Read this post.

Don’t buy the new iPhone. Buy Apple stock instead.

It sounds clever, like the owner mindset version of thrift… skip the toy and own the company instead.

Here’s the problem with that advice when you scale it.

If everyone skipped the phone and bought the stock, Apple stock would get crushed (and also, there is nothing wrong with buying things to upgrade your life).

The company is still paid by people who open their wallet for hardware and then pay for recurring Services.

In FY2025, iPhone was still about 50% of Apple’s revenue. Services about 26%. Services is the beautiful high-margin bit.

The share price is a claim on future cash flows. Cash flows come from spending in the real economy. Stop the spending and you don’t get a cleaner portfolio but you do get a smaller business.

iPhone brought in about $210bn. Services about $109bn. Everything else is the supporting cast.

So when someone says skip the upgrade and buy the equity, they’re treating the customer and the shareholder as two different planets. They’re the same loop. Fewer phones sold over time means a weaker installed base, slower Services growth, and a multiple that has to reprice.

The viral line flatters the reader. It says you’re smarter than the sheep queuing for a craze.

The process line is duller and truer.

You are buying a business that needs customers.

If your thesis on Apple (or anything else) assumes infinite shareholders and zero buyers, you don’t have a thesis.

It is the same mistake as confusing a cheap chart with a reason to size up in a stock.

Labels instead of context.

Ask the boring questions. Who actually pays this company? What happens to the equity story if that spending slows? Am I sizing a brand I like, or a cash-flow machine I understand?

Mid-September

Second week of the month and the sector table is still uneven. The useful tell this week is not the index print. It is where the money actually went.

Energy is still the one working. About +2.3% on the week, +9.3% on the month, about +45% year to date. XLE took in roughly $567m over five days, the largest single-sector inflow in that pack. Oil refiners (CRAK) led at about +5.7% on the week. Global shipping (BOAT) about +3.8%. Oil services (OIH) is still up mid-30s year to date even after giving back from the May high. We spent weeks ignoring oil as a Hormuz headline sport. The sector kept grinding anyway. We notice. We are still not willing to get involved due to the unpredictability of the geopolitics.

Separate from crude, the power-for-AI story is where capital is concentrating: Bloom on fuel-cell speed-to-power for data centres, and nuclear / generation platforms (Vistra, Constellation) locking multi-year PPAs with hyperscalers. That is grid and capacity, not a Hormuz bet.

Crypto stays bullish on flows. Latest full week print: US crypto ETFs about $1.2bn in. Spot bitcoin ETFs about $987m of that. Ether ETFs about $215m. Bitcoin took most of the money. Price has chopped into this week and a couple of daily ETF prints flipped negative, but the institutional bid has not gone away. We got a momentum signal on ETH yesterday.

Tech is mixed and the flows disagree with the price. XLK about +1.1% on the week, basically flat on the month (−0.3%), still up around 30–37% year to date. Information Technology saw about $2.0bn leave over five days. IGV (software) was among the weakest at about −4.5% on the week with roughly $545m of outflows. SMH (semis) shed about $355m even while semis are up. One green session after a drawdown is not a regime. Semis are stretched versus their own history, soft on the month, and still the thing people want to declare back.

Healthcare is still making ground with real bids behind it. About +0.4% on the week, +6.2% on the month, +15.4% over three months. XLV took in about $230m. Useful software and cybersecurity sit in the same paid-this-year bucket. CrowdStrike sits here (see Stock in Focus).

Industrials look tired: soft on the week, about −6% on the month, and the industrial ETF segment saw hundreds of millions leave. Discretionary is worse: about −1.8% on the week, still negative year to date (around −2.7%). That sits next to this week’s Big Read. Choosier consumer spending shows up in XLY first.

Financials are fine. About +0.3% on the week, XLF took in about $453m. Cash machines being cash machines. Not a theme.

The Lights

🟢 Energy (price and flows) – Top of the table on both. XLE +$567m. Refiners and shipping did the niche work. We notice. We are still not getting involved while geopolitics creates the vol.

🟢 Crypto – About $1.2bn into US crypto ETFs on the latest full week. Bitcoin took most of it. Momentum signal on ETH yesterday. Still bullish on flows.

🟡 Healthcare / useful software / cyber – XLV +$230m, constructive month and quarter. CrowdStrike sits here.

🔴 Software beta + industrials + discretionary + semis bounce-chasers – IGV −4.5% and −$545m. Industrials and discretionary soft. One bounce in semis after a drawdown is not a regime flip.

CrowdStrike (CRWD)

This sits next to What We’re Ignoring on purpose (read What We’re Ignoring right after).

We are ignoring the Anthropic-alumni doom thread because it has no mechanism and no detail. Markets do not price that fantasy. They price the equal and opposite reaction… if enterprises push AI into every workflow, someone has to secure the endpoints, identities, cloud workloads and data those models touch. That someone is increasingly CrowdStrike.

George Kurtz said it after Q2: AI adoption needs security, and that is CrowdStrike. The invoice trail is clearer than the doom thread.

The numbers (Q2 FY27, quarter ended 31 July 2026)

Guidance they raised into

Full-year FY27 ARR $6.603bn–$6.612bn. Full-year revenue $5.991bn–$6.011bn. Net new ARR growth outlook lifted by 630bp to about 34% at the midpoint. Q3 ARR guide roughly $6.184bn–$6.188bn, Q3 revenue $1.523bn–$1.529bn.

Around $210 a share as of 11 Sept, market cap about $218bn. Expensive on GAAP multiples. Less silly if you are buying accelerating net new ARR into an AI-security spend cycle.

Why it earns the slot this week

Because the doom story and the security budget are the same coin. If AI is powerful enough to scare people, it is powerful enough to force CISOs to buy platform coverage. CrowdStrike already prints the ARR from that impulse.

What we need to see

Net new ARR staying elevated after the record quarter. Gross and net retention holding. Guidance that does not walk back the 34% net new ARR growth story.

Size

Theme vehicle sits inside useful software/cyber, not a belief trade or just punting on software generally. If net new ARR rolls over while the AI narrative stays abundant, the story is intact and the business is not. You already know what to do with that.

If you don’t, time to join the Academy.

Not financial advice. Figures from CrowdStrike Q2 FY27 results and company guidance.

We’re ignoring the kid that came out and said he left Anthropic after two months because AI is going to kill all humans.

We simply think this is a view that is entirely arrogant and isn’t founded upon reality.

It’s very much a bubble type view, and pretty typical of the AI people.

You can find the thread here.

What you might note is there is no detail on exactly how AI will kill us.

There is no specific reasoning.

This is akin to someone saying ‘the market is going to implode’, just because it has gone up a lot.

Anthropic recently released this paper, which is quite interesting on the future of work and our economies.

Now why exactly are we ignoring it?

Well we see everything as having an equal and opposite reaction.

Since we understand there is a biological incentive to live, anything deemed a threat to this will have a desire to be fought back against, especially if you can make money from it.

Which is why we have mentioned Crowdstrike as our stock of the week this week.

If you truly perceive AI to be a considerable threat, it is likely there will be an incentive from someone else to prevent said threat from being biologically risky.

Sure, labour might be at risk and the way our economies function could change, but no on really knows whether this means doom - it could mean everything is sped up and we face huge productivity gains.

The doom and gloom is rather Luddite-speak in my view, mixed with arrogance and blinkers.

What happens to oil?

We’re seeing high prices in oil again.

But there are some telling signs that I think it’ll come back down, maybe before midterms.

Trump will not like rates going up into the MTs, and largely, a lot of the rate prospects on a forward looking basis are dependent on the oil price.

I had a discussion with Natali from Leverage Shares on this yesterday as a wider part of our discussion on the Anthropic IPO, and her view is it’s pretty damn hard to know what will happen since it’s a bit of a geopolitical football right now.

She had mentioned that high prices generally conclude in a glut of oil, and with Trump capturing 67bn barrels of oil in Venezuela, perhaps this is something the market hasn’t yet priced for whatever reason.

What’s your take?

Oil higher in 6 months or do we find lower prices?