Well, I’ve been promising a lot of continued volatility, and that’s what we’ve been getting.
And, unfortunately, I was right about my prediction of an escalating war in the Middle East. The Israel/Iran conflict is worsening, and the U.S. is now involved, whether it likes it or not. We can only hope for a deal that will bring the war to an end. Oil is now trading above $78 a barrel as a result.
Domestically, in a bewildering but expected move, Fed Chair Powell kept the Fed Funds rate at 4.5% again despite what the data is telling us. A 50-basis-point drop would have been easily justified, and yet the Fed is leaving rates artificially high. This will continue to negatively impact the housing market as well as the Treasury’s ability to manage its debt and interest expenses.
It hasn’t all been bad, though.
The U.S. Senate passed the GENIUS Act stablecoin bill.
The “Guiding and Establishing National Innovation for U.S. Stablecoins of 2025” – or GENIUS Act – is a bill that lays out a federal regulatory framework for payment with stablecoins.
It clearly defines stablecoins and discusses a licensing framework and some protections for them, such as reserve requirements, disclosures, and oversight.
We’ve been tracking the bill’s progress with regular updates in The Bleeding Edge. And now, it has passed the Senate. This is incredible news for the digital assets industry. It now goes to the House. We’ll see what happens there, but the Act is on track for signature before the August recess.
And the artificial intelligence (AI)-specific semiconductor industry is still on fire right now. Marvell held a very bullish AI event this week with nothing but good news. It speaks to the incredible strength in continued investment and spend in AI infrastructure, as well as the exponentially expanding number of AI application-specific semiconductors.
This was also the topic of one of our questions in this week’s AMA below regarding thermodynamic computing systems.
Something very special is happening right now in the world of high tech. I’ve never seen a confluence of technologies and the pace of technological development that we’re seeing right now. Opportunity is everywhere.
We have so much to look forward to.
Now, onto our AMA…
Hello,
I hope this is the correct email to send general questions to, but if not, please let me know where to properly send queries.
I am an Exponential Tech subscriber and am curious about your thoughts on Thermodynamic chips. I have read a few articles about a startup called Extropic, which is working on this, and it seems they are pretty close to having some sort of product ready.
It would obviously take time to scale, and it calculates things differently, so it seems like it is more of a complement to traditional GPUs (at least to start).
Do you know anything about these and what sort of timeline it all may be on? And the effects it might have on the AI supply chain of semis and including the sort of picks and shovel names that are in the Exponential Tech portfolio?
Keep up the good work!
Best regards.
– Kevin G.
Hi Kevin,
You found the right place…
Even better, I’ve been invested in Extropic (a private company) since its very first funding round and have been following the company’s progress very closely.
Extropic is one of the most interesting semiconductor companies that I’ve seen in the last few years. It’s been around since 2022 and was in stealth for quite some time, but recently got some more widespread attention in a May article in Wired.
At a high level, Extropic’s thermodynamic computing is easy to understand. Its accelerators will, in theory, be able to use thermal fluctuations to perform probabilistic calculations at speeds significantly faster than digital processors (i.e., GPUs, CPUs) and at dramatically better power efficiencies.
It’s a fascinating idea that is radically different than the current framework for artificial intelligence-specific semiconductors, which is why I’m so interested in what Extropic is doing. But to be clear, that doesn’t mean Extropic will be successful.
This is a deep tech research and development company working on something that has never been invented before. Extropic may very well be unsuccessful in commercializing its technology, but it is worth following.
Its first prototypes were developed in 2024. They were nano-fabricated out of aluminum, were very tiny, and needed to be operated at extremely cold temperatures to achieve superconductivity. This was required to more accurately measure the temperature fluctuations of its thermodynamic computing system.
But a quantum-like system requiring extremely cold temperatures is not very scalable, which is why the company has been hard at work on a room temperature semiconductor design that will act as an accelerator for AI applications.
Extropic’s Thermodynamic Accelerator Prototype | Source: Wired
Extropic’s current prototype, shown above, is a room-temperature CMOS semiconductor, which means it uses standard semiconductor manufacturing processes and is highly scalable.
A CMOS architecture ultimately sacrifices energy efficiency but provides a path towards scale. And with a market showing the kind of explosive growth like AI-specific semiconductors, that’s the correct path to move forward.
Extropic has said that it is going to be making its “Alpha” product available during the summer of 2025, which is now. But that also means that it will be very selective in terms of which companies it will share the technology with.
I have zero insider knowledge on this point, so I’m only speculating here. The purpose of releasing an alpha product is twofold: to collect feedback from key strategic potential customers and to get investors excited (presuming Extropic can demonstrate the tech works or has a believable path towards working).
Again, I’m speculating here, but assuming the alpha launch goes well, I fully expect that Extropic would raise a large venture capital round, develop a beta product by the end of this year, and target general availability of its first commercial product around the second quarter of 2026.
It’s important to note that this is a very accelerated time scale that I’m speculating based on public comments from the founder and how quickly the company is moving.
As for your specific question about our model portfolio in Exponential Tech Investor, there will be no short- to mid-term impact on the portfolio at all from the developments at Extropic.
Extropic is still in research and development mode, needs to raise a lot of additional capital, has no commercial product, and has a material possibility of failing. This is not to mention that it will take years to scale up manufacturing in a way that will have any material impact on the overall AI-specific semiconductor industry.
None of this is a criticism of Extropic. It’s just the realities of designing, testing, and, ultimately, contract manufacturing a radical new semiconductor design.
I certainly hope that Extropic is successful. After all, a magnitude improvement in performance and dramatically better energy efficiency would be great for the AI industry. But no matter what, it will be a tough road ahead for the company.
In the meantime, we’ll stay focused on under-the-radar, small-cap companies that are well established, generating healthy free cash flows, growing exponentially, and serving markets that are seeing incredible growth.
Jeff,
Thanks so much for your constant insight into the world of technology. I have been wondering about the use of Optimus as a personal assistant. I assume that, as such, I would provide direction verbally. My interface is established. In addition, I am assuming that Optimus will be connected somehow to xAI. (Or could be.) So, in this case, why couldn’t my interface with xAI be Optimus? I know there are a myriad of issues that would have to be addressed, but do you think that Elon and his crew are considering some aspects of this for future iterations of Optimus? After all, they seem to be long-range thinkers.
Just a thought.
– Carl R.
Hello Carl,
I like the way you are thinking.
And basically, I think you’re right. One way we can think about Optimus’s software is to break it down into two components:
So, it makes perfect sense that xAI’s latest developments will be accessible with Optimus to perform tasks, engage in conversation, act as an encyclopedic knowledge base, have a personalized memory of its conversations with us, and become an incredibly useful personal assistant.
It’s happening, in the very near future.
In the coming months, I fully expect a new update from Tesla about Optimus Gen 3 and all its incredible capabilities. Tesla is already mass-producing Optimus in its Fremont, California, factory, and using Optimus in its offices and factories.
Next year will be a major production year and also when Optimus will be made available to the general public.
As a faithful subscriber, I want to say I love your content. It would be of great interest to me to see your research and to read your analysis of the emerging Enhanced Geothermal Systems (EGS) industry.
– Allen D.
Hello Allen,
Thank you for the compliment, it’s great to have you as a subscriber. I appreciate the question as well. I’m pretty sure this is the first one I have received about Enhanced Geothermal Systems (EGS).
For everyone’s benefit, EGS is an emerging form of geothermal energy that works similarly to today’s geothermal energy. The major difference is that EGS doesn’t require a site above any kind of natural geothermal resources.
EGS doesn’t require a natural hydrothermal reservoir to operate, which greatly expands the number of locations where EGS technology can be deployed. All EGS needs are very hot rocks to create an artificial reservoir used to great geothermal energy.
Source: U.S. Department of Energy
To develop an EGS, it is necessary to drill typically to depths that range from 3-10 kilometers to find rocks around 200 degrees Celsius in temperature.
The U.S. Department of Energy has already funded an EGS project called Utah Forge, which is managed by the Energy and Geoscience Institute of the University of Utah.
Utah Forge EGS Site | Source: Forbes
No breakthroughs are required to commercially develop EGS power plants. And the process is pretty straightforward:
So, given that there are no major technological hurdles, this really comes down to regulatory support for the drilling of EGS power plants and the economics of doing so.
Earlier this year, I saw a report from Stanford that forecasted the costs for EGS around $80 per megawatt-hour by 2027. That would put EGS on par with average electricity prices in the U.S. At this cost, EGS would produce electricity at prices double those of the cheapest areas of the country, but would be within a normal range of some of the more expensive parts of the U.S.
EGS’s largest competition, in terms of clean energy, will come in terms of upfront capital expenditures. EGS is many multiples more expensive than solar, and also compared to wind.
Neither wind nor solar is perfect, which is why they are often paired with batteries, and also supplemental to any power grid. EGS’ benefit would be the ability to produce baseload power, assuming regulatory support and capital to fund these projects.
Fervo Energy (private) is the most well-funded player in the EGS space. It is the most advanced and uses horizontal fracking technology to target a 400 MW plant in Utah by 2028.
Commercialization timelines are not far off from those of the most advanced nuclear fusion companies. Nuclear fusion will probably cost about two to three times more in capex per megawatt, but it will not carry any of the geological risk, that being seismic activity and performance of the geothermal well.
Nuclear fusion’s risk is technological and operational. I believe that both will be overcome very quickly, but it remains to be seen how operationally efficient fusion reactors can be run.
Regardless, both technologies will be pursued as I believe both will receive funding to commercialize.
In the end, I do believe that the power of the sun will win out at scale, though.
I hope everyone has a great weekend.
Jeff
P.S. I also wanted to put on everyone’s radar an upcoming opportunity from my colleague, Larry Benedict…
Next week, he’s revealing a strategy he’s developed to profit from Trump’s market chaos using a secret calendar of 52 pre-scheduled government events he calls “Trump’s 24-Hour Profit Windows.”
Each single-day window gives you the opportunity to profit using just ONE ticker, over and over again.
He’ll get into all the details on Thursday, June 26, at 2 p.m. ET. Just go here to automatically add your name to Larry’s guest list.
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.
