NEW ALL IN*: $NVTS @ 13.95.
FIRST ENTERED & ALERTED TO SUBS LAST FRIDAY.
THESIS ACTUALLY PRETTY BIG BRAIN. NO TLDR.
---
# WHAT DO THEY DO
They make a critical chip to convert 800V to 6V directly skipping the 48V intermediary.
$3.8B market cap down -57% from ATHs with $557M cash and no debt.
Formerly known for GaN mobile chargers (cute), they're pivoting entirely to higher-margin AI and power infra.
Showcased an 800V-to-6V board inside NVIDIA’s MGX AI Factory ecosystem (aka the standard for data centers) that's targeting 97.5% peak efficiency and 2100 W/in power density.
Stock price shot up to $34 after that btw.
Now it's at ~$14.50 after bouncing off strong support line.
# WHAT IS 800VDC
AI racks are getting extremely power hungry.
NVIDIA is moving toward 1MW+ AI racks in 2027, and getting power efficiently into those racks is becoming a serious bottleneck.
This is MISSION CRITICAL for unlocking the next 10x in AI compute and NVIDIA is pushing this hard https://t.co/QFXqZDVz2k
Jukan reports timeline moved up to 2H26 https://t.co/naCjPUBqsy
800VDC feels a lot like CPO ~12 months ago: the ramp up is coming, and the stock price usually moves first.
# POTENTIAL CATALYSTS
Several potential dates, all coming soon:
- Aug 18 Rosenblatt Technology Summit
- Aug 25 Jefferies Semiconductor Conference
- Aug 26 NVIDIA earnings
- Aug 27 Deutsche Bank Conference
They also said its 6.5kV SiC tech will be unveiled “very soon” and a 10kV SiC device announcement with a lead customer “in coming weeks.”
I have no idea what that means but I like the sound of it.
# THE PLAY
My favorite swing trades are entering near support and let it bobble for a bit until some good news makes it pop.
$RAM I entered on 8/7 and it was essentially flat until explosion on 8/12.
Feels like there could be an announcement around 800V any day now, especially with $NVDA earnings on Aug 26 (only 9 days away).
So the play is to chill here until the market realizes the opportunity.
---
*As a reminder: this is my challenge account where I restarted with $45k to go all-in swing trading 1 stock at a time to $10M again.
Important disclosure: I currently own 4,611.980599 shares of NVTS and wrote this after establishing the position. I may buy or sell at any time without updating this post. This is not investment advice or a solicitation to trade securities. NVTS is highly speculative, and you could lose your entire investment. Any forward-looking statements or scenarios are speculative and may not occur. Do your own research and review company filings.
Video from last Friday. Screenshot from just now.
Switched from Claude because I got sick and tired of wasting my time on reading indecipherable language.
I think there are a couple of things I would improve:
-Ability to merge projects/chats into a new project/chat. There are a few cases where I have two separate but related chats and then I constantly have to cross reference them with the other one to make sure I'm still pulling on the same threads
-Better global search. I know I may have had a back and forth about something, but I can't find the original chat. The search function is pretty bad as it only exact matches whatever I'm searching for. I would prefer to be able to search like this: "We talked about something related to XYZ and you said ABC, but I can't find it, where is that chat?"
-I can't seem to get codex to work in Chrome. Something is wrong with the plugin but I can't figure it out. Claude was seamless on that but ChatGPT seems to have issues
Thanks!
I was shorting $SPT in the $60/share range a few years back (h/t @akramsrazor), doubled the short in the $30/share range once it collapsed and rebounded when management was spreading rumors in the media about selling themselves. Finally covered it around $12/share a couple years later
At every point during that period, people were bulled up on the company and its trajectory. Now that I trades at close to 0.5x sales and 5x EBITDA and literally less than 1x cRPO, it's crickets.
With 80% gross margins you could just fire everyone and suck out the cash here. Even with a complete AI disruption to their business over time, there are still contracts here that are going to deliver value over time and clients that will stay with them.
Honestly, hard to see how you lose (caveat that I have no idea what is going on with the business in the last year). There's a price for everything
@TidefallCapital Well, technically none of it will be dilutive where the stock price is
Unfortunately this also means no one talented will want work there any more
@Crussian17 100% hard agree
Even if he is the greatest CEO that adds trillions in value just from his existence, that just tells you how risky investing in his companies is. No one worries about that for Amazon, Alphabet, Microsoft, etc because they've built systems to avoid key man risk
I highly recommend
I built a lot of projects on Claude but it started becoming draining and time consuming to reading so much explanatory text
Fable and Opus 5.0 have just made everything worse and noticeably slower
I was easily able to import everything I ever did in Claude and Codex ended being many times better. For the most part, it is much faster and gives a better result on 5.6 Sol Max than anything in Claude
Where Claude is noticeably better is design and presentation which I think causes people to be more drawn to it's responses/creations
Also this platform is very anti Altman, so I wouldn't be surprised if good publicity around Codex is suppressed
Would you trust a CEO that just invested $150m if his own money to buy his company's stock only to see it collapse 50% based on a revenue trajectory he was likely aware of? $TTD
Seems odd that $IRTC would be allowed to acquire VitalConnect which is basically a direct competitor. The acquisition seems highly anti-competitive. Anyone have thoughts here?
Let's ban the CFO section from earnings calls
Ideally all prepared remarks will be banned, but the CFO section is especially useless. What's the point of having someone read a bunch of financial metrics that were released in the PR?
Completely useless 10-15 minutes of time
Pretty sure this is a way to sequester capital away from public markets so that they can't be redeemed
If you could keep $400m of AUM and maintain 2% management fees ($8m/year) then obviously that is something you would do rather than keep it in redeemable liquid equities
@dalibali2 Even the fact that it gives you live transcripts is incredible! If you don't like the native summary, just copy paste into your preferred chatbot/AI tool
Codex is infinitely better at everything other than design/formatting where Claude clearly wins.
I literally cannot read Claude's outputs anymore, they make me delirious
How did people get such a high opinion of Claude? Is it because it sounds "smarter" to normies because its more verbose? Is it just because Elon hates ChatGPT so he has demoted all positive comments on Codex?
The lack of commentary on Codex vs. Claude, Kimi, DeepSeek, etc. is astonishing. Codex is at least 2x better at everything and gets 10% of the media attention at this point. I'm super bullish on OpenAI now and they probably surpass Anthropic in ARR next year if not by the end of this year