Tweet épinglé
Longview Research
273 posts
Longview Research
@Longviewres
Technology researcher with a focus on AI power and infrastructure. Views are my own. Not investment advice.
New York, NY Inscrit le Nisan 2026
96 Abonnements329 Abonnés
@sambath47 Also if 3MW was the real max, why would the company mgmt and investor presentations be even entertaining a data center deal over 3MW? Seems kind of offbase for mgmt and/or Monarch to recap a company with that kind of limit per site.
English
this is not the new form factor product they have developed that is 800VDC standard, this is the prior gen product spec that is AC. The 800VDC C250S that can go to 5x C1250S is 1.25MW in lab tests with C1000S at 1MW already developed. They dont have a product page up for the 800VDC products since they arent shipping yet.
English
Very important to review emissions and efficiency of turbines. I am starting to believe that gas turbine popularity for AI deployments may drop dramatically given backlog lead lead times and them really being bridge power solutions, not permanent durable power sources. This is why $CGEH and $BE reacted so sharply to newsflow this week.
English
@Longviewres $BWA will likely be a similar name this year. They’re starting to make turbines in collab with Endeavor energy who have leased a data center to $CRWV. 2027 capacity of 2GW with revenues expected to starting coming in 2026.
English
$CGEH thoughts. Report on track for Monday. My report will include more than 10 data tables comparing $BE, $CGEH, and $GEV mainly for product specification and base power deployment installation (the perspective that matters to a hyperscale/colocation buyer). It has taken me months of reading to collect all this data hidden on company websites and product spec sheets to gather (not even web search through AI pro plans helped in this aggregation, too many misfires in the data collection). Please follow me to get this updated report Monday. I am also trying to figure out how to democratize the $CGEH model I built and I may have found a way to do do this using an AI Tool (like Gemini) with some prompts. If I get to a few thousand followers after this $CGEH long-form equity research report I will publish the prompt to build it and all the files needed to upload to produce the model. As long as you know what you are building it is easy to prompt for it. Also, if investors have questions on $CGEH before this report comes out please comment below so I can track, I already have a follow up report I need to write that sheds more light on the NIMBY (Not In My Backyard = NIMBY) dynamics that seem to be getting amplified by the day given the K-shaped economic dynamics of AI/data center development to non-tech employees. Read this punchy read for the preview linked below. Not investment advice, do your own due diligence, know what you own. x.com/Longviewres/st…
English
@_LittleFire1 I think people finally taking a look at this one, but the trading volumes are not institutional yet. $BE probably drove the entire modular AI Power peer group higher, hence the attention this past week.
English
Read the article and waiting for the deep dive, thanks!
Microturbine can only generate 3 MW even after packaging multiple turbines together.
This leads to additional space and maintenance to host 100s of micro turbines on the site.
Not sure if hyperscalers would be really interested in this.
English
@TSteenstrup Let’s get the big report out first and we can discuss this
English
@Longviewres Would be great to see you do a livestream on this stock and we could ask questions. Or send me a DM I’d like to get on a call
English
I dont even think it has peaked... I actually think this is the first move of the structural breakout as it absorbs the total addressable market away from gas turbines and other lower quality bridge power solutions for AI Power. Buckle up! also see $CGEH, it is a mini $BE, on my profile I outlined the key areas to focus on.
English
It is not as bad as you would think after the Monarch Recap. the Monarch Recap removed the Goldman controls which is a huge overhang removed. They even got balance sheet capital true up to handle major MW microturbine orders... So really the dilution is not that bad. This stock still trades below 2x sales, $BE trades at 12x. $GEV at about 3.5-4x Sales... so in a crap scenario you are up 50-100% on just sales multiple arbitrage spread. Worth the risk-reward.
English
@Longviewres What do you think of the dilution overhang? Necessary price to pay?
English
@_LittleFire1 Really not a fan, it is has struggled at every major tech cycle entry to become relevant, that is not going to change cause the mgmt team and IP has not changed. They make big announcements, but fail to execute or deliver. Not a fan of $FCEL.
English
I might do this in a later report, I view all the diesel players as bridge solutions, nobody actually wants this permanently in their power installations, its just simply what is available. Same applies to supersonic engines that are also somehow getting deployed for AI Power. I describe these as very low quality bridge power solutions. The structural AI Power longs are $BE and $CGEH. Look at BE, that is the analog of $CGEH minus the ultra sexy Silicon Valley aura.
English
@grackychan You are not too late, just wait a couple more months to let them get things cooking with the factory layout redo to support 7 MW per week of output.
English
@Longviewres This name has come onto my radar from your posts, too bad I sat on my hands for two weeks since. Mcap is still in infancy stage though, so tons more upside. Look forward to your research.
English
Go for it, you will lose your shirt this summer. The press release with $ORCL from last Monday sets a new sector structural precedent: clean power available now or data center deployment cannot commence. If this is the market analog, $BE will become sponsored by every hyperscaler explicitly and they have capacity to service the demand versus the alternatives. Good luck, but yes go for it short it so you can post your tears here later. Again, you are fighting a STRUCTURAL AI POWER thesis now, not a vibe.
English
How about this, you seem like a smart plugged-in poster on this area of the power market. Give the Monday report a read, LMK your thoughts, I think you will find that the entire thesis strung together with all its components may inform a different view with less resistance to equity sponsorship. If your opinion shifts, would love to hear why.
English
@Longviewres Not sure what you mean Bloom doesn't have absorption chiller technology. See here -
@outspokengeek/note/c-214343765" target="_blank" rel="nofollow noopener">substack.com/@outspokengeek…
@outspokengeek/note/c-206562448" target="_blank" rel="nofollow noopener">substack.com/@outspokengeek…
Air permits are not just about NOx. Need SOx disclosure. I agree they might be better than other turbines and can use dirty fuels better.
English
@OutspokenGeek I stand corrected, you are right they do have absorption chillers. I may have mixed $BE with another fuel cell company when I was checking this a few days ago.
English
Longview Research@Longviewres
English
I've started to broadly map out the power space with some good posts by @NuttyCLD
First thing to understand is that the transition to 800V requires datacenters to be built from the ground up for true HVDC support.
The mid-term approach will be to use power sidecars to handle large power levels per rack with high voltage. Longer term, datacenters will distribute power = no sidecars.
The whole power situation can be monitored by looking at four umbrellas
1. WFE (Aixtron, Veeco)
2. Power Devices (Infinion, TI, Navitas, Power Integrations, ST, ON, Wolf)
3. Box builders (eaton, vertiv, schnieder electric)
4. Hardware providers (APH, TEC, Rosenberger)
Most interestingly, the EV supply chain now has a massive AI DC pivot, and a lot of them will be looking to pivot.
English
You are gravely mistaken, 800VDC standard was not originally architected by EVs, but it was originally $BE that was trying to standardize the power architecture but at the time everyone's reaction was "why would we need this much power?!..." Then EVs saw it and thought "what a great way to accelerate charging..." so it picked up with EVs. This same conversation has resurfaced for $NVDA Kyber and Vera Rubin GPU rack systems since they use so much power density. Two companies that are key architects of the data center 800VDC architecture are $BE and $CGEH. $BE is well known, no surprises, but $CGEH is flying way under the radar, see the quick punchy pitch here: x.com/Longviewres/st…
English
The whole 800V thing in datacenters comes from EVs which use a series-parallel combination of batteries to produce 800V.
Datacenters use that value because electronics is availalbe.
Same physics at play here.
Higher voltage = lower current, for same power = lower losses
In EVs, this gives better driving range. In AI datacenters, is is better power efficiency, better TCO.
I have a whole post on EV traction inverters, how they work, and where SiC plays a role ... if you are interested.
open.substack.com/pub/viksnewsle…
Vikram Sekar@vikramskr
I've started to broadly map out the power space with some good posts by @NuttyCLD First thing to understand is that the transition to 800V requires datacenters to be built from the ground up for true HVDC support. The mid-term approach will be to use power sidecars to handle large power levels per rack with high voltage. Longer term, datacenters will distribute power = no sidecars. The whole power situation can be monitored by looking at four umbrellas 1. WFE (Aixtron, Veeco) 2. Power Devices (Infinion, TI, Navitas, Power Integrations, ST, ON, Wolf) 3. Box builders (eaton, vertiv, schnieder electric) 4. Hardware providers (APH, TEC, Rosenberger) Most interestingly, the EV supply chain now has a massive AI DC pivot, and a lot of them will be looking to pivot.
English
Very good point came up by another X poster. See below for sneak peak at one of the data tables that are expanded upon in the Monday report, this post on Air Permit triggers associated with Title V. x.com/Longviewres/st…
Longview Research@Longviewres
You bring up great points, but any emissions output compared to $BE is going to be extreme since you are comparing to near zero. But I would push back: $BE does not have this Absorption Chiller tech because they do not have heat outlet exchangers and nobody else has a microturbine system up to 1MW. In my Monday report below is the following table which actually outlines how despite having combustion emissions, the volume of emissions isn't high enough to trigger Title V air permit requirements up to certain MW sized installations. This allows buyers/hyperscalers/developers to not file Title V air permits and therefore can get other approvals fast for up to 380MW installation deployments of $CGEH. Now importantly, $CGEH should not be 100% of a 1GW data center campus, this would be a build mistake, they should only be one-third maximum to leverage the Absorption Chiller tech integrated into Energy Surplus Program (ESP) and thereby lower PUE from 1.3 ballpark down closer to 1.15 or below (both these calculations and explanations outlined in my Monday report). And let other power sources handle the other two-thirds of power installation. Keep the Qs coming sir, people think $CGEH is some widget power company... they are mistaken, the math and resistance to deployment is materially stronger than everyone realizes. People aren't absorbing this stock because they are so fixated on either 1) small cap on OTC Markets or 2) it went into bankruptcy.
English
@OutspokenGeek Tagging @pernasresearch for them to review my math/framing. Would be open to feedback from the early pros on $CGEH.
English
You bring up great points, but any emissions output compared to $BE is going to be extreme since you are comparing to near zero. But I would push back: $BE does not have this Absorption Chiller tech because they do not have heat outlet exchangers and nobody else has a microturbine system up to 1MW. In my Monday report below is the following table which actually outlines how despite having combustion emissions, the volume of emissions isn't high enough to trigger Title V air permit requirements up to certain MW sized installations. This allows buyers/hyperscalers/developers to not file Title V air permits and therefore can get other approvals fast for up to 380MW installation deployments of $CGEH. Now importantly, $CGEH should not be 100% of a 1GW data center campus, this would be a build mistake, they should only be one-third maximum to leverage the Absorption Chiller tech integrated into Energy Surplus Program (ESP) and thereby lower PUE from 1.3 ballpark down closer to 1.15 or below (both these calculations and explanations outlined in my Monday report). And let other power sources handle the other two-thirds of power installation. Keep the Qs coming sir, people think $CGEH is some widget power company... they are mistaken, the math and resistance to deployment is materially stronger than everyone realizes. People aren't absorbing this stock because they are so fixated on either 1) small cap on OTC Markets or 2) it went into bankruptcy.
English
Also on NIMBY, this is just not great for gas turbine companies and the only reason it may not be reflected stock prices yet is because of the backlogs that go through 2030. But after that, I would be concerned on $GEV, Mitsubishi Heavy and Siemens Energy. Even $CAT might be at same risk given engine sources/emissions.
English