AI Wonder Material?

I hope that all of my readers in the Northern Hemisphere are enjoying their summer. Mine has been extremely busy, intense, and full of challenges.

To that end, I’ll be taking some time off next week to recover.

I won’t be writing any issues of The Bleeding Edge, but we will still be producing content for you. My team will step in and be writing for me next week about a bunch of interesting topics consistent with our mission here at Brownstone Research.

I will be in touch with my team daily, though, as we monitor the markets and our model portfolio positions in real time, as always. I’m fortunate to have a fantastic team around me where I don’t have to worry about all the right things getting done.

Have a great week next week, keep in touch with The Bleeding Edge, and I’ll be writing to you next on Monday, July 28.

Sincerely,

Jeff

Graphene for AI?

Dear Jeff

What are your thoughts on whether graphene chips, with layers of indium selenide, will dramatically reduce the energy consumption of AI?

Will the technology be in mass production in the early part of the 2030s?

Regards.

– Robert A.

Hi Robert,

Thanks for writing in. I’ve actually written quite a lot about graphene in previous Bleeding Edge issues. This material is often looked at as a breakthrough material for semiconductors.

For the benefit of new readers or for anyone who wants a refresher, graphene is an incredibly thin, light, and highly conductive material made of pure carbon.

It’s 200 times stronger than steel and five times lighter than aluminum. It’s especially popular for its conductivity – it can handle heat up to 1,300 degrees Fahrenheit.

Graphene comes up in conversation a lot when we talk about heat management in semiconductors. The main problem with graphene is that, economically, it’s not a good solution. There are some technical issues as well that make it difficult to use in semiconductor manufacturing.

Here’s what I wrote recently in The Bleeding Edge – Reason to Be Optimistic About Nuclear Energy…

Graphene production as of now is just too expensive and too energy-intensive to apply it widely in the industries where we’d hypothetically see the most use of it, among other issues. I previously wrote…

Graphene can cost as much as $200,000 per ton. The industry is, of course, working towards a breakthrough that could radically reduce the costs of producing graphene. But until that happens, it is very difficult to scale production for commercial manufacturing.

And not to get too technical, but one of the reasons that graphene is so good at conducting electricity is because it doesn’t have an energy band gap.

There is an energy band gap in silicon, which is the difference between the valence band (which is filled with electrons) and the conduction band (which is an empty state). And without a band gap – as is the case with graphene’s natural state – graphene can’t stop conducting electricity.

This makes it hard to use for applications like transistors and thus semiconductors.

There are some solutions to this problem, but they have not been commercialized. And it would make the manufacturing process more complex, which means even more expensive.

It’s an incredible material, no doubt. I’m still very excited about future applications. But as of now, the economics just don’t make a lot of sense.

However, I’m always keeping an eye on the materials science industry as any breakthroughs directly impact a wide range of industries we follow here at Brownstone Research.

With that understanding, Robert, I’d like to now get to your question. Given your specificity, I’m sure you know the purpose of indium selenide is to potentially solve the band gap problem I mentioned above.

Indium selenide (InSe) is one of a handful of materials that shows promise for semiconductors at an extremely small scale, below 3-nanometer manufacturing. This is possible because of indium selenide’s low effective mass, high mobility, and highly tunable band gap.

These properties could theoretically result in a two-dimensional semiconductor capable of outperforming silicon. That’s why we hear a lot of buzz about graphene, or some combination of new materials, from time to time.

In the case of indium selenide, there are two major issues that negatively impact the manufacturing of InSe semiconductor wafers. There is a lot of phase diversity with InSe, which is difficult to control and results in significant deviation across the material, as well as degradation.

The second major problem is that it is hard to manufacture high-quality crystallinity of InSe films. This is due to the vapor pressure of selenide being about seven orders of magnitude greater than that of indium. That makes it very hard to find a balance between the two materials.

To be fair, a lot could happen within the next five years that could address these challenges. But one thing that I’m sure of is that the issue of economics will almost certainly not be solved.

Because of that, I predict that if graphene with layers of InSe is used in 2030, it will only be used for niche applications where cost is not a concern.

The reality is that current semiconductor manufacturing materials and technologies continue to make significant improvements in terms of reducing energy consumption per unit of compute.

I believe that this progress will be significant enough, and at low cost, which will mean that it is unlikely that the current course will be changed by 2030.

To unseat a technological approach, it would take a radical improvement in total performance in addition to much lower costs of manufacturing.

At the moment, working with InSe for semiconductor applications is nothing more than a laboratory experiment. But I’ll continue to watch this space and those of other promising materials, always looking for a breakthrough that just might cause an inflection point in the industry.

When it happens, you can be sure I’ll be writing about it. 

What’s Going on With Day One Investor?

Mr. Brown, are you still doing your Day One Investor service? I signed up as a lifetime member of Brownstone to receive this service. I tried contacting you through Brownridge, but the voicemail box was full. I have not seen any news for some time now.

Regards.

– Jerry P.

Hi Jerry,

I appreciate you writing in to ask about that. I have been meaning to put out an update, in addition to updates on portfolio companies. I have been keeping in touch with the portfolio company CEOs, as well as speaking with new potential Day One Investor companies.

Unfortunately, I’ve had a bit of a challenging time with my health the last several months as I’m trying to cure myself from my cancer. My therapeutic protocol has been much more difficult than I expected in terms of side effects, and that has limited my productivity. I hope to have some good news on that front in a few weeks or so.

I still have some heavy lifting to do with returning Day One Investor to full health, which I am committed to doing. I hope to have some updates for you soon. Please bear with me as I put the right infrastructure in place and return to health and my normal levels of super-productivity.

Any News on Cerebras?

Your major sales pitch in late May about the possible IPO of Cerebras was the main reason I subscribed to Exponential Tech Investor. Unless I’ve overlooked something, I have not seen any updates on the status of the IPO as to whether you think it’s still on the company’s drawing board and simply postponed, or if the company has dropped pursuit of it. I strongly suggest you periodically update subscribers on this matter. I await your response. Thank you.

– John W.B.

Hi John,

Thanks for joining us at Exponential Tech Investor. Happy to have you on board.

We’ve had quite a few people writing in wondering about the Cerebras IPO. I’m glad to see so many readers as excited about it as we are.

I wrote about this in a recent AMA issue when someone wrote in wondering What’s Up With Cerebras…

[Cerebras] will be one of the largest and most important IPOs of 2025.

It has already chosen its underwriters, namely Citi, Barclays, and UBS. I know that the demand is through the roof, and Cerebras is carefully picking its window to go public.

There haven’t been any recent filings with the SEC that would provide us with specific details, so I can only speculate on timing and reasoning. I’ve been at this for a few decades, so I have a pretty good feel for what’s likely going on.

Companies as prominent, promising, and in a very hot sector like AI are in the best position to maximize their public offering. Good timing allows them to raise more capital – exactly the amount they want to raise.

Executives looking to take their company public tend to avoid highly volatile markets and windows with a lot of negative “noise.” There was a lot of that in the first half of the year with the new Trump administration, DOGE, Ukraine/Russia, the breakout of conflict in the Middle East, tariff negotiations, and persistently high interest rates.

Put simply, the last six months have not been ideal for taking a company public. There has been too much risk and too many distractions with everything that was happening.

But we did see some positive progress in the IPO markets this quarter, with 42 IPO’s raising about $7 billion. Two exciting tech companies, Circle (crypto) and Chime (fintech), received particularly strong demand.

If it were my decision, I’d wait until the Federal Reserve starts cutting interest rates.

This is long overdue, and a cut of at least 50 basis points is warranted at the next meeting. But then, that’s been the case for a while now, and so long as Fed Chair Jay Powell continues his misguided campaign against rate cuts, we don’t know if that will happen yet.

But if it were me at Cerebras, I’d be looking for the sector rotation into smaller capitalization stocks, which will only happen when rates start to come down.

The next FOMC meetings are July 29-30 and September 16-17.  If I had to make a prediction, I would say we might look forward to an IPO in the September/October time frame.

My team and I are tracking its progress closely. The timing for this IPO is so critical, and it makes sense that the team over at Cerebras would wait for the opportune moment to go public. I stand by my initial forecast. This fall seems the ideal window for a Cerebras IPO. But, as always, we’ve got our eyes and ears on anything that might signal an earlier move.

We’ll absolutely keep you posted of any changes and alert our readers over at Exponential Tech Investors as soon as it’s time to move.

Deep Access Update

Message for Jeff –

Don’t give up on Deep Access!

Actually, keep refining it… once the current “melt-up” is finished, probably in the next 2-4 years, the downturn will be historic. And just like the dot-com bust, with the next downturn, I’ll be looking for the most over-hyped, worthless companies (probably in the AI frenzy, I’m guessing this time) to short.

Hopefully, you’ll have those already discovered in Deep Access!

And the next step… I’m counting on you to already know the solid companies that will survive the carnage to become the next massive winners (like the Amazons) for the next market cycle. Thanks for all the work you and your team are doing!

– James R.

Hi James,

Don’t worry! Deep Access isn’t going anywhere.

We’re just holding back on new recommendations for a couple of weeks or so, looking for an opportunity to stack the deck in our favor for our next trades. Markets are overextended, and everything’s running pretty hot at the moment, which means even trades in companies with questionable fundamentals or bad news can still climb higher.

As it goes, a rising tide lifts all boats.

But the Deep Access AI is tirelessly seeking our next opportunity. As my senior analyst Nick Rokke and I wrote in our recent update to Deep Access readers…

Deep Access isn’t about swinging for the fences on every trade. It’s about staying in the game long enough to capture the big winners that come when the Deep Access AI picks up massive institutional short positions in a company, signaling that the smart money knows something is going to crack.

So for now, we’ll let this rally run. The markets are already overextended right now, and they definitely could go higher given the potential positive catalysts mentioned earlier.

We continue to track new signals daily. We’re going to be very picky in this environment. And when the time is right and market conditions simply return to more normal trading conditions, we’ll be ready.

And if there is a silver lining when conditions do get this overextended, it’s that they tend to snap back sharply when there is bad news, especially for individual equities.  And larger swings downward mean larger profits on trades.

Deep Access has been years in the making, and it’s a fantastic neural network. And just in the last couple of days, we’ve seen the signal strength on several assets start to climb.  We believe that this indicates that some institutional capital believes many assets are way overvalued now, and they are building their short positions.

There will definitely be major pullbacks even during bull markets. And there will always be companies that take a turn for the worse and get it all wrong. We’ll still be able to build profitable trades when that happens. And the Deep Access AI will be working night and day to spot those patterns for high probability signals.

We’re just getting started… And just a reminder, we’re never going to put out a trade just to put out a trade. When the Deep Access AI has a high probability signal that we believe in, we’ll get it out to you as quickly as we can.

Have a great weekend,

Jeff