What another incredible week it has been for the artificial intelligence industry.
Just when we might think that it couldn’t get more interesting and more exciting… wham! More incredible news.
The week kicked off with another incredible deal. This time, it was OpenAI committing to acquire 6 gigawatts of AMD’s (AMD) GPUs over the next few years.
It’s an odd thing to now quantify the scale of a massive GPU deal by gigawatts. Why not just name the quantity of the intended purchase?
Either way, it is one of the largest semiconductor deals in history and a game-changer for AMD. It also highlights the reality of NVIDIA’s contract manufacturing constraints.
I’ve long maintained that AMD’s guaranteed manufacturing production from its manufacturing partner is something that the industry simply won’t be able to overlook. AMD is capable of delivering products on par with NVIDIA, and it also has the best high-end CPUs for AI data centers in the market.
AMD is a powerhouse, and now it has the guaranteed business to scale even further, which will result in more business for its GPUs. It’s also good for maintaining a competitive marketplace. In the absence of that, NVIDIA’s gross margins would become even higher than they are today.
Making the deal especially unique is a provision that OpenAI will receive warrants for up to 160 million shares. The warrants are tied to vesting schedules, AMD’s share price, and the order volumes achieved by OpenAI. If all warrants are received and exercised, OpenAI will end up with about 10% of AMD.
It’s a steep price to pay by AMD, but the limited dilution will be worth it. The business from OpenAI is large enough, and the resulting future business from other hyperscalers will make it a smart deal when all is said and done. The market certainly liked the deal, with AMD rising as high as 30% after the deal was announced.
If that wasn’t exciting enough, Tesla released its first major update to full self-driving (FSD) in a year – version 14.
Early reviews of the technology are stunning. I won’t spoil it, as I plan on writing about it early next week, but this release is clearly something special.
In this week’s AMA, we’ll revisit some popular topics of the last few weeks, specifically SMRs, AR glasses, and how digital tokens appreciate in value.
Hope you have a wonderful fall weekend.
Jeff
Greetings, Jeff,
Given the distress around nuclear reactors in the war zone in Ukraine, it’s clear that Russia could unleash a Chernobyl-like or Fukushima-like event at Zaporizhzhia with as little as a $1000 drone. This speaks to the risk of legacy nuclear reactors worldwide (but several in the U.S.) acting as “pre-deployed weapons of mass destruction.”
To me, this speaks to the wisdom of decommissioning these and removing the fissile material, but instead, as you have mentioned, these are being extended or recommissioned like Three Mile Island.
My question: Would small modular reactors similarly be subject to explosion with widespread nuclear pollution if attacked or sabotaged? Or are they safer in such an event?
Would fusion reactors be safer, or would the collapse of plasma confinement in such a sabotage scenario be damaging to the neighborhood to a similar magnitude?
Long-time Bleeding Edge reader,
– Richard S.
Hi Richard,
That’s an important question, and always a concern, especially for the older legacy nuclear fission reactors.
It’s pretty incredible that in the decades of using nuclear fission reactors for emission-free, clean energy production, there have only been three major accidents:
That’s pretty remarkable considering more than six decades of operation of legacy nuclear reactors. Better yet, compared to all other forms of energy production, nuclear power per unit of energy produced as the lowest number of deaths.
To your question about the risk associated with a potential drone attack on one of those facilities, Sandia National Laboratories ran an interesting test in 1988. It took a 27-ton F-4 Phantom fighter jet and sent it hurtling into a 3.7-meter slab of reinforced concrete at 765 km/h to simulate an attack on a nuclear fission reactor.
What happened? Well, 96% of the fighter jet’s kinetic energy went into its own destruction, and only 4% was absorbed by the 3.7-meter slab. The maximum penetration of the concrete slab was just 60 mm.
So, to answer your question, it would be highly unlikely that drones could do any significant damage.
The much greater risk boils down to potential security issues. If bad actors gain access to the interior of the facility and can disable key safety systems, then it becomes very dangerous.
And no, it just doesn’t make any logical sense to decommission these plants, unless they need significant safety system upgrades or have neared the end of their life. It is economic suicide to do so in the absence of a replacement.
Germany is a perfect example. It decommissioned all of its nuclear reactors in the name of “green energy” and became heavily dependent upon Russian natural gas.
When Germany decommissioned its reactors, electricity prices rose significantly and made Germany exporters less competitive.
Then the pipelines were sabotaged, and Germany had an energy crisis. Then, Germany did the unthinkable… it started producing electricity using coal. And it continues to use coal, and it is also buying Russian natural gas through intermediaries, defying supposed EU sanctions on Russia.
How’s that for “clean energy”?
The results of these policies to shut down the nuclear power production have resulted in complete economic stagnation. Germany’s economic output is roughly at the same level as it was in 2019. This year has been another disaster for Germany’s economy. And it is still using coal and Russian natural gas.
Shutting down emission-free energy production from nuclear reactors is economic suicide. Until new small modular reactors (SMRs) can be brought online – along with nuclear fusion reactors – the world needs to maintain this clean, consistent source of power generation.
And yes, SMRs are inherently safer, assuming good site security. The key point about SMR designs is that they have passive safety systems.
No human involvement is required to shut down an SMR. The biggest risk is if someone gained access to the interior facilities of an SMR and somehow was able to disable the passive safety systems.
Nuclear fusion is actually the safest because when a nuclear fusion plasma becomes unstable, the fusion reactions simply shut down.
New energy production is so critical right now to support the largest infrastructure project in history – namely, the AI data centers that are required to achieve both AGI and run artificial intelligence applications. This is why decommissioned reactors are being upgraded and recommissioned.
It’s also why the U.S. is aggressively working to test and build SMRs, as well as a range of nuclear fusion reactor designs. Once SMRs and fusion reactors are producing grid-scale energy, in time, older fission reactors will find their time to be decommissioned and replaced with these exciting new clean energy technologies.
Hello Jeff,
I have some comments and questions regarding Meta’s new AR glasses. First, from [the recent] AMA, I love that prescription glasses that are also AR are being developed.
I’ve recently reached the age where I have to wear reading glasses to clearly see a lot of things on my iPhone. The text size modifications are just not good enough.
Second, in relation to Meta’s AR glasses and their business plan of selling their “customers” data, is there anyone trying to develop a platform-agnostic version of AR glasses that can be used anywhere?
I was honestly kind of surprised that you didn’t even mention the negative privacy considerations, given your views on World Coin and Google’s practices over many years, which I strongly agree with.
Thank you for the great work you do.
– Synthya G.
Hello again Synthya,
I am certainly in agreement with you about the negative privacy concerns.
I’ve written so much over the years about the horrible data surveillance and negative privacy practices of WorldCoin (now just World), Meta, Google, Microsoft, and TikTok that I just didn’t mention it last time.
To answer your question, yes, there are alternatives that claim they are compliant with CCPA and GDPR privacy regulations. Even Realities is one player that produces prescription-compatible augmented reality (AR) glasses. XReal is another similar company that has been at it for a while.
Both companies, however, are based in China, and China-based electronics companies have a horrible record of violating privacy regulations and sending data back to China.
A U.S.-based option is Vuzix, which, historically, has been enterprise-focused but has recently been working on smart glasses for consumer use as well.
But the biggest and likely best alternative will come from Apple. We explored Apple’s recent strategic pivot on Monday in The Bleeding Edge – Someone Give Apple a Pair of Glasses.
The only downside of Apple’s pivot is that we’ll probably have to wait a couple of years before they become available.
Hi Jeff, on your most recent Friday AMA, when talking about Civic and the CVC token, you stated, “There are many cases where a company is working on useful technology, but the value of that technology doesn’t accrue to the token.”
For example, the Brave browser has shown massive growth from 1 million users in 2018 to 100 million users today, yet the BAT token has gone nowhere.
My question is, with these crypto projects, how can you tell if a promising project will translate into the token price?
– Jarod R.
Excellent question, Jarod.
There are a lot of variables to consider when trying to discern whether a project alone will be successful, let alone whether or not its token will be a quality investment to buy and hold.
Brave is a classic example. The browser it created is incredible. This month, it hit 100 million active monthly users. For comparison, Safari has 1 billion, which adds up to about 17-19% of the global market. Brave is focused on blocking ads and trackers to improve the web browsing experience.
The browser also includes tooling to help protect privacy and even navigate on Web3 applications with greater ease. It’s a great product.
The token was initially used to help bootstrap users. Users could opt in to receiving ads and, in exchange, receive their native BAT token.
In time, the BAT token’s value dropped significantly, and the BAT rewards used for opting in were not substantial enough to incentivize ad engagement. But more importantly, two issues proved too difficult to tackle for the token to maintain sustainable value.
The first was supply. The continuous rewards program meant more and more tokens were entering circulation.
This would be like the U.S. government constantly issuing dollars like it had done during the COVID pandemic. The result for the U.S. dollar was inflation of around 9%. Inflation is another way to say that a dollar can’t purchase as much as it could before. It’s a loss in purchasing power.
The same goes for a cryptocurrency. When a project increases the supply of a token, it’s reducing the token’s value. This is the blockchain equivalent of monetary policy. And it starts to hint at one of the hardest issues for new projects. They are not only building bleeding-edge technology, but also managing the supply of a currency.
This is referred to as cryptoeconomics or tokenomics. They must learn some of the nuances that come into play in minting a currency. And most teams are not economists or former Federal Reserve officials. It’s also why most projects tend to overinflate their supply and, in turn, damage the value of the token.
The second issue for Brave was around economics. This is balancing the supply and demand aspect for a currency.
We touched on the supply issues in the prior talking point, but demand is equally important. If there is no need to essentially hold a token or use it for some purpose, then what’s the reason that new users or buyers would hold the token? This is where sound token design comes into play.
Many projects have increased their focus on this issue. Many of the projects we recommend in Permissionless Investor have clear demand drivers for the token.
Users might need to stake the token to validate activity of the protocol, other projects might use a portion of revenue to buy supply off the market and remove it from circulation, and others might pass along yield to the token to make it a productive asset.
Brave didn’t incorporate any demand drivers with any success. And it’s in part why the BAT token has become an afterthought in the digital asset space.
If you take a look around the Brave website, you’ll see most discussions around their BAT token have to do with creating utility for the token. It is respectable that the team is still attempting to revive its token, but in all truth, many projects don’t recover from this mismanagement of their token.
That’s what we see with Civic. And instead of trying to even tackle their prior errors, they’ve let time take its course to make the token a forgotten component of the project.
This has happened more often than most realize in cryptocurrencies. And it’s why whenever finding a new project or token to invest in, we must look at the supply and demand drivers of the token.
It’s not enough to simply view the technology and the problem being solved… Or the team behind the project… We need to understand if the proposed economics will result in value accrual to the token that is enough to match or exceed the supply of the token.
That’s why when we recommend a token in Permissionless Investor, readers will often get to know about what demand drivers exist for the token. And if none currently exist, we state that, but then explain what is coming down the pipeline based upon governance discussions that could pull that demand lever to drive value to the token.
It’s not enough to understand the technology of Web3. We need to look at these tokens with the eye of an economist to determine if the value the project realizes will translate to the token’s price.
Thanks for giving us an opportunity to dig deeper into this interesting topic.
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.
