- Generative AI is coming to our phones…
- Can energy be “teleported”?
- The EU’s CBDC is a stalking horse…
Last year was a tough year for Intel.
Intel has dominated the semiconductor industry for decades since it went public in 1971. And since 1987, it has consistently generated free cash flow. This has been the key to enabling the relentless and necessary levels of research and development required to stay relevant in the semiconductor industry.
But the tech giant has struggled for the last decade. The stock is currently trading at levels we haven’t seen since 2014. And while Intel’s stock rose close to 150% between 2014 and its most recent high in early 2021, that compares with the thousands of percent increases in the share prices of competitors like AMD and NVIDIA.
What happened? Intel missed the boat. It completely missed the shift to mobile computing and smartphones. And an even bigger miss, if that’s possible, has been the shift towards artificial intelligence (AI) and machine learning. That’s the reason why companies like AMD and NVIDIA have done so well. They were ahead of that trend, not behind it like Intel has been.
During the first two years of the pandemic, Intel’s management team boasted about how well the business was doing. That was primarily based on a resurgence of newfound personal computer revenues saying the “PC is more essential than ever.” Of course, that was just temporary as the world had to adjust quickly to remote work. It wasn’t meant to last, and what happened next wasn’t pretty.
Since its recent high in April 2021, Intel’s share price is down more than 60%. This is a surprising drop for one of the greatest blue-chip stocks of all time. It was always seen as a safe place for capital because of its decades of free cash flow generation and consistent record of raising quarterly dividends.
But 2022 was different. For the first time since 1987, Intel generated a negative free cash flow of -$9.6 billion. Ouch. And the next two years don’t look promising either with a combined negative free cash flow of about -$10 billion.
And it just got worse. Intel recently announced that it cut its dividend by 66% from $0.365 a share to just $0.125 a share. Given the negative free cash flow, this was a necessary move, but it makes Intel a lot less attractive to institutional capital.
While Intel has plenty of cash and the ability to raise even more debt, the reality is that it will have to go through a multi-year period of spending tens of billions in an effort just to catch up with the rest of the industry.
And given the consistency with which Intel has missed its production targets for its new products, and brought out lackluster products to market, it’s far from certain that Intel can catch up with giants like Taiwan Semiconductor Manufacturing Company (TSMC) and Samsung on the foundry side, or AMD and NVIDIA in cloud-based computing and artificial intelligence.
Where Intel is focused on new products, the industry is shifting away from Intel semiconductors. The major providers of hyperscale cloud services have been migrating to custom-designed semiconductors for years. This trend will continue.
And this latest explosion in development in AI will result in the need for even more AI-centric application-specific integrated circuit semiconductors (ASICs), a product category where Intel is extremely weak.
Worst yet, the big shift that will take place over the next five years will be towards the edge of the network. As semiconductors for mobile and remote devices become more powerful, and software code is optimized, software will run more and more at the edge (example: on smartphones) than in the cloud where Intel’s CPUs power servers run.
Intel’s acquisition of Altera for nearly $17 billion back in 2015 was in part an effort to address Intel’s weakness in this area. The acquisition has not been managed well, and Intel’s current products for edge computing remain particularly weak.
A stock like Intel might look “cheap” in a market like this, but with massive cash flow burning for the next couple of years, intense levels of necessary R&D spend, and a recent history of weakness in new product development, it’s hard for me to get excited about the next few years given the current competitive environment.
Growth will soon return with a fervor in the semiconductor industry, but sadly, I won’t be looking at Intel.
A personalized generative AI for all…
Yesterday we talked about how Microsoft just pushed generative AI to the edge of the network – onto our smartphones.
Of course, it all started with OpenAI’s ChatGPT. This is the generative AI that can produce content and have intelligent conversations with humans.
At first, ChatGPT was only accessible through a web browser. Then Microsoft integrated it into its Bing search engine. And as we discussed yesterday, Microsoft’s latest move was to incorporate ChatGPT into mobile apps like Skype.
This puts generative AI onto mobile devices operating with a mobile-friendly user interface… but the AI still relies on cloud computing for processing. When users make a request, it’s sent to the cloud, processed, then sent back to their device.
The significance here is there’s a slight delay when interacting with the AI. What’s more, there are cloud computing costs involved.
But that’s all about to change.
Semiconductor powerhouse Qualcomm just published some interesting new research. The company demonstrated a semiconductor capable of running generative AI on a smartphone with no ties back to cloud computing resources.
Specifically, Qualcomm showed that its new semiconductors could run text-to-image generator Stable Diffusion right on a phone with no cloud computing support.
Naturally this eliminates the delay we experience when an inquiry has to be sent to the cloud for processing before it’s returned to our phones. But arguably more important, when the software resides and runs on our mobile device, no cloud computing costs are required.
Little or no costs means even more rapid adoption of a new technology. We’re moving closer to a world where everybody has their own super-intelligent and personalized digital assistant. This is something I’ve been predicting for years now. Late last year, the technology was already impressive. But with these latest developments in large language models (LLMs), the technological capability has now jumped into hyperdrive. What we’re about to see will be a dramatic upgrade to the digital assistants like Apple’s Siri, Google’s Assistant, or Amazon’s Alexa that we’re used to.
Just as an example, Apple’s Siri can do basic tasks: Set alarms, check the weather, read text messages, play songs, etc. But its utility drops off quickly from there.
A personalized digital assistant powered by generative AI will be significantly more useful. It will be capable of performing tasks for us that are menial and time-consuming throughout our days. It will understand our lifestyles, habits, and even our needs. Knowing this, it will be able to proactively offer support and recommendations at precisely the time that we will benefit from them. It will be a remarkable feeling, to feel like we have a high-powered assistant catering to our daily needs, performing valuable tasks, and saving us time throughout the day.
And that world’s not far away.
OpenAI’s GPT-4, its next version of a large language model, is already in beta testing. And a handful of companies are already working to further productize this kind of intelligent technology targeted at consumers. We’re going to see both the hardware and software rapidly develop by the second half of this year, enabling services most of us wouldn’t have imagined just a few years ago.
The quantum teleportation of energy…
There’s been an absolutely incredible development in the world of quantum physics.
Fifteen years ago, a Japanese theoretical physicist made what was considered to be a ridiculous claim. He said that energy could appear in a vacuum… essentially out of nowhere.
The industry largely laughed at this. The thought of energy appearing out of nowhere in a vacuum was thought to be fringe science at best.
But this physicist was dead serious. And he conducted years’ worth of research to test his hypothesis.
Well guess what? His research has been proven valid twice over the last few weeks.
Physicists are calling this principle the quantum teleportation of energy. And as crazy as it sounds, it’s actually a rather simple idea. And the principle maintains the conservation of energy.
The research shows that, due to the nature of quantum mechanics and quantum entanglement, it is, in fact, possible to pull energy from one location and teleport it into another.
Now, this has only been done in a highly controlled lab setting. It’s not something we can go out and start doing in the real world immediately. But the theory has been proven.
We talked about the Universe and its formation on Tuesday… well, one of the biggest questions we haven’t been able to answer is “how did it all get here?”
Sure, we’ve got theories such as the Big Bang that posit how our Universe was formed… but we can’t explain how the Big Bang came about. Where did the energy and matter come from?
Perhaps this principle of quantum teleportation has something to do with it all.
More practically, it’s easy to envision how this breakthrough could be a boon for human civilization.
Imagine a world in which we all have devices that teleport energy from an extremely energy-dense location – say a black hole – and then these devices teleport that energy into a vacuum where it can then be used to produce clean electricity.
I know this sounds like science fiction, or perhaps magic to some, but something like this is possible if we can figure out how to teleport energy at scale using quantum physics.
This is how so many crazy breakthroughs happen. Someone comes up with a “crazy” theory, it’s proven, and then in the years that follow, others figure out how to bring the idea to life and utility. Of course, we’re a long way away from being able to do that. But now we know it’s theoretically possible.
It’s not out of the realm of possibility that this kind of technology is developed within our lifetimes. Hopefully, we can develop it without creating a tear in the space-time continuum. I’m half joking of course, but we don’t know what impact the teleportation of large amounts of energy will have on the fabric of space-time.
While I’ll continue to monitor developments in quantum physics, our far more near-term solution to limitless, cheap, and 100% clean energy will come in the form of nuclear fusion, which is an area that we’ve been watching closely in The Bleeding Edge.
A central bank digital currency is coming to Europe…
A few days ago, the European Central Bank (ECB) put out a presentation that details its plan to roll out a central bank digital currency (CBDC) for the European Union (EU).
As a reminder, a CBDC is a digital currency designed to mimic national fiat currencies. The difference is, they aren’t held in a bank account. Instead, they are held in a digital wallet that is fully controlled by a central bank.
The ECB has been working on its own CBDC for about two years now. And its latest proposal suggests the CBDC could be ready as soon as the third quarter. It’s almost here.
And the ECB plans to take a systematic approach to implementation.
It will start by initially focusing on person-to-person transactions. This is the equivalent of Venmo or Cash App payments between individuals.
The ECB also wants to initially deploy its CBDC for e-commerce transactions. It will provide the infrastructure necessary for e-commerce shops to accept the CBDC in exchange for goods.
From there, the ECB will migrate its CBDC to physical stores with point-of-sale (POS) systems. Then eventually to government payments.
So they’re going to tactically focus on innocuous day-to-day transactions first to drive user adoption. The idea is to get people comfortable with the new monetary system and how it works.
But there’s a dark side at work here…
It’s no secret that certain factions in the EU want to implement carbon footprint trackers, and potentially even social credit scores, into the CBDC system and consumer digital wallets.
This would entail creating incentives that encourage citizens to engage in those behaviors the government deems desirable. And it would punish those who do not go along with the establishment’s narrative.
In other words, the EU wants to import something similar to the social credit system that communist China has already rolled out. We’ve heard reports of Chinese citizens being restricted from travel and routine purchases as punishment for not behaving in ways the regime wants.
CBDCs could enable governments to become even more authoritarian, tyrannical, and hostile towards citizens. This is something we’ll want to keep a very close eye on.
Editor, The Bleeding Edge