The Principal’s Edge

$1M → $75M+ARR/Exit. Operator turned investor. Using AI as my second brain to find asymmetric opportunities across AI infrastructure. Capital. Conviction. Edge.

[The Principal’s Edge @JRConviction]

I’m not building a portfolio just to watch a number on a screen get bigger.

I’m building freedom.

Freedom to control my time.
Freedom to be present with my family.
Freedom to walk away when I want to.
Freedom to live life on my terms.

The vehicle is a concentrated AI infrastructure portfolio.

The goal is to build this particular port to $5M+ by 2030.

Current 2030 portfolio framework:

NBIS ~ 32%
MU ~ 20%
SKHY ~ 8%
CRDO ~ 5%
NVDA ~ 6%
AVGO ~ 5%
FN ~ 2%
BE ~ 4%
CRWD ~ 6%
Cash ~ 12%

Total ~ 100%

Not through copy trading.
Not through chasing every ticker.
Not through pretending risk doesn’t exist.

The strategy is simple: own high-quality businesses where the fundamentals support the thesis, concentrate when the opportunity is asymmetric, keep cash available, use options strategically, and be willing to change my mind when the facts change.

AI is my second brain.

It helps me research deeper, challenge my own assumptions, track the portfolio, model outcomes and make better decisions ~ but the capital, conviction and accountability are still mine.

And as important as the portfolio is, money isn’t the scoreboard.

For me, life comes down to the four Fs:

Fitness ~ build a body capable of enjoying the life you’re working for.

Finances ~ build enough freedom that money becomes a choice, not a constraint.

Family ~ success means very little if the people you’re doing it for only get what’s left of you.

Faith ~ stay grounded in something bigger than your portfolio, career or net worth.

That’s what I’m building toward.

Capital. Conviction. Edge.

[The Principal’s Edge @JRConviction]

This is exactly why I try not to get married to any ticker and I have done it in the past.

$IREN can still have real value in power, land and bare-metal infrastructure while also having legitimate weaknesses in managed cloud. Both can be true.

Dismissing independent user feedback as a “hit piece” because it conflicts with the thesis is how investors stop doing research and start defending a position.

The job is to update the thesis when the evidence changes ~ not attack the evidence.

[The Principal’s Edge @JRConviction]

Information is supposed to strengthen or weaken your thesis along the way ~ not be selectively accepted or dismissed depending on whether it makes your stock look good.

I own $NBIS with high conviction, but I’m not married to it. If new information challenges my thesis, I want to understand it, not defend the ticker.

Conviction without intellectual honesty eventually becomes a cult.

[The Principal’s Edge @JRConviction]

This is exactly why I try not to get married to any ticker and I have done it in the past.

$IREN can still have real value in power, land and bare-metal infrastructure while also having legitimate weaknesses in managed cloud. Both can be true.

Dismissing independent user feedback as a “hit piece” because it conflicts with the thesis is how investors stop doing research and start defending a position.

The job is to update the thesis when the evidence changes ~ not attack the evidence.

[The Principal’s Edge @JRConviction]

This is the part of the $NBIS thesis I think the market still underestimates.

Most people model Nebius like this:

MW × GPU rental rate

That may end up being too simplistic.

SemiAnalysis recently modeled that serving open-source models can generate $100M+ per MW per year at the inference layer. That does not mean Nebius automatically captures $100M/MW ~ but it shows how much economic value can exist above raw infrastructure

Now combine that with what SemiAnalysis just said about Nebius:

~ Platinum-tier managed cloud
~ premium pricing
~ rising revenue/MW
~ strong customer feedback
~ short-term capacity auctions
~ prepayments that can approach 100% on 1-year commitments Pasted markdown(4).md

That’s why I keep coming back to this:

The real upside for $NBIS may not be in simply renting GPUs.

It may be in moving up the stack:

power → GPUs → managed cloud → inference → tokens → agent infrastructure

If Nebius keeps executing and captures more value per MW while scaling active capacity, the numbers can get very large, very fast.

My current bull-case path:

2027 ~ $500+
2028 ~ $850
2029 ~ $1,100
2030 ~ $1,300+

Now for the crazy part.

If Nebius eventually proves it can monetize several GW of active compute at $40M-$50M+ per MW, while pushing margins higher through inference and software, then a $500B-$1T market cap stops sounding mathematically impossible.

Depending on future dilution, that kind of outcome could put the stock somewhere around:

$3,000-$3,500+ per share

Again ~ not my base case, not even my normal bull case.

But the math shows how extreme the upside can get if Nebius evolves from a neocloud into a full AI infrastructure + inference platform.

The KPI I care about most now:

Revenue per active MW + margin per MW.

If both keep rising while capacity scales, that’s where the real asymmetry lives.

[The Principal’s Edge @JRConviction]

@DeepValueBagger Wen deal.

[The Principal’s Edge @JRConviction]

This is the kind of independent validation I care about.

SemiAnalysis moved $NBIS into its Platinum tier alongside CoreWeave and called Nebius “unquestionably an industry leader.”

But the bigger takeaway for me is underneath the ranking:

Premium pricing.
Strong customer feedback.
Revenue/MW climbing.
Short-term capacity being auctioned.
Prepayments that can approach 100% on 1-year commitments.
Technical execution strong enough to compete at the top of the market.

That’s the thesis.

$NBIS isn’t just adding GPUs ~ it’s building a differentiated AI cloud that customers are willing to pay a premium for.

Capacity matters.

But quality + pricing power + utilization is where the real edge is.

[Nebius @nebiusai]

"Nebius is unquestionably an industry leader with strong offerings in every category."

SemiAnalysis just gave Nebius a Platinum rating, its highest tier for AI clouds. Less infrastructure firefighting, more building.

Full results: https://t.co/BY5r1LBEfv

[The Principal’s Edge @JRConviction]

@SCHDaccumulator $5M.

At 10%, the portfolio is producing $500K/year.

That’s my definition of escape velocity ~ when my capital can replace my labor.

Everything after that is optionality.

[The Principal’s Edge @JRConviction]

This is the $META thesis in one post.

Models will commoditize. Distribution won’t.

Meta already owns the attention, the intent and the customer relationship. Muse just gives it an agent capable of turning all three into transactions.

That’s an incredibly difficult flywheel to compete with.

[Rihard Jarc @RihardJarc]

It will be hard for anyone to compete with $META Muse. You either compete on the product, distribution or price.

$META is really strong in all three categories:

- The product reviews for Muse are 40k ratings with a 4.88 score (truly rare) telling you users love the product.

- In terms of distribution, $META owns the biggest ecosystem by far.

- In terms of price $META has the most spare infra capacity + enough profits from its core to subsidize if needed.

[The Principal’s Edge @JRConviction]

@JUST_KAWS $GOOGL and it’s not particularly close for me.

Better AI positioning, better optionality, and I’d rather own the company aggressively investing into the next computing platform than the one still trying to prove where its next major growth engine comes from.

[The Principal’s Edge @JRConviction]

@DiligentPlane Fear is only an opportunity when the fundamentals haven’t changed.

Otherwise you’re just buying a falling knife with a motivational quote.

[The Principal’s Edge @JRConviction]

@GuyTalksFinance Great advice for most people. Terrible advice for everyone.

Some of us actually enjoy doing the work.

[The Principal’s Edge @JRConviction]

@SCHDaccumulator Pathetic would be being 30 and not caring.

You’ve got capital, time and the desire to improve. Keep contributing, keep learning and let compounding do its job.

The first $100K is supposed to be the hard part.

[The Principal’s Edge @JRConviction]

@jpinsights Cash doesn’t need a target allocation. It needs a hurdle rate.

Make the market earn it.

[The Principal’s Edge @JRConviction]

@hamids Different trades. $NVDA is the better business today. $AMD is the bet that the gap narrows. $MU is the bet that they both keep needing an absurd amount of memory.

I’ll take the picks and shovels feeding both. 👀

[The Principal’s Edge @JRConviction]

@FindleysFinance $NBIS. $MU. Cash.

Having $1M to deploy doesn’t mean I need to deploy $1M today.

Price matters. The market still has to earn my cash.

[The Principal’s Edge @JRConviction]

@YodaStockInvest Dividends aren’t overrated. Chasing yield is.

I’d rather own a great business compounding capital at 20% than a mediocre one paying me 5% to stick around.

Total return is the scoreboard.

[The Principal’s Edge @JRConviction]

The Edge Report ~ Sept. 23

Busy day today.

The market gave me exactly what I want ~ volatility, red tape in quality names, and a chance to put cash to work without chasing.

I made several moves:

• Sold 1x $MU Oct. 30 $980 put for $4,200
• Sold 1x $MU Oct. 30 $900 put for $2,031
• Sold 1x $META Oct. 30 $680 put for $1,835
• Sold 1x $QCOM Oct. 30 $185 put for $700
• Sold 1x $NVDA Oct. 30 $220 put for $606
• Bought 1x Jan. 2028 $NVDA $170 LEAP for $7,737
• Bought 1x Jan. 2028 $CDNS $230 LEAP for $11,250
• Added 50 shares of $CDNS at $309.22
• Sold the small $AMD common position

The big theme today was capital allocation. I’m not interested in owning every great AI company just to fill out a roster. I want capital concentrated where I think the risk/reward is best.

Right now, $MU continues to stand out. The stock is around $1,070, but I’m getting paid thousands of dollars to potentially buy more at effective bases around $938 and $880.

That’s exactly how I want to use cash.

$META is another good example. The stock has ripped from the $500s into the $700s, so I’m not chasing it. Instead, I sold the $680 CSP for $18.35 and sold a second one last week: $640 CSP collecting a $23.94 premium.

If META keeps running ~ I keep the premium. If it retraces ~ my effective acquisition price is roughly $662 & $616/share

Same philosophy with $QCOM and $NVDA.

I don’t need the stock today. I want the stock at my price.

I also increased exposure to $CDNS because the more work I do on Cadence, the more I like the setup.

EDA duopoly.
Mission-critical software.
AI accelerating chip complexity.
Strong margins.
Strong balance sheet.
A stock still well below its highs.

So I added common and a long-dated LEAP.

Meanwhile, I exited the $AMD setup.

AMD is a fantastic company, but at the current valuation I simply don’t think the risk/reward is as attractive as the alternatives.

That’s the entire point of the process. The real question is ~ "Where does my next dollar have the highest expected return?"

The portfolio now has $420K of cash working through CSPs, but every one of those puts is attached to a company I’d actually want to own if assigned.

$MU
$META
$NBIS
$NVDA
$QCOM

That matters. Premium is the consolation prize.

The shares at the right price are the trade.

Now I let the book work and volatility is paying me to wait.

[The Principal’s Edge @JRConviction]

@DeepValueBagger Testing!

[The Principal’s Edge @JRConviction]

The real story with Muse isn’t “another AI assistant.” It’s distribution.

$SHOP
$PYPL
$EXPE
Instacart
OpenTable
$SPOT
Ticketmaster
Gmail

All plugging into $META Muse.

That means Muse is starting to sit above commerce, travel, reservations, payments, music and communication as the consumer intent layer.

You tell Muse what you want. The platforms underneath execute it.

That’s the part I think the market is beginning to understand.

Amazon blocking Muse makes sense from a customer-ownership standpoint ~ but it also shows how valuable this layer could become.

If Muse becomes the place where people search, decide and transact, $META isn’t just monetizing AI through ads anymore.

It may be building a new consumer operating layer for the internet.

And that’s a much bigger opportunity than most people were underwriting a month ago.

[Tevis @FunOfInvesting]

MUSE MANIA TAKING OVER THE INTERNET

In the past few days we had:

Shopify $SHOP Shop Pay + Shopify Catalog; eligible US merchants on by default unless they opt out. Announced ~Sep 21.

Stripe / Link - Launch-day payments: instant checkout at 1M+ Link merchants; elsewhere a single-use virtual card. Muse never sees the card details; user approves the total in chat.

Paypal $PYPL Announced today: shop + check out across PayPal merchants worldwide. Deal is public; no live rollout date yet.

Expedia $EXPE Muse can hit Expedia pages for hotels the way a person would

OpenTable - Restaurant reservations (launch connector; approval before it books).

Spotify $SPOT, Ticketmaster, Gmail — Third-party app links Meta called out at launch (music / tickets / email), not commerce rails.

All partnering with $META Muse.

Meanwhile, Amazon $AMZN blocked them. I think this is the WRONG move. That's coming from a shareholder. They're swimming against the current.

Can't stop this wave!