

NVDA is going up more than it should, true, but the company keeps coming up with new software and state-of-the-art hardware to support this rise. As you have said yourself, AI is a very strong driver. Do you really believe that NVIDIA will drop back down to 200 or below?

Gordon E.

Hi, Gordon. I’m glad you wrote in with this question as it will give us a chance to explore some of the nuance around a situation like this. For additional context, my team and I were actually discussing NVIDIA the day it joined the “$1 Trillion Club” (readers can catch up here).

It’s true that artificial intelligence (AI) will be a major driver of productivity improvements and technological breakthroughs in the months and years ahead. I’ve been writing about the technology for years, recommending stocks with strong exposure to AI in my portfolios in The Near Future Report and Exponential Tech Investor, and more recently writing about the technology just about every day in The Bleeding Edge since December. I doubt you’ll find a bigger bull on AI than me…

So it’s possible that we are bullish on a sector and even bullish on an individual company like NVIDIA and its prospects, but bearish on it as an investment due to being grossly overvalued.

The nuance in your question is around time frame. In the short term, NVIDIA’s share price is way ahead of itself, and it will fall. But if we look out three years or more, it is likely that NVIDIA will grow into its trillion-dollar valuation.

These things can all be true at the same time. It’s the time frame that’s different. At NVIDIA’s current valuation, I just don’t see NVIDIA’s current share price as being sustainable in the months to come.

Longtime readers may remember that I officially recommended NVIDIA in the pages of The Near Future Report all the way back in January of 2019. Here’s what I said about its valuation at the time:

NVIDIA’s valuation was so high in 2018, it was destined to collapse. Take its fiscal year 2018 as an example. The company was trading at an enterprise value to sales ratio (EV/Sales) of 14.7. 

And collapse it did. Over the course of the last three months, NVIDIA has collapsed 56% from its highs in early October to its bottom in December – losing over $98 billion in valuation.

We went on to buy NVDA at an EV/Sales of around 6. The stock was trading for about $35. It was great timing, and we enjoyed a great rally over the next several years.

But where is NVDA trading today? The stock now trades at a 38 EV/Sales. In other words, it’s two and a half times as expensive as it was before the crash in October of 2018.

True, the larger trend of artificial intelligence has been a remarkable tailwind for the company. And as I said, I’m very bullish on AI in general. But with a volatile market, terrible inflation, an economy that is arguably in a recession, and an aggressive Federal Reserve, I just don’t see this as being sustainable. NVDA could easily lose 50% of its value from here. That would put the share price below $200.

For historical perspective, when I first recommended NVIDIA back in early 2016, it was trading at a valuation of about 2.5 EV/Sales! Between 2017 and 2020, its EV/Sales range was typically between 7 and 14. So to see any semiconductor stock trading at a 38 EV/Sales is astronomical.

I would love to see NVDA trade below an EV/Sales of 10. I doubt that will happen amidst the current AI hype cycle, but if it got there, it would be a great re-entry point. But I just can’t get excited about NVIDIA at these elevated levels. Buying any stock at a 38 EV/Sales is not investing, it’s rolling the dice. When NVIDIA’s share price falls, it’s going to be terrible news for the market.

An important thing to understand about indices like the S&P 500 is that they are market-cap weighted. That means that larger companies are given more “weight” in the performance of the index. Apple, Microsoft, Amazon, Google, and NVIDIA combined make up 20.4% of the S&P weighting.

The reality is that the vast majority of the returns year-to-date have been powered by just a handful of stocks like NVIDIA that have soared so far this year. But when that reverses, it will bring the rest of the market down with it.

The exponential growth of nuclear fusion…

Hi Jeff,

I love reading The Bleeding Edge. It has the most info-per-word of any newsletter I’ve seen.

I’ve been reading your updates on nuclear fusion with excitement and anticipation, not just from an investing perspective, but also from excitement for what this technology can do for both clean and inexpensive energy production.

Just this past week, I met a new friend, Steve, who’s a retired nuclear engineer. I asked him when he thinks nuclear fusion will become a productive source for energy production, and with no hesitation, he said, “Not for at least 100 years.”

I was taken aback by his answer and asked why he thought it would take so long. He told me he got into nuclear engineering 40 years ago to build nuclear fusion rockets to fly people to Mars in less than six weeks.

In all his years of working in the field, he has seen attempt after attempt to control and channel the energy emitted from nuclear fusion fail. He hasn’t seen anything that makes him think we’re anywhere close yet.

What have you seen that makes you think they are close and have overcome those big obstacles? What is Steve missing that you have learned?


Mike H.

Hi, Mike. Thanks for writing in. And your comments from your friend gave me a chuckle. I hear these kinds of comments from industry insiders with decades of experience all the time.

It reminded me of a genetic editing conference that I attended in Boston back in 2015. I started following genetic editing back in 2013 just after reading the seminal paper on CRISPR technology that was published at the end of 2012. I knew at that time the technology was transforming and would change the world. And I knew that the challenges of using the technology would be overcome.

At that conference, I remember clearly sitting at a round table about 30 feet back from the speaker’s podium. During one of the breaks between presentations, I spoke with the person beside me, a PhD in genetics, who told me I was wrong about how quickly the progress would be with CRISPR and that it would take decades before any CRISPR therapy would work.

Well, it was only a few years after that when a woman, Victoria Gray, was cured of sickle cell from a CRISPR genetic editing therapy. And the first FDA approved CRISPR therapy will most likely happen within the next 9 months.

It wasn’t the scientists’ fault. It’s just that they have a normal human bias. They are anchored to past challenges. Anchored to what they know. And they most often are not up to speed on the latest technological developments.

I can’t tell you how many times I’ve cited exciting research with doctors, physicians, and scientists in their own field of specialty, and they aren’t even aware of it. There are only so many hours in a day, and busy professionals work with what they know. Very few spend their free time trying to stay ahead of the curve. It’s not a fault. It’s just reality.

That’s the reason for the gap in what I share with my readers and what your friend had to say. Here’s a fun experiment for you. Ask your friend to perform a simple task. Ask him to name just five promising private nuclear fusion companies. Just their names, nothing else. 

He won’t be able to. Whenever I speak with a PhD or an expert in their field, I’ll always ask simple questions to politely test whether or not they are aware of what is happening in their own field. They are always simple, obvious questions that anyone “in the know” would be able to answer in a heartbeat.

I don’t do it to embarrass anyone. I just do it so that I can understand what topics I can discuss with someone. If they’re not aware of these advancements but seem curious, I can share them with them. If they seem uninterested, we can talk about something else.

But to answer your question more specifically, there are a few things that have changed in the last few years that have set nuclear fusion on an exponential growth curve.

At a fundamental technology level, we’ve seen major breakthroughs in both materials science and plasma physics that have opened up the doors to several paths toward nuclear fusion.

The materials are critical as a magnetic chamber has to be built to control a plasma reaction. And the plasma physics developments have been fascinating as they have led to several paths to creating and sustaining a plasma reaction. Here are just a few examples:

  • Magnetic confinement fusion

  • Inertial confinement fusion

  • Magnetized target fusion

  • Field reversed configuration

  • Stellarator

  • Orbitron

All six of these distinct approaches to nuclear fusion are currently being built by private companies. It’s happening, and the first prototype reactors will be tested as early as next year. The 2024 to 2026 time frame is going to be ridiculously exciting in this space.

The other critical technological nuance is that the software systems, specifically machine learning technology, have advanced at a doubly exponential rate. Machine learning (a form of artificial intelligence) is the key to controlling the magnetic fields required to sustain a fusion plasma.

And because of all the great technological progress, private capital has started to flood into the industry. This has long been the missing element. The U.S. government is currently uninterested in supporting nuclear fusion development. This is deeply ironic considering that one of its major policy positions is all about “clean energy.” It has aggressively supported the electric vehicle industry and completely ignores that the majority of U.S. electricity used to “fuel” electric vehicles comes from burning coal and natural gas.

To put this into further context, 2022 was a record year for investment in nuclear fusion companies. In total, $4.86 billion was invested in the sector. And of that $4.86 billion, only $117 million was from public (government) funding. That represents just 2.4% of the total, basically nothing.

The entire industry has hit an inflection point. Technological advancement combined with the capital to bring this breakthrough technology to life has finally coalesced (note: I got a chuckle out of using this word; after all, the first four letters are “coal”). When these two things happen at the same time, there will always be major breakthroughs.

You’re going to have some incredibly exciting, and concrete, developments to share with your friend in the years ahead. I’ll make sure that you’ll know about it as soon as they happen. And I’m confident that you’ll be able to get a kick out of the moment when you share something with him, and the lightbulb goes off above his head – he’ll suddenly realize that it’s happening now.

The AI will take your order…

Hi Jeff,

Read your article about Wendy’s AI for ordering fast food. I actually experienced AI with Del Taco (my favorite fast food restaurant).

A little over a month ago, I visited my local restaurant. I pulled up to the order area and was quickly greeted by a “female” in an extremely clear and pleasant voice. It was Del Taco’s AI!

I wear a hearing aid, but I always have problems hearing the words that the person at the other end of the speaker is saying to me. But the speech was so incredibly clear I could have understood it without my hearing aid.

I ordered my food, but I immediately realized that I forgot to tell them to add extra sauce. So I quickly said, “Extra sauce on the first burrito.” I thought they would never get that last bit, but it was worth a try.

The “voice” quickly spoke back my order and said, “The chicken burrito with extra ‘Secret Sauce.’” I thought the whole experience was incredible. Every fast food place needs this technology!

Doug S.

Thanks for sharing your story, Doug. Awesome! I always enjoy hearing from readers who interact with the technologies we cover out in the real world. Boots on the ground research… very much appreciated.

As we have been sharing the past several months, one of the first practical applications of AI we’re seeing is in fast-casual restaurants. Wendy’s is teaming up with Google to use a generative AI to take orders at the drive-through.

But – as Doug just reported – Del Taco is also adopting this technology. Specifically, the restaurant chain is teaming up with Presto, an early-stage company that specializes in automation for the restaurant industry.

Presto is also deploying its technology into Checkers, another popular franchise. A study of Presto’s technology at Checkers showed that the AI was able to complete 99% of orders with no human intervention.

Doug, you’re absolutely right. This technology is simply too compelling for restaurants to ignore. And the experience for customers is much better. It’s only a matter of time before this technology becomes a “must have” for every fast food restaurant.

And if any readers experience this technology “out in the wild,” feel free to send me a note. I’m always happy to hear from subscribers, and I deeply appreciate hearing about real-world experiences with tech.


Jeff Brown
Editor, The Bleeding Edge