It’s AMA day at The Bleeding Edge…
Every Friday, I take on your questions that we receive in our Brownstone mailbag. It’s a great chance for me to address questions that I may not have addressed in normal issues of The Bleeding Edge.
The questions also let me know what’s front of mind for many of you, so I really appreciate you taking the time to write in.
I’ve got a great round of questions to get to today. From Intel’s recent “collapse” to Tesla’s robotics competitive advantage to the potential for hydrogen-fueled transportation…
But before we get into it, I want to put something on your radar…
I’ve got an important announcement coming up.
You might have noticed this week at The Bleeding Edge that we’ve been very focused on the future of artificial intelligence and what the next leap in this transformative technology looks like.
Over the last couple of years, AI has taken the world by storm. We’ve seen an incredible boom in AI stocks with shares of companies like Nvidia skyrocketing to unbelievable highs.
For those subscribers who bought shares in NVIDIA when I first recommended it in early 2016 and held on through to today, you’d be sitting on more than a 20,000% gain.
But this is just the beginning. We’re nowhere near the end of this once-in-a-generation trend. And I want to get my subscribers well-positioned for what’s coming next in AI.
There’s a major industry event coming on August 28 that I believe will trigger a seismic shift in the market. So, next Wednesday, August 14, at 8 p.m. ET, I’m sitting down with MarketWise CEO Porter Stansberry for an AI Emergency Meeting.
There, we’ll talk about the event that’s about to send shockwaves through the entire AI industry… and how we can play it in our portfolios. Just go here to get your name on the attendees’ list. We want to make sure you don’t miss this.
Now, let’s have some fun with the mailbag…
Re: The $7 Trillion Investment of a Lifetime
Thanks, Jeff, for this great article.
What about Intel and the massive investments put in place in new fabs in the U.S. and Europe and its external foundry strategy? Do you see it as a component of this supply chain resiliency you talk about?
Again great material! keep it coming!
– Guillermo T.
Hello Guillermo,
You’ve hit a hot button for me – Intel!
This is a company that should have been doing extremely well during the last five years. After all, we’ve seen a boom in leading semiconductor companies… especially those related to artificial intelligence.
And yet, let’s have a quick look at Intel’s share price over the last five years. Ouch!!!
What happened?
Intel’s story is so sad, especially for such a storied and foundational company in the high-tech industry.
The iconic semiconductor pioneer has fallen due to complacency, groupthink, bureaucracy, and lack of innovation for so many years. Its failures, flops, and product delays over the last decade are simply too many for me to list.
Intel’s revenues are down from $74 billion in 2021 to $52 billion this year. Revenues dropped this year when the opposite is happening with NVIDIA, AMD, and so many others. While Intel has about $35 billion in cash, it also has $53 billion in debt and will have about a negative $14.7 billion in free cash flow this year.
That’s why the share price has dropped from nearly $70 to less than $20 right now. And just since this January, the stock is down more than 60%.
As much as I want to like Intel, I just can’t get excited about it… not even as an income-generating stock. There could still be more downside.
With all that said Guillermo, your point about supply chain resiliency is definitely a bright spot. One of the key reasons that it has had negative free cash flow since 2022 is because it has been spending heavily on expanding its own manufacturing capacity.
Unlike NVIDIA and AMD, Intel manufactures most of its semiconductors. It requires tens of billions in investment to keep up. And its new manufacturing capacity will bring improved supply chain resiliency to North American and European markets.
Intel is also receiving help in the form of grants through the CHIPS Act. In total, $8.5 billion in grants and $11 billion in loans across the following projects:
This will certainly help improve manufacturing resiliency, specifically for semiconductors related to personal computers and servers.
I’m glad to see the company making these investments. But from an investment perspective, Intel is trailing years behind the kinds of semiconductors that TSMC or even Samsung is manufacturing… and it has been largely irrelevant in mobile computing and anything related to artificial intelligence –the two largest growth markets in the last decade.
You write (without giving a reason): Tesla is so far ahead of the competition… how can you be so sure?
– Staehli M.
Hi, Staehli. Thanks for writing in.
For readers who missed it, this is referring to a July issue of The Bleeding Edge – Did Tesla Mess Up Its Optimus Deadline?
I was reflecting on some of the developments from the World AI Conference in Shanghai earlier this year… specifically, the various humanoid robot prototypes on display there and how they all seemed to be chasing what Tesla has already achieved with its Optimus robot.
This Bleeding Edge issue was published shortly after Tesla CEO Elon Musk announced an updated plan to hold off on Optimus production for external customers until 2026 – which delays external sales by roughly a year.
This led to speculation that, by falling behind on its deadline, Tesla was falling behind the competition.
A wild notion considering, as I said in the essay, just how vastly far ahead of the rest Tesla is in the autonomous robotics industry.
I actually wrote on this very topic earlier this week in The Bleeding Edge – “If I Only Had a Brain”…
The hardware for a humanoid robot, while complex, is technologically an easier problem to solve for.
The really hard part about developing an autonomous general-purpose humanoid robot is the software – the brain.
And this is where Tesla’s Optimus is lightyears ahead of the competitive field.
After billions of miles of data collected by Teslas driven on auto-pilot and full self-driving mode, Tesla has an artificial intelligence capable of seamlessly driving any machine from point A to point B, with the ability to infer the correct actions in a split second, even in completely unstructured environments in areas where no Tesla has driven before.
This autonomous software is the foundation for Optimus’ “brain.” And just like the human brain, that software has been trained on video inputs, just what the cars “see” in real-world conditions.
The inputs for Optimus are the same. Computer vision is used to “see” the environment so that it can carry out the tasks that it has been instructed to perform.
Optimus processes the world much in the same way that we do. It sees its environment and then reasons the most efficient way to get a task completed. No other player comes close to what Tesla is doing right now with humanoid robots.
Tesla may have entered the humanoid robot race late, but its progress has been extraordinary. And that’s because it had already developed the most complex and difficult technology related to a humanoid robot – the autonomous software built using artificial intelligence.
Musk and his team solved the most difficult problem first, and now they are building the hardware that can manifest AI and become a productive member of society.
We’re going to stay very plugged into developments in this industry. There will be more interesting developments between now and the end of the year.
Jeff,
Thank you for your great insight and experience. Your Walmart story was a testament to brick-and-mortar companies because they will eventually overtake internet sellers.
Consumers are becoming tired of buying products they are not satisfied with and using their credit card information for a lot of reasons.
To another point, FedEx recently announced their two-day service to anywhere in the continental USA delivered by FedEx drivers on their trucks. I think they were feeling the pressure from the USPS Priority Service that Amazon was mostly handling for them.
Do you think FedEx, UPS, or USPS will get into Drone delivery?
– Fred M.
Hi Fred,
Yes, definitely. And all three parties that you mentioned have been evaluating and even experimenting with the technology.
FedEx started experimenting with Wing, a subsidiary of Alphabet, back in 2019 for last-mile delivery. And it also has been partnering with Elroy Air since 2022 for middle-mile delivery to quickly move packages from one sorting or distribution center to another.
UPS also started experimenting with a last-mile drone service back in 2019. It even has a division called UPS Flight Forward, which is responsible for drone delivery of small packages. UPS Flight Forward uses Matternet M2 drones.
UPS has also announced that it plans to purchase eVTOL aircraft – that’s electric vertical take-off and landing aircraft like we wrote about in The Bleeding Edge – from Beta Technologies for middle-mile logistics as well. While Beta’s initial aircraft will be piloted, its aircraft will eventually become autonomous.
I also remember back in 2019, the USPS was also investigating using drone technology for last-mile delivery. I’m not aware of any new major developments with the USPS, and I expect the USPS to be a laggard in this space compared to Amazon, FedEx, and UPS…
But given the USPS’s infrastructure footprint, it would be feasible and logical for it to launch similar services in the future.
I’m curious. Do you believe in any limits at all on “free speech?” I’d like to see your answer, b/c it seems like you don’t.
– Michael S.
Hello Michael,
This is the first time anyone has written in about this, so I appreciate the question. And there are limits as defined by the law. I’ll give some examples below.
With regards to the U.S., we have to start with the First Amendment:
Congress shall make no law respecting an establishment of religion, or prohibiting the free exercise thereof; or abridging the freedom of speech, or of the press; or the right of the people peaceably to assemble, and to petition the Government for a redress of grievances.
There’s a reason that freedom of speech is the First Amendment. If a society cannot speak freely, and share thoughts, opinions, and scientific research, there is no true foundation to build upon.
There are natural limits, however, defined by the law. For example, the First Amendment states that we can peaceably assemble, but we are not permitted to assemble and call for genocide against a race.
Sadly, that has not only happened in the last year, but it was also permitted, despite being against the law.
We cannot stalk and threaten to kill one another. An absolutist might say that’s still freedom of speech, but it is clearly against the law. Pedophilia and child pornography are clearly illegal and are limited by the law in both action and speech.
There are many other examples, but the key point is that freedom of speech must be consistent with the laws that a country sets.
An orderly and civil society can maintain both freedom of speech and the enforcement of laws against criminal behavior.
Unfortunately, right now, we are seeing many states that are completely ineffective and/or unwilling to enforce laws against criminal behavior and at the same time employ tyrannical policies – including censoring speech – against those that abide by the laws (and also pay taxes). This is also happening at the national level.
This situation is untenable, and if allowed to persist will lead to a horrible Orwellian outcome.
Hi Jeff,
Saw an interesting article on Apple using TPUs vs GPUs, built by Google.
What are your thoughts on this development? How will this affect the industry? How does this position Apple?
Thank you.
– Michael M.
Hi Michael,
Great to see you staying on top of things. This was an interesting development, or more specifically, it was interesting that Apple made this publicly known. Apple tends to be far more secretive about its own development process.
What happened is that Apple uncharacteristically published some research related to Apple Intelligence on July 29, which specifically shared that Apple trained its own foundational model on 8,192 TPUs (version 4) from Google.
If you’re interested, you can find that research, appropriately titled, right here: Apple Intelligence Foundation Language Models.
To some, this seems like a bigger deal than it really is.
I don’t know why Apple chose Google’s TPUs over NVIDIA’s GPUs. Given the backlogs for NVIDIA GPUs and the fact that Apple is not a big NVIDIA customer – Apple designs its key semiconductors and has them manufactured by TSMC, including its CPU, GPU, and neural processing unit (NPU) – it may have just been easier to get Google’s TPUs for training.
The reality is that NVIDIA GPUs, AMD GPUs, Google’s TPUs, and a number of other AI-specific semiconductor companies all have fantastic semiconductors well-suited for training foundational models.
NVIDIA’s advantage lies in having the largest software development community with its CUDA software platform, which is the software platform upon which companies can use NVIDIA’s GPUs.
As we know, Apple has been behind in artificial intelligence, and it has been playing catch up. That’s why it chose to make its research public. It was published to give software developers confidence that Apple is doing legitimate work with artificial intelligence, and now it is…
Apple may not be on the bleeding edge right now, but it has the financial resources to be a major player in artificial intelligence. And the Apple ecosystem of products gives it an amazing distribution channel that will ensure widespread adoption of its AI products.
Additionally, Apple’s tight integration of both hardware and software, from the semiconductors on up, is a competitive advantage in the industry. Less important is what semiconductors Apple uses to train on. More important is what semiconductors Apple uses for inference, running the AI locally on iPhones, iPads, and Apple laptops.
I’m excited to see where Apple goes from here with Apple Intelligence and how it weaves AI into its consumer products.
Jeff, I have recently read that a new way of producing hydrogen has been discovered which is much cheaper than the old way. Also, I have read that both Korea and Japan have developed hydrogen engines for vehicles.
I would like to hear your take on both of those new developments and how it relates to the possibility of hydrogen fuel replacing fossil fuels in the future.
– Hank P.
Hi Hank,
I’m often asked about hydrogen as the next great source of clean energy to fuel transportation. It’s such an exciting topic. The thought of powering our cars, trucks, trains, and even planes with the most abundant element in the universe, whose byproduct is just water when used, sounds like the solution to greatly reducing global carbon emissions.
And it’s true, hydrogen seems perfect on the surface. It stores three times as much energy per unit of mass compared to gasoline. When it is combined with air, the energy released can power a vehicle, and it combines with oxygen to produce water. We can’t do much better than that, right?
Yet that’s not the whole story. We have to understand how hydrogen is made to see the whole picture.
Hydrogen is produced from water. About 70 million tons of hydrogen are produced each year which is primarily used for ammonia fertilizer. And 96% of hydrogen production is made by a process known as steam-methane reformation.
Here’s the problem… this process uses energy created by natural gas, coal, and oil to produce that hydrogen. In all, the industry produces 830 million metric tons of carbon dioxide every year to produce this “clean” hydrogen fuel.
Not so clean after all.
In fact, I would argue, what’s the point? If we have to burn massive amounts of carbon-based fuel just to put hydrogen in our cars, we aren’t helping the environment at all. We are only displacing where the carbon emissions take place, not whether or not they happen in the first place.
It is really no different than fueling our electric vehicles with electricity that is produced from coal, natural gas, or oil. It is nonsensical to think that we are helping the environment by doing so.
As for the remaining 4% of hydrogen production, it is produced using electrolysis which uses electricity to split the hydrogen out of the water.
Again, at the surface, this sounds better than steam-methane reformation; but we must ask… where does the electricity come from? And again, the answer is almost entirely from fossil fuel power plants or nuclear fission power plants (radioactive waste).
So where does this leave us?
We can use carbon capture technology to capture the carbon dioxide that is emitted when we produce hydrogen. That works, but it is not perfect. At least 10%, and often more than 20% of the emissions are not captured in this process.
It’s better than 100%. But still, it is not clean energy production. We can make up for the difference by using carbon offsets, for example, by planting more trees halfway around the world; but the fact remains that we’d still be burning fossil fuels in the production of the hydrogen.
Worth noting is that using carbon capture technology will also result in hydrogen that is about twice as expensive than that produced using steam-methane reformation.
And if we can make the leap to producing hydrogen with 100% renewable energy, and that is a major leap because of how much power is required, the costs are four times higher than steam-methane reformation.
Why is this important? Well… hydrogen is fuel, and consumers and businesses have to pay for it. We can think of a gallon of gas as having about the same amount of energy as a kilogram of hydrogen. But a hydrogen fuel cell system is twice as efficient as a gasoline system.
We can expect close to 70 miles per kilogram of hydrogen fuel for a small hydrogen fuel cell car.
To put things in perspective, it takes about 50–55 kilowatt hours of electricity to produce a single kilogram of hydrogen fuel. That’s the equivalent of about two days of electricity consumption for an average home in America.
We should think about that, two days of an entire household’s energy consumption just to produce one kilogram that provides enough fuel to travel 70 miles.
Where hydrogen is problematic is in its volume. It takes up a lot of space, so we can only carry about 5–6 kilograms of hydrogen in our tank. The other tricky nuance is that hydrogen molecules are so tiny, that they easily leak out of most containers.
The average price of hydrogen fuel in California, which has the highest number of hydrogen cars, is about $16 per kilogram. That works out to the equivalent of about $5–6 a gallon of gasoline. And we have to remember that almost all of the hydrogen was produced by burning fossil fuels.
Without billions of dollars in subsidies, hydrogen just doesn’t make economic sense; and because of where the energy comes from in the production of hydrogen – mainly fossil fuels – it doesn’t even make environmental sense.
For hydrogen fuel cells to be both environmentally sustainable and economical, the world must address how it produces baseload power. This is the kind of power required to manufacture the 70 million tons of hydrogen produced every year.
The most promising technology to achieve this is a kind of nuclear fusion (not fission) technology that does not produce any radioactive waste. Carbon-free emissions with no radioactive waste is a sustainable energy production strategy and one that we would all benefit from. Something very tangible that the industry can work towards.
That’s it for today’s mailbag. Thanks for writing in. If you have your own questions or comments, you can reach me and my team right here.
And please don’t forget to go here to sign up for my for my AI Emergency Meeting next Wednesday, August 14, at 8 p.m. ET.
If you missed the first wave of profits in this AI boom, you won’t want to miss this meeting.
I hope you have a great weekend.
Jeff
The Bleeding Edge is the only free newsletter that delivers daily insights and information from the high-tech world as well as topics and trends relevant to investments.
The Bleeding Edge is the only free newsletter that delivers daily insights and information from the high-tech world as well as topics and trends relevant to investments.
