Things Noticed

I run an active stock portfolio. Noticing when the market's story about a company stops matching the business. $tickers, moats, narrative lag.

[Things Noticed @thingsnoticed]

$INTU is down 20% and I keep going back and forth on this one.

The numbers weren't bad. They actually raised guidance. So what are people selling?
I think it's this -> TurboTax grabs your inputs, spits out a return, doesn't explain anything. A general AI model does the same job but talks back. That's a hard gap to close when you're a legacy product.

But then -> 100 million customers, 40 years of tax data, regulatory relationships built over decades. That's not easy to replicate. Maybe they just become the interface layer for AI agents rather than fight them.

They cut 17% of staff and signed with Anthropic and OpenAI the same week. I don't read that as panic. I read that as a company that sees what's coming and is trying to be on the right side of it.

Still not sure how this plays out. Doesn't feel like a company that disappears. Feels like a company in the middle of figuring out whether its moat is an asset or just a slower way to lose.

[Things Noticed @thingsnoticed]

$GOOGL raised ~$85B by selling stock. The obvious reaction is: why is a company this profitable diluting shareholders?

Because $73B in free cash flow last year sounds enormous until you see $175–185B in CapEx guidance for 2026. The AI infrastructure bet is roughly 2x the size of the entire cash engine.

And $BRK.B bought $10B in the same transaction. I guess both can be right at the same time - one says the bet needs outside capital, the other says that's exactly where they want to put theirs.

[Things Noticed @thingsnoticed]

@aleabitoreddit The market sold $NVDA execution risk but the people actually building it said timelines are accelerating, I guess that's the gap worth paying attention to.

[Things Noticed @thingsnoticed]

@CapexAndChill This one snuck up on me, every AI shopping agent is a new fee event for $MA on every leg of every split basket....

[Things Noticed @thingsnoticed]

@dariandemetri The thesis makes sense, though I'd really want to see $CRM base at a higher level first - the market needs to confirm it before the story does.

[Things Noticed @thingsnoticed]

AI expands the menu. It doesn't replace the chef

KPMG is hiring more people while deploying more AI. Not despite it. Alongside it.

The market is pricing $ADBE $CRM $INTU like the chef is getting replaced. History says the chef just gets a bigger kitchen.

[Things Noticed @thingsnoticed]

$MU needed a breather after running 200%+, I get that.

But the selloff mechanism isn't what I'm watching. Stocks that climb fastest give back the most when sentiment turns, that's just how it works.

June 24 is the only number that matters. Does $MU confirm the HBM cycle is locked multi-year, or does guidance show the first crack?

Everything else between now and then is just price.

[Things Noticed @thingsnoticed]

$INTU down 63%. $CRM down 42%. $ADBE down 31%. All beat earnings, all raised guidance on the way down.

$NOW recovered. Same sector, same AI threat.

The market prices stories about businesses, not businesses. Not sure why that still surprises anyone.

[Things Noticed @thingsnoticed]

Everyone screens $BPOP against US regional banks. Same industry bucket, same valuation framework, same peer group.

But Popular operates in Puerto Rico with no serious competition - the kind of market position US regional banks haven't had in decades. Q1 EPS came in at $3.78, beat estimates by ~15%, and the company is buying back stock while trading below peer group multiples.

I think, the market is pricing the category, not the business. Kind of makes you wonder how long that holds.

[Things Noticed @thingsnoticed]

@FinanceJack44 35x assumes $AMD Lisa Su can hit numbers that depend on hyperscaler decisions she doesn't control.

[Things Noticed @thingsnoticed]

@oguzerkan The assumption is that 4GW of $META compute is interchangeable with what others are paying 3x for... I am not sure about it.

[Things Noticed @thingsnoticed]

@SJosephBurns Being precisely wrong still requires a precise right answer to be wrong about.

[Things Noticed @thingsnoticed]

@StockSavvyShay $NBIS skipping continental Europe for the UK says more about EU AI regulation than it does about the UK.

[Things Noticed @thingsnoticed]

Disincentives over incentives -> the founder edge

Founder alignment is on every investor's checklist. But the mechanism isn't the upside.

A founder with 40% of net worth in one business can't afford to be wrong. That's consequence alignment. The attention it creates is something no bonus structure, no stock option package, no long-term incentive plan ever replicates.

Professional managers lose their job when things go wrong. Founders lose potentially everything.

[Things Noticed @thingsnoticed]

@JonahLupton The flight data keeps accelerating but $TMDX Q1 EPS missed by 50%, at some point volume has to show up in the margins.

[Things Noticed @thingsnoticed]

The market reprices stocks every second. It reprices its story about a company maybe once a year - if that.

That lag is what I spend most of my time looking at. And while semis and AI plays are selling off hard, these five names are quietly holding up - because the market hasn't updated its story on them yet:

$BPOP - priced like a mid-tier US regional bank. Operates like a near-monopoly in a territory with no serious competition.

$LGND - priced like a biotech with drug risk. Collects royalty checks from other companies' drug sales. No manufacturing, no sales force, no clinical risk after the deal is signed.

$MNST - priced on US energy drink saturation fears. 45% of sales now come from China and India -> markets that haven't reached peak penetration yet.

$PSX - being valued on cyclically depressed refining margins. The midstream transformation is already generating cash. The market is pricing the old business.

$AXSM - still being read as a money-losing biotech. Sales growing 50%+ year-over-year, three products approved across four indications, loss narrowing. The commercialization curve is bending.

One deeper look per name this week. I'll track price performance on all five and report back.

I don't give buy signals — what you do with this is your call.

[Things Noticed @thingsnoticed]

$GEV is the most interesting position in here -> it sells turbines to the utilities, not electrons at regulated prices, which puts it upstream of the margin cap everyone else in this basket is subject to.

[Sergey @SergeyCYW]

AI infrastructure is becoming a power trade, not only a chip trade

The Defiance AI & Power Infrastructure ETF, $AIPO, focuses on companies building the physical backbone required for AI data centers: power generation, grid equipment, cooling systems, construction, utilities, nuclear fuel, and AI hardware.

The fund has around $780M in AUM and holds 78 companies. Its structure is different from many AI ETFs because it is not dominated only by software or semiconductor names. Industrials account for about 56% of the portfolio, followed by technology at 19% and utilities at 16%. U.S. companies represent roughly 88% of holdings.

The top holdings show the theme clearly.

$PWR, $ETN, $GEV, and $VRT are the largest positions. Their roles are tied to transmission lines, power management, turbines, cooling, backup systems, and electrical equipment. In other words, they support the infrastructure layer behind AI compute.

$BE adds exposure to on-site fuel cells, which can help data centers bypass grid bottlenecks. $CCJ and $CEG add nuclear exposure, which is becoming more relevant as hyperscalers look for reliable, carbon-free baseload power.

The ETF also includes AI hardware and connectivity names such as $AVGO, $NVDA, $AMD, and $MRVL. These companies support GPUs, custom accelerators, networking, and data movement inside large AI clusters.

Smaller holdings like $STRL, $MTZ, $NVT, $HUBB, $VST, and $GNRC broaden the portfolio into site development, transmission construction, rack enclosures, power distribution, merchant power, and backup generation.

AI demand is creating pressure across the entire physical stack. More models require more data centers. More data centers require more electricity, cooling, transmission, and backup power.

$AIPO offers a concentrated way to track this infrastructure theme. The main risk is that many holdings already reflect strong AI-related expectations, so valuation discipline still matters.

[Things Noticed @thingsnoticed]

@danielisdizzy $GOOGL raising $80B doesn't pull cash from equity markets - it pulls from bond markets, which is a different plumbing problem entirely.

[Things Noticed @thingsnoticed]

The agricultural revolution parallel for AI

Everyone's looking for the right historical parallel for AI. 1999? 1987? The dot-com bubble?

The right comparison might be the late 1800s - when machinery replaced 40% of agricultural workers. Not a market cycle. A workforce restructuring that took decades.

The stocks are pricing a cycle. The reality might be something longer and harder to trade around.

[Things Noticed @thingsnoticed]

@Couch_Investor $MELI is deliberately compressing margins to pull users in more often.

[Things Noticed @thingsnoticed]

@cryptopunk7213 $AAPL routing prompts to better models is a distribution play, not an AI one.