The Finance Flow Newsletter

The Bond Market vs The Stock Market

The Finance Flow — Sun, 30 Aug 2026 04:04:47 GMT

Kevin Warsh had his first appearance as the new Fed Chair at Jackson Hole, Nvidia earnings brought the whole market up, and we look into economic data to show how weak the labor market truly is.

This Week’s Numbers

S&P 500: 7,711 (+0.5%)
Nasdaq Composite: 26,402 (+0.8%)
Dow: 53,560 (+0.5%)
2-year Treasury: 4.352% (+13bp Friday)
10-year: 4.724% | 30-year: 5.209%
September hike odds (CME FedWatch): ~57.5%, up from 35.4% Thursday

Warsh’s First Jackson Hole Speech

Kevin Warsh gave his first Jackson Hole keynote Friday morning. He has been chair since May, has held two meetings, and has held rates steady at both while refusing to give markets details on his thoughts about where rates should land.

He didn’t give one Friday either. Early in the speech he joked that his remarks could be called an outline or a trail map, “just don’t call it forward guidance”.

What he did say was this: while the summer's inflation readings came in better than expected, they do not tell him that underlying trends have meaningfully improved. And if the Fed can't be confident inflation is moving to target at sufficient speed, "we have work to do”. Two-year yields jumped 13 basis points.

Where’s the Economy at Today?

Inflation is Stuck
July PCE held at 3.7% year over year, a tenth above the 3.6% consensus. Core came in at 3.3%. That is the fourth consecutive month at essentially the same level: 3.3% in April, 3.4% in May, 3.3% in June, 3.3% in July. The energy shock from the Iran conflict pushed headline to a three-year peak of 4.1% in May.

The Labor Market isn’t Growing
July payrolls fell 23,000 against a Dow Jones consensus of +83,000. The prior two months were revised down by a combined 103,000 — May cut from 129,000 to 63,000, June from 57,000 to 20,000. The unemployment rate "improved" to 4.1%, but only because 264,000 people left the labor force. Participation hit 61.4%, the lowest reading outside of covid, since 1976. Average hourly earnings grew 3.2% year over year, the weakest since May 2021. Against 3.7% inflation, the average worker got a pay cut in real terms.

Market Rose on Nvidia Earnings

Nvidia reported earnings Wednesday after the close. Revenue $96.2 billion, up 106% year over year. Data center revenue $89.0 billion, up 117%, now 92% of the company. Gross margin 75.0%. Q3 guidance of $108 billion against a Street looking for roughly $104 billion. Jensen Huang said AI has reached its inflection point and demand is accelerating.

The stock rose 8.7% Thursday and dragged the entire tape with it. Broadcom +4.5%, Intel +4.4%, and Salesforce closed up roughly 22.6% on its own results and an expanded Anthropic partnership.

My Take & What I’m Watching

The Housing Market Today

July new home sales fell to a 607,000 annualized rate, down 10.5% from June and 6.3% year over year, well below the roughly 620,000 economists expected.

The median new home price dropped to $393,800, the lowest since July 2021. Housing supply went up to 9.6 months from 8.5. Housing starts fell 12.4% from June and 13.5% from a year ago.

Builders are not sitting still. 63% offered incentives in August and 35% cut prices outright. Demand is slowing, as the share of consumers intending to buy a home in the next six months fell to 5.2% from 6.5%, the largest single-month drop in more than five years.

Important Dates Ahead

September 1: Ternus takes over at Apple
September 4: August jobs report (confirm BLS date)
September 9: Apple event; Treasury’s expanded buybacks begin
September 15–16: FOMC decision September 16
September 30: August PCE
October 6: Marvell investor day

Thanks for Reading

If you’re not subscribed yet, it’s free to join and you’ll get every deep dive, breakdown, and market take delivered straight to your inbox. Or become a paid subscriber for my takes, a look into my portfolio, and podcast episodes!

Subscribe now

Disclaimer: This newsletter is for informational and educational purposes only and does not constitute financial advice. Finance Flow is not a registered investment advisor. The portfolio and trade content shared with paid subscribers reflects personal investment activity and opinions only. Always conduct your own research and consult a qualified financial professional before making any investment decisions. Past performance is not indicative of future results.

The Capex Story Today

The Finance Flow — Mon, 10 Aug 2026 17:54:00 GMT

The Capex Story Across the Market

For twenty years, Alphabet was the closest thing public markets had to a perpetual motion machine: search ads in, free cash flow out, every quarter, without fail.

In Q2 2026, that machine posted −$5.9B in free cash flow — its first negative quarter since the 2004 IPO. Operating cash flow of $39.1B couldn’t cover $44.9B in quarterly capex.

Alphabet’s response was not to slow down. In the same earnings report, it raised 2026 capex guidance to $195–205B, halted buybacks, doubled long-term debt to $98.2B, and — weeks earlier — executed the largest equity raise in US corporate history ($84.75B priced, up to $90B with options exercised).

That’s the capex story today in one company: the most cash-generative businesses ever built are now spending faster than they earn, and reaching for every layer of the capital stack to keep going.

The four largest US hyperscalers plan $725B of combined 2026 capex, up 77% from ~$410B in 2025 (Goldman Sachs), funded increasingly by debt and off-balance-sheet structures. Bain estimates the sector is ~$800B short of the annual revenue needed to fund the 2030 buildout, and the BIS named an AI capex bust and circular financing among its top financial-stability risks.

2026 capex guidance across hyperscalers ratcheted up all year:

Alphabet $195–205B
Microsoft ~$190B
Amazon $220B
Meta $130–145B.

Memory-cost inflation is now a line item that moves budgets:

Microsoft attributes ~$25B of its $190B calendar-2026 plan to component prices
Amazon's entire $20B guidance raise (to $220B) was attributed to higher memory costs.

The pure-play labs have committed far beyond their revenue:

OpenAI ($13.1B 2025 revenue, ~$38.5B net loss) has ~$1.15T in vendor commitments and walked its infrastructure number from $1.4T to ~$600B (Feb 2026) and back up to ~$750B (mid-2026).
Anthropic raised $65B at a $965B valuation on a $47B revenue run-rate.

The Hyperscalers Behind the Buildout

Alphabet/Google:

Capex History:

2024 capex: ~$52.5B.
2025 capex: ~$85B (initial 2025 guidance was ~$75B, raised through the year).
2026 capex guidance: entered 2026 around $175–185B; after Q1 2026 raised to $180–190B; after Q2 2026 (July 22, 2026) raised to $195–205B.

CFO Anat Ashkenazi: “We are updating our full-year 2026 CapEx guidance range to $195 billion-$205 billion, up from our previous estimate of $180 billion-$190 billion. The increase in the range is primarily due to an acceleration in the delivery of capacity to meet growing demand.” She warned 2027 capex will “increase significantly.”

Q2 2026 actuals:

Capex $44.9B (double the $22.5B in Q2 2025; up from $35.7B in Q1 2026). H1 2026 capex ~$80.6B.
Revenue $119.8B (+24%)
Operating income $40.8B (+30%)
Net income $112.1B

About 60% of capex to servers, 40% to data centers/networking.

Free cash flow: Q2 2026 operating cash flow $39.1B could not cover $44.9B capex → −$5.9B FCF, first negative quarter since the 2004 IPO (Q1 2026 FCF was +$10.1B). MLQ

Balance sheet: long-term debt more than doubled since December 2025 to $98.2B; raised $20.3B in senior notes in Q2; ~$242.5B cash and marketable securities. Contracted purchase commitments $811B at end-June 2026 (up ~$500B from March), $200.7B due within one year.

Microsoft:

Capex History:

The $110–120B figures in trackers refer to fiscal-year capex (Microsoft's FY runs July–June). The ~$190B figure is a calendar-year 2026 plan disclosed on the FY26 Q3 call (April 29, 2026).

CFO Amy Hood: "For calendar year 2026, we expect to invest roughly $190 billion in capital expenditures which includes approximately $25 billion from the impact of higher component pricing" — up ~61% from 2025 (vs. a Visible Alpha consensus of $154.6B). On the FY26 Q4 call (July 29, 2026), Microsoft noted a shift from finance to operating leases "adjusts our expectation to approximately $175 billion" for calendar 2026 (a classification change, not a spending cut).

Azure/backlog: Azure grew 40% constant-currency in FY26 Q3, the fifth consecutive quarter of acceleration; Azure FY2026 revenue exceeded $100B for the first time (+41%); Hood guided 45% constant-currency Azure growth for FY26 Q1 (Sept 2026).

Microsoft is the one major hyperscaler not tapping bond markets heavily, funding from cash flow.

Leave a comment

Amazon

Capex History:

2024 capex: ~$83B.
2025 capex: ~$132B.
2026 guidance: $200B (set February 2026, held in April), raised to ~$220B on the Q2 2026 call (July 30/31, 2026); CEO Andy Jassy attributed the entire $20B increase to higher memory costs.

Q2 2026 actuals: capex $54.2B (up from $32.1B YoY; ~68% increase). Trailing-12-month capex ~$173B; TTM free cash flow −$7.6B (from +$18.2B a year earlier). Revenue $200.6B (+20%), operating income $27.5B (+43%), net income $62.6B.

AWS: revenue $42.2B (+37% YoY), fastest growth in 18 quarters, operating margin ~39.4%. Backlog ~$496B (up ~$132B in a single quarter), demand reserved through 2028.
Jassy: “we will still not have enough capacity to meet all the demand we have in 2026, and I believe this dynamic will also be true in 2027, too. In fact, the demand we already have for 2028 is striking.”
Balance sheet: long-term debt ~$119–128.9B (nearly double the $65.6B at December 2025 — an ~81% jump reported for Q1 2026). Sold $25B of bonds in July 2026 (weakest hyperscaler demand since Meta’s Oct 2025 deal; investors demanded wider spreads); ~$37B raised in March 2026.

How the Bill is Getting Paid

1. Cash flow. Alphabet FCF negative for the first time since IPO. Amazon TTM FCF negative. Meta at $784M for the quarter. Capex now consumes nearly all of what operations generate.

2. Debt. Alphabet’s long-term debt doubled to $98.2B in six months. Amazon’s nearly doubled to ~$119–129B. Meta issued ~$24.9B net in H1.

And demand is softening: Amazon’s July bond sale required wider spreads to clear.

3. Equity. Alphabet’s $84.75B raise (up to $90B) was the largest in US corporate history. Companies with fortress balance sheets don’t sell stock unless the spending plan exceeds what debt markets will comfortably absorb.

4. Off-balance-sheet. Vendor purchase commitments, pre-commencement leases, and JV financing vehicles now total an estimated ~$1.65 trillion across Big Tech (Nikkei) — more than their combined reported debt.

Meta’s Blue Owl structure is the template: the capacity gets built, the obligation is guaranteed, and the balance sheet stays clean. This leverage doesn’t disappear by being moved out of sight, and the BIS flagged exactly this in June, warning that terms of such deals are poorly disclosed, with risks of the same asset being “pledged multiple times.”

And one layer down, the pure-play labs have committed the most relative to what they earn: OpenAI ($13.1B of 2025 revenue against a ~$38.5B net loss) carries ~$1.15T in vendor commitments and has walked its headline infrastructure number from $1.4T to ~$600B and back up to ~$750B inside nine months.

Anthropic raised $65B at a $965B valuation on a $47B revenue run-rate. The hyperscalers at least have cash machines under the capex. The labs have the hyperscalers.

Red Flags

1. Purchase commitments are compounding faster than revenue. Alphabet’s contracted commitments jumped ~$500B in a single quarter to $811B, with $200.7B due within a year.

2. Bond markets are getting choosier. Wider spreads on Amazon’s July deal, the softest hyperscaler demand in nine months. The marginal dollar of this buildout costs more than it did in February.

3. Dividends and buybacks are becoming casualties. Alphabet halted repurchases. Meta’s dividend exceeded its quarterly FCF. The shareholder-return era that justified these multiples is being quietly suspended to fund the buildout.

The Bull Case

Every hyperscaler says the same thing: demand exceeds supply.

Pichai says cloud revenue would have been higher with more capacity. Jassy says Amazon won’t catch up to demand until after 2027. The growth is real — Google Cloud +82%, Azure +40%, AWS +37% (its fastest in 18 quarters) — and the backlogs are enormous: Oracle at $523B, AWS at ~$496B.

That matters, because demand-exceeds-supply is a fundamentally different risk profile than speculative overbuild. The fiber companies of 1999 were building for traffic that didn’t exist. These companies are building for contracts they’ve already signed.

PIMCO’s May assessment is the fair one: genuine risks — “uncertain monetization, potential overbuild, shortening asset lives, and growing reliance on debt” — but “a more disciplined and far more financeable cycle than the late-1990s telecom boom.”

The bear rebuttal is that signed contracts tell you demand exists today; they don’t tell you the capital gets repaid before the hardware depreciates.

Thanks for Reading!

If you made it this far, you already think differently about markets than most people.

If you’re not subscribed yet, it’s free to join and you’ll get every deep dive, breakdown, and market take delivered straight to your inbox.

Google Keeps Losing People

The Finance Flow — Thu, 06 Aug 2026 16:06:21 GMT

What Just Happened at Google?

On Wednesday, Alphabet (NASDAQ: GOOGL) lost four of the most-cited AI researchers on the planet on the same day, in the same announcement. Chief scientist Jeff Dean is leaving after 27 years. Oriol Vinyals, Quoc Le, and Sanjay Ghemawat are going with him. Together they’re launching Discovery Loop, a public-benefit startup that wants to fully automate multistep science and engineering experiments — machine learning research first, then hardware design, drug discovery, and clean energy

The market’s reaction was immediate and violent. Shares touched $381.81 before the news broke, bottomed at $355.16 once traders digested it (a nearly 7% intraday swing). On a $4.61 trillion company, that’s roughly $190 billion in market value gone by the close, and closer to $260 billion at the intraday low.

Alphabet is a founding investor in Discovery Loop. Google will maintain a research and cloud partnership with the startup its own chief scientist just left to build. The market treated Wednesday as a capability loss. Google is treating it as a spinout it gets to own a piece of.

How They’re Reorganizing

Demis Hassabis, Google’s Nobel laureate and arguably its single most important AI figure, is stepping back from day-to-day control of Google DeepMind — but he’s moving up, not out. He becomes Chair of DeepMind and Chief Scientist of Alphabet, with a broader remit across the whole company, while continuing to run Isomorphic Labs. Koray Kavukcuoglu — a 13-year DeepMind veteran who led early breakthroughs like WaveNet and DQN — takes over as SVP, reporting directly to Sundar Pichai, with ownership of Gemini model development, frontier research, and the Gemini app.

In his note to staff, Hassabis called Google “the only company that has the full stack” — infrastructure, custom silicon, cloud, and frontier models. Whatever you think of the spin, the TPU claim isn’t rhetoric. It’s one of the few genuinely full-stack AI positions in the industry.

Noticing a Pattern

Wednesday wasn’t a one-off. It’s the third act of a pattern:

June 22, 2026: Alphabet closed down 5.1% — its worst session in a year at the time — after two other researchers departed within days of each other
Earlier exits: Character.AI co-founder Noam Shazeer to OpenAI; Nobel laureate John Jumper to Anthropic
August 5, 2026: Four departures in one day, including the chief scientist

Talent risk has become a recurring, tradeable pressure point for this company. The market has now sold Alphabet hard twice in six weeks on researcher departures alone.

The Numbers

Revenue: $119.8B, up 24% YoY
Operating income: $40.8B, up 30% — operating margin widened to 34%
Google Cloud: $24.8B, up 82% YoY (from $13.6B) — cloud operating margin 35.6%
Cloud backlog: $514B, up more than $50B sequentially; just over half expected to convert to revenue within 24 months
Search: $63.3B, up 17% | YouTube ads: up 13%
Gemini app: 950M monthly active users
Valuation: ~17.8x trailing P/E — near historic lows for the company. Analyst targets average $425–428.
Q2 capex: $44.9B — exactly double a year ago, and a company record
FY2026 capex guidance: raised to $195–205B, from $180–190B just one quarter earlier — with management explicitly warning 2027 will “increase significantly”
Free cash flow: negative $5.9B — the first negative FCF quarter in Alphabet’s history as a public company
Balance sheet: $242.5B in cash and securities, $98.2B in long-term debt, plus a recent $100B debt raise to fund the buildout

Bull & Bear Case

The Bull Case: The Market Sold People, Not Fundamentals

The business didn’t change on Wednesday. Revenue is compounding at 24%, cloud is accelerating ahead of AWS and Azure, and the backlog gives Alphabet half a trillion dollars of contracted demand it hasn’t recognized yet.

Hassabis — the most decorated AI leader in the industry — expanded his role and publicly reaffirmed the full-stack position. Gemini development passed to a 13-year insider. And the departing researchers didn’t join a competitor: they built a company Google funds, partners with, and runs on Google Cloud. At ~17.8x trailing earnings — near historic lows — the bull argument is that you just got a rare entry point on a hyperscaler priced like a legacy business.

Leave a comment

The Bear Case: The Cracks Were Already Showing

The bear case doesn’t start with Jeff Dean. It starts with negative free cash flow. Alphabet just posted its first FCF-negative quarter ever, raised capex guidance by $15B mid-year, and told you 2027 will be significantly bigger — an open-ended commitment with no defined ceiling. The four hyperscalers will spend ~$725B combined in 2026, up 77% from 2025, and every one of them has traded down on capex despite beating earnings. Now layer on the talent story: when the researchers who built your infrastructure keep leaving — to OpenAI, to Anthropic, and now to their own ventures — the “full stack” advantage has a leak at the layer that matters most.

What to Watch Next

Q3 backlog print (October): Another $50B+ sequential add validates the capex; a stall is the first real bear signal
FCF trajectory: One negative quarter is a milestone, but two could shake the stock
Discovery Loop details: Terms of Alphabet’s stake and cloud commitment — is this a loss or a satellite?
Further departures: The market has shown you exactly how it reacts. Any follow-on exits are now tradeable events
2027 capex guidance: The first hard number will move the stock more than any single researcher

Thanks for Reading!

If you made it this far, you already think differently about markets than most people.

Subscribe now

If you’re not subscribed yet, it’s free to join and you’ll get every deep dive, breakdown, and market take delivered straight to your inbox.

Apple is Cooked.

The Finance Flow — Mon, 03 Aug 2026 00:27:06 GMT

What Happened to Apple Stock?

On Tuesday, Apple crossed $5 trillion, making it the most valuable company that has ever existed over Nvidia. Thursday it posted the best June quarter in its history. Friday it fell over 9% before ending the day down 7.35% to $308.91, the worst post-earnings drop in thirteen years.

Created with TradingView

The headlines for earnings said “weak guidance.” Apple guided the September quarter to 9–11% revenue growth against a 12.1% consensus and the entire shortfall is a supply problem. Cook told analysts Apple continues to expect high levels of demand, but with less flexibility in the supply chain than normal, the impact from supply constraints will increase significantly sequentially. Then he said it again, more bluntly: this isn’t a partner or supplier issue. The issue is incredibly strong demand. He described the cause not as a supply failure but as a demand forecast failure. iPhone and Mac are doing remarkably better than Apple planned for, and as a side note, I recently purchased a Mac Mini and love it so far!

This Last Quarter for Apple

Revenue: $109.4B, +16.4% — a June-quarter record, above the ~$108.7B consensus
Diluted EPS: $2.02 vs. $1.89 expected — $0.11 of it from tariff refunds
Gross margin: 50.1% — but roughly 2 points of that was the same tariff item. Adjusted, it was in line.
iPhone: $54.3B, +21.7% | Mac: $10.4B, +28.7% (vs. $8.74B expected)
Greater China: $18.8B, +22.4% — first real reacceleration in years, though short of the ~$19.6B Street expectation
Services: $30.7B, +12.1% — missed, and declined sequentially for the first time since 2022
9M free cash flow: $110.2B on $101.5B of net income — 109% cash conversion
Net cash: $62.2B

The Supply Constraint Story in Financials

Management’s story about supply is testable, and you can see it in their recent 10-Q. Component inventory went from $2.12 billion to $7.65 billion over nine months — up 260% — while finished goods actually fell.

That is the exact opposite of what a demand problem looks like. When demand softens, finished goods pile up in the channel. Here, Apple loaded up on raw components and cannot convert them to product fast enough to keep shelves stocked. It’s also the quantified version of the “carry-in inventory” Cook kept citing as a margin cushion. Apple saw the memory spike coming and front-ran it.

The constraint has two sources, and only one gets covered. Cook described the memory market as a “100-year flood” in pricing and said Apple reluctantly raised iPad and Mac prices as a result. But he also called out limited availability of the advanced chip nodes Apple silicon is built on — a TSMC allocation problem, not a RAM problem. Apple has already discontinued several high-memory Mac Studio configurations. Two independent bottlenecks, both driven by the AI buildout consuming the same wafers.

Will Users Choose Apple, Even with Price Hikes?

There are two ways to miss a quarter: a customer who buys their iPhone in February instead of October (revenue deferred), and a customer who gets tired of waiting and buys a Samsung (revenue destroyed, along with years of Services income attached to them).

The market just priced Apple as if this is mostly the second kind. Cook stressed this isn't a supplier issue, it's incredibly strong demand Apple can't build fast enough to meet. The math of that disagreement is stark: the guidance miss amounts to roughly $2–3 billion of revenue on a ~$113 billion quarter, and if those sales simply arrive a few months late, the true cost of the delay is well under $100 million, yet the market erased $460 billion.

The bulls' counter is that every month of shortage adds people to a waiting line, so when supply recovers Apple sells to the normal crowd plus the backlog. The catch, and the real bet in this stock: those waiting customers return just as Apple raises prices, and nobody knows how many will decide not to buy now. The December-quarter guide in late October is the first hard evidence either way.

Leave a comment

Apple’s Spend vs Everyone Else

Here is the contrast almost nobody is drawing. Apple spent $6.8 billion on capex over nine months. Only 1.9% of revenue, and down 28% year over year. For calendar 2026, Amazon, Microsoft, Alphabet and Meta plan roughly $725 billion combined, up 77% from last year’s $410 billion. Capex as a share of sales runs 86% at Oracle, 54% at Meta, 47% at Microsoft, 46% at Alphabet and 25% at Amazon. Apple’s is about 2%.

Analysts expect Microsoft’s free cash flow to turn negative in the fourth quarter for the first time since at least 2001, and Amazon’s is projected negative for the year. Apple converted 109% of net income into free cash flow while returning $62 billion in buybacks and $12 billion in dividends over the same nine months.

So the AI boom is hitting Apple through its cost of goods — a memory bill that is painful, cyclical, and self-correcting when fabs come online. It is hitting everyone else through depreciation schedules on assets with a three-to-five year useful life, funded increasingly with debt and off-balance-sheet structures. One of those is a margin problem. The other is a capital structure problem.

The Handoff of Management

This was Tim Cook’s last quarter as CEO of Apple. John Ternus takes over September 1. Ternus is a 25-year Apple engineer who ran the Apple Silicon transition — as strong a hardware pedigree as exists. He has also never run a public company, never run Services, and has limited regulatory experience. He inherits the Siri rebuild launching this fall on Google’s Gemini models under a roughly $1 billion annual license — an arrangement widely read as Apple conceding its own models aren’t frontier-grade.

His first ninety days contain the iPhone 18 and foldable launch, whatever iPhone price increase the memory crunch forces, the Siri reset, and a supply chain his predecessor just called the tightest in memory. Cook is staying on as executive chairman, which softens it. It’s still the most loaded CEO transition in the S&P.

My Year-End 2026 Price Targets

I'm using a forward-multiple framework here rather than a DCF, because a five-month price target is a question about what the market will pay, not about perpetual intrinsic value.

Bear ($200.36) — the supply shortfall becomes a demand problem. Price hikes meet real elasticity, memory stays tight past FY28, App Store regulation compounds, Siri slips again, and the multiple compresses toward its pre-2020 range. I’m weighting this lightest because it requires several independent things to go wrong at once, and the one piece of hard evidence we have — a 260% component inventory build against falling finished goods — points the other way.

Base ($311.96) — deferred revenue is recaptured through calendar 2027, memory peaks in the first half of FY27, price increases hold, and the multiple stays roughly where it trades today. My FY27 EPS of $9.12 is below the $9.57 consensus because I model gross margin at 47.0%; the Street implicitly assumes a faster recovery.

Bull ($399.71) — the record eligible-upgrader base plus lean channel inventory makes the iPhone 18 and foldable cycle the strongest in a decade, and Apple re-rates as the one mega-cap generating enormous free cash flow without an AI capex hole. I’m giving this real weight — nearly a third — because it doesn’t require anything crazy. The demand already exists by management’s own account, and the two catalysts (supply recovery, product cycle) are scheduled, not speculative. My bull case lands at $400, which is exactly where TD Cowen took its target after earnings, citing wafer supply constraints and cost.

I’m giving a probability weighted price target of $322 by year end, about 4% above today's price. That's not a table-pounding buy — it says Friday's selloff roughly overshot fair value by the size of one bad quarter, and the skew now favors the upside: the weighted downside to my bear case is about 5 points of the target, while the weighted upside to the bull case is about 9.

The Street average sits at $323.

A Few Red Flags

1. Services broke a four-year streak. Revenue fell sequentially from $30.98B — the first sequential decline since 2022 and the slowest growth since Q2 2025. CFO Kevan Parekh conceded App Store regulatory changes are beginning to affect growth — the first time Apple has directly acknowledged regulatory pressure hitting its financials on a call. Apple currently cannot charge commission on US external-link purchases post-Epic, with the Supreme Court hearing its appeal, and has cut fees in Japan, Brazil and the EU. Services carries a 75%+ gross margin.

2. “Deferred, not destroyed” is a hypothesis, not a fact. Nobody knows the price elasticity of a $100 iPhone increase. JPMorgan models $50–100; TechInsights’ margin-preserving estimate is closer to $270. If Apple takes the smaller hike, margin suffers. If it takes the larger one, units might.

3. The shortage may outrun my base case. SK Hynix warned in July it could persist past 2030, and new Micron and SK Hynix capacity won’t hit volume production until 2027 at the earliest. My base assumes relief in FY28. That is the assumption most likely to be wrong.

4. I’m below consensus on FY27 and above it on FY28. If the memory cycle breaks earlier than I model, I’m too bearish near-term. If it breaks later, I’m too bullish on the recovery.

Catalysts

September 1 — Ternus takes over
September — iPhone 18, foldable, Siri on Gemini. Watch the pricing and demand.
Late October — Q4 FY26. Two things matter: did gross margin land in the 47–48% guide, and did Services stabilize sequentially?
December quarter guide — the first read on whether constrained demand actually came back
Supreme Court on the App Store link-out appeal
Monthly DRAM contract pricing — now an Apple input, which remains a strange sentence to write

Thanks for Reading!

If you made it this far, you already think differently about markets than most people.

Subscribe now

If you’re not subscribed yet, it’s free to join and you’ll get every deep dive, breakdown, and market take delivered straight to your inbox.

Why Are the Biggest Tech Companies Crashing Today?

The Finance Flow — Thu, 23 Jul 2026 15:25:41 GMT

What Happened?

AI stocks got hit today. Alphabet dropped 7% after unveiling a massive capex forecast during their earnings yesterday and a new news report is spreading the selloff across the largest AI infrastructure builders.

The trigger isn't one bad earnings report. It's a Nikkei Asia study that put a number on something investors have been squinting at for months: the five biggest AI builders — Alphabet, Microsoft, Amazon, Meta, and Oracle — are carrying an estimated $1.65 trillion in off-balance-sheet obligations. That's more than the $1.35 trillion they officially report. And it's grown roughly 8x in four years.

These companies are clearly taking a huge bet on AI demand by dipping into more debt than ever seen before, and now investors are asking the question, “What if the demand doesn’t meet expectations?”.

Where is the Debt?

1. Footnote obligations. Long-term data center leases and GPU supply contracts get disclosed in the fine print of SEC filings, not on the balance sheet.

2. SPVs — Special Purpose Vehicles. This is the aggressive slice, and it’s the part drawing concern.

A tech giant and a private credit firm create a separate legal entity
The SPV borrows billions, owns the land, the building, the chips
The tech company leases it all back as the sole tenant
The debt sits in the SPV. Not on the tech company’s books.

Roughly $120 billion+ has moved off balance sheets this way in under two years. Oracle leads with ~$66B shifted. Meta ~$30B. xAI ~$20B.

Meta’s Hyperion data center in Louisiana. Blue Owl Capital and Meta put equity into a separate structure that took on $27 billion in debt. Meta is the only tenant. Meta argues the debt isn’t Meta’s problem. That $27 billion never appeared in an earnings report.

Why is This a Concern?

1. Rates. Treasury yields just hit their highest level since January 2025. The entire SPV model was built for a low-rate world. Every refinancing now costs more — and cash-flow-negative AI builders refinance constantly.

2. The lenders are blinking. Blue Owl — the same firm behind Meta’s Hyperion — reportedly walked away from a $10 billion Oracle data center deal in Michigan. The reported reason: concerns over Oracle’s ballooning debt and “unfavorable” terms. When the people funding the SPVs start passing, pay attention. S&P has already downgraded Oracle’s debt.

3. The obligations are coming home. Here’s the part nobody’s pricing: this debt doesn’t stay hidden. When the GPUs get delivered and the data centers switch on, those lease obligations move onto the balance sheet. Depreciation starts. And if AI demand disappoints? These companies could be in a lot of trouble.

Thanks for Reading!

If you made it this far, you already think differently about markets than most people.

Leave a comment

If you’re not subscribed yet, it’s free to join and you’ll get every deep dive, breakdown, and market take delivered straight to your inbox.

Is Microsoft Dead?

The Finance Flow — Mon, 20 Jul 2026 18:01:18 GMT

What’s Happening with Microsoft?

Microsoft (NASDAQ: MSFT) is down roughly 16% in 2026, its worst start to a year since 2000. June alone erased over 21%, the steepest monthly drop since December 2000, wiping out more than $500 billion in market value. Shares bottomed near $349 and have crawled back to around $395, still miles below the 52-week high of $555.

The narrative driving the selloff: Investors are worried the company is spending itself into a hole building AI infrastructure while, at the same time, AI tools threaten to replace the very software Microsoft sells. Getting squeezed from both ends. The only problem is, their numbers aren’t reflecting those fears.

Microsoft’s most recent quarter (Q3 FY26, ended March 31):

Revenue: $82.9B, up 18% YoY
Diluted EPS: $4.27, up 23% — beat estimates by 5%
Operating income: $38.4B, up 20%, with margins expanding to 46%
Microsoft Cloud: $54.5B, up 29%
Azure: up 40% — and accelerating from the prior quarter
AI business: $37B annual run rate, up 123% YoY
Commercial backlog (RPO): $627B, up 99% YoY

$627 billion in contracted future revenue. Nearly doubled in a year. That’s not a company being displaced. That’s businesses signing longer, bigger commitments to Microsoft than ever before, meanwhile the stock trades at ~24x earnings. It’s at the cheapest valuation in years, and revenue grew 18%. The entire decline is multiple compression driven by fear, not reflected in fundamentals.

Where’s the Fear Coming From?

Fear #1: The capex bill. Microsoft plans to spend roughly $190 billion on AI infrastructure in calendar 2026 (a ~60% jump). Last quarter’s capex came at a direct cost to free cash flow, which contracted 10%. Cloud gross margins slipped from 69% to 66%. Wall Street’s frustration is real: buybacks and dividends shrink when every spare dollar goes into data centers. One prominent analyst compared Microsoft’s treatment by investors to someone dressed completely wrong for the weather.

Fear #2: Displacement. The theory that AI chatbots and “vibe coding” make Office, Windows, and traditional software licenses obsolete. Why pay for Excel when AI can just... do it?

Are these problems truly worth selling Microsoft stock? Let’s see.

The Case

Here’s the thing about AI displacing Microsoft: AI has to work off of something. When an AI agent builds a financial model, it builds it in Excel. When it drafts a deck, it drafts it in PowerPoint. When it schedules your meeting, it’s touching Outlook and Teams. The file formats, the workflows, the compliance systems, the identity management — it’s all Microsoft, embedded so deep in corporate infrastructure that ripping it out would cost more than any AI subscription saves.

The numbers back this up:

M365 commercial cloud revenue grew 19% last quarter — ahead of Microsoft’s own expectations
Paid M365 seats are still growing — up 6% YoY, in the middle of the supposed AI extinction event
Copilot passed 20 million paid seats
Roughly two-thirds of last quarter’s capex went to GPUs and servers that generate revenue immediately — not speculative land grabs

The capex fear has nuance the headlines skip: about $25B of the 2026 budget is component price inflation, not extra volume. Microsoft is also the rare hyperscaler where demand still exceeds capacity — Azure grew 40% because they brought capacity online early, and management says customer demand continues to outstrip what they can build.

You don’t spend $190B on infrastructure nobody wants. You spend it because you can’t keep up with orders. They’re trying to keep up with the demand they are seeing.

Subscribe now

Things to Watch

1. Free cash flow is genuinely compressing. If AI revenue doesn’t scale fast enough to cover the buildout, this becomes a multi-year drag on shareholder returns. The AI run rate is $37B against a $190B capex plan. That gap needs to close.

2. Copilot monetization is unproven at scale. 20 million seats sounds big until you remember Microsoft has hundreds of millions of Office users. Enterprise adoption has been slower than the hype suggested.

3. Headcount is shrinking. Voluntary buyouts offered to ~7% of U.S. employees, on top of 15,000+ layoffs last year. Efficient, or a company bracing for something.

Earnings for Microsoft are set for July 29th and I will be looking closely into any insights that might help bring clarity to the problems investors have expressed above.

Thanks for Reading!

If you made it this far, you already think differently about markets than most people.

Leave a comment

If you’re not subscribed yet, it’s free to join and you’ll get every deep dive, breakdown, and market take delivered straight to your inbox.

What Happened to Netflix?

The Finance Flow — Fri, 17 Jul 2026 23:21:06 GMT

What Happened?

Netflix (NASDAQ: NFLX) reported Q2 2026 earnings after the bell yesterday, and on paper, there’s not much to complain about. EPS beat. Revenue grew double digits. Net income was up 9% year over year. The company even executed its largest quarterly buyback in history.

Created with TradingView

The stock fell nearly 9% after hours anyway — to roughly $68, below its 52-week low of $70.86 and the lowest level since September 2024. Shares are now down about 45% over the past year.

The Numbers

Revenue: $12.56B, up 13.4% YoY — a hair below the $12.58B consensus
EPS: $0.80 vs. $0.79 expected
Net income: $3.40B, up from $3.13B a year ago
FX-neutral revenue growth: 11%, down from 12% last quarter
Q3 guidance: $12.86B (+11.7%) — below Street expectations
FY26 guidance: Narrowed to $51.0–$51.4B
Buybacks: $4.7B in the quarter, a company record

That Q3 guide is the surface-level explanation for the selloff. Growth decelerating from 13% to below 12% isn’t a crisis, but at a premium multiple, deceleration gets punished. It seems that subscribers aren’t happy lately with new shows on Netflix, even with Netflix reporting that season 2 viewership of most of their shows has increased. With that being said, they also announced that they will only report on season 2 viewership annually now instead of quarterly.

The Disappearing Scoreboard

This was the first quarter in Netflix’s modern history reported without subscriber numbers. The company killed membership reporting after Q1 (when it had 325M+ paid members) and shifted to regional revenue figures instead.

Today, it went a step further: the “What We Watched” engagement report — the semi-annual viewership disclosure — is being cut to once a year starting in 2027. Netflix says the goal is to keep the focus on financial metrics like revenue and operating profit.

Maybe. But investors have a well-earned instinct here: companies emphasize the metrics that flatter them and retire the ones that don’t. Nielsen data showed Netflix’s U.S. TV viewership share slipping to a multi-year low of 7.8% in April. Viewing hours grew just 2% in the first half. And now the engagement scoreboard goes dark for most of the year.

The Bull Case Is Still Real

To be fair to the other side:

The ad business is scaling fast. On track for ~$3B in ad revenue this year, roughly double 2025, with 250M+ monthly active users on the ad tier.
Free cash flow is the best it’s ever been. FY26 FCF guidance sits around $12.5B, helped in part by the $2.8B breakup fee from the collapsed Warner Bros. deal.
Live events punch above their weight. Live programming is just over 5% of content spend and only ~1% of view hours — but live events account for six of the top ten sign-up days in the last five years. The 2027 FIFA Women’s World Cup, an expanded NFL slate, WWE, and MLB are all coming.
The runway argument. Management says Netflix reaches under 45% of addressable households and captures just 7% of its revenue opportunity.

A ~24x P/E for a business guiding to a 31.5% operating margin and record free cash flow is not the bubble pricing of years past. The average analyst target still sits far above the current price of around $70/share.

The Real Question

The company is asking investors to grade it on revenue and margins right as both engagement and growth visibility get harder to track. If the ad business keeps doubling and live events keep converting sign-ups, this two-year low looks like a gift in hindsight. If engagement keeps eroding behind a curtain that only opens once a year, today’s selloff was the warning shot.

Watch the Q3 print. With no subscriber numbers and no engagement report, revenue growth is now the only scoreboard left for investors to judge them on and Netflix just guided it lower.

Thanks for Reading!

If you made it this far, you already think differently about markets than most people.

Leave a comment
If you’re not subscribed yet, it’s free to join and you’ll get every deep dive, breakdown, and market take delivered straight to your inbox.

The $4 Trillion IPO Wave That Could Break the Market

The Finance Flow — Mon, 01 Jun 2026 21:50:30 GMT

Upcoming IPOs

In the span of under two weeks, the two most anticipated IPOs of the decade are set to debut. SpaceX filed its S-1 with the SEC on May 20, 2026, targeting a roughly $2 trillion valuation which, if reached, would be the largest IPO on record. Then this morning, Anthropic confidentially submitted a draft Form S-1 to the SEC for a proposed IPO, with the number of shares and the price not yet set. Throw in OpenAI, which confidentially filed around May 22 and is eyeing a September 2026 debut north of a $1 trillion valuation, and you have a tech IPO wave with combined paper value north of $4 trillion.

Subscribe now

SpaceX, Anthropic, & OpenAI Numbers

SpaceX

SpaceX runs the world's dominant launch business (Falcon 9), the fastest-growing satellite internet service (Starlink), and as of February the xAI artificial-intelligence operation is now merged in.

The 2025 numbers: $18.7 billion in revenue, of which Starlink was $11.4 billion, or 61% of the total. The launch-and-connectivity side is genuinely profitable.

2025 revenue: $18.7B | Starlink share: ~61%
Starlink operating profit (2025): $4.4B
xAI segment loss (2025): $6.4B | Space segment loss: $619M
Q1 2026 net loss: $4.28B | Accumulated deficit: $41.3B
Target valuation: $1.75T–$2T | Most recent tender (Dec 2025): ~$800B
IPO price range: blank in the S-1 — to be set during the early-June roadshow
Control: Musk holds ~41% of shares but ~85% of voting power via dual-class stock
**Not all numbers may be completely accurate

Goldman is leading the deal with 21 banks, and retail is earmarked for 30% of the float, three times the usual mega-cap norm.

Anthropic

Anthropic is the maker of the Claude AI models and is now, by valuation, the most valuable AI startup in the world. Today's filing follows its $65 billion Series H round at a $965 billion valuation, with an annualized revenue run-rate that recently crossed $47 billion. It's targeting a possible listing as early as October 2026 and brought on Wilson Sonsini — the firm that ran Google's 2004 IPO.

Valuation (Series H): $965B | Round size: $65B
Revenue run-rate: ~$47B annualized
Lead investors: Altimeter, Dragoneer, Greenoaks, Sequoia
Strategic backers in the round: Samsung, Micron, SK Hynix (the memory-chip trio)
Filing type: confidential draft — shares and price not disclosed
Targeted listing: as early as October 2026
**Not all numbers may be completely accurate

The market is assigning a near-trillion-dollar value off a run-rate figure and a funding-round price, not audited public statements. We won't see real numbers until a public S-1 lands, likely a month or so before any actual offering. And one detail from SpaceX's filing shows how circular this AI-infrastructure economy has become: Anthropic agreed to use SpaceX's Colossus 1 data center and will pay SpaceX $1.25 billion per month through May 2029. One of these IPO darlings is a billion-dollar-a-month customer of the other.

OpenAI

OpenAI is the maker of ChatGPT and the most widely used AI platform in the world. It confidentially filed its S-1 with the SEC on Friday, May 22, 2026, targeting a Q4 listing (possibly as early as September) at a valuation between $852 billion and $1 trillion, with Goldman Sachs and Morgan Stanley leading the deal.

Current valuation (March 2026 round): $852B | Round size: $122B — the largest private funding round on record
IPO valuation target: $852B–$1T+ (a $1T+ debut would be the largest tech IPO in history)
Revenue run-rate: ~$25B annualized (Feb 2026), up from ~$20B at end-2025
2025 actual revenue: ~$13.1B | ~$2B/month
Q1 2026: loses ~$1.22 for every $1 of revenue
Projected 2026 operating loss: ~$14B | Not expected to be cash-flow positive until ~2030
Compute commitments: ~$600B in five-year spending through 2030
Backers: Amazon, Nvidia, SoftBank | Filing type: confidential draft
Targeted listing: as early as September 2026
**Not all numbers may be completely accurate

Here's the tension at the center of the OpenAI story. The $852 billion private mark already priced the company at roughly 35x its run-rate; a $1 trillion IPO pushes that past 40x — for a business that doesn't expect to be cash-flow positive until around 2030. The October 2025 restructuring into a Public Benefit Corporation removed the old 100x investor-return cap and cleared the legal path to list, and a jury dismissed Elon Musk's lawsuit against the company two days before the filing, clearing the last big obstacle.

The Cautionary History of Big Waves of IPOs

The dot-com boom (1999–2000). The Nasdaq peaked at 5,048.62 on March 10, 2000 — a level it would not see again until March 2015. From the 1995 run-up to that peak the index rose over 500%, then fell roughly 78% by October 2002, wiping out about $5 trillion in market value.

The poster child: Priceline went public in March 1999 at $16, spiked to $88 on day one, and carried a $9.8 billion valuation despite racking up $142.5 million in early losses. Pets.com, Webvan, eToys, and Boo.com went bankrupt; the survivors who justified the whole era Amazon, & Microsoft were the exception, but still both experienced large drops in valuation during the time.

The 2021 IPO boom. 1,073 companies IPO'd in 2021, raising $317 billion; including SPACs, global proceeds hit a record $594 billion. Then the hangover: 2021 deals fell 14% on average in the six months after listing, versus a historical average gain of 14%. By mid-2022, of the ten largest 2021 IPOs, nine were down between 40% and 73%. Rivian briefly topped Ford and GM in market cap, then gave back its entire pop and traded below its IPO price.

What to Watch

SpaceX’s price range. It’s blank for now. The roadshow in early June sets it based on institutional demand — that range, against the December ~$800B tender, tells you how much hype is baked in before you can even buy.
Anthropic’s public S-1. Today’s filing is confidential. The real risk assessment can’t happen until audited numbers are public. Don’t underwrite a $965B valuation on a run-rate.

Thanks for Reading!

If you made it this far, you already think differently about markets than most people. That’s worth something.

Subscribe now

Leave a comment

If you’re not subscribed yet, it’s free to join and you’ll get every deep dive, breakdown, and market take delivered straight to your inbox.

This Space ETF Is Taking Off!

The Finance Flow — Wed, 27 May 2026 21:54:47 GMT

The NASA ETF

The Tema Space Innovators ETF (NYSE: NASA) launched earlier this year as the first pure-play space ETF to provide direct SpaceX exposure. It's an actively managed fund holding ~36 companies across the global space economy (launch, satellites, propulsion, defense optics, in-orbit services) with one structurally different position from anything else on the market: a pre-IPO stake in SpaceX held through an SPV.

Created with TradingView

Since inception, it's up 74.32%. The fund crossed into broader attention when Tema announced it had passed $2B AUM in under three years across its lineup, and $NASA is the flagship driving the recent flows.

Subscribe now

The Holdings

Top positions as of the most recent disclosure (top 10 = ~113% on a leveraged/notional basis given the SPV structure):

ASTS – AST SpaceMobile – 15.87% (direct-to-cell satellite)
RKLB – Rocket Lab – 15.24% (launch + Neutron)
PL – Planet Labs – 13.27% (earth observation)
SATS – EchoStar – 10.65% (satellite comms / spectrum)
FLY – Firefly Aerospace – 10.48% (launch + lunar)
VNP (TSX) – 5N Plus – 9.99% (specialty space-grade materials)
FTC.L – Filtronic – 9.94% (RF / antennas, supplies SpaceX)
OHB SE – 9.72% (European satellite manufacturer)
LUNR – Intuitive Machines – 9.27% (lunar landers, NASA CLPS)
SpaceX SPV (Preferred) – 8.73%

When you buy NASA you’re buying a barbell of high-conviction US-listed launch and satellite names, international exposure (Japan, Germany, Italy, Canada, Sweden, UK), and the SpaceX SPV as the asymmetric kicker.

The SpaceX Position

As of May 22, 2026, $NASA holds 232,192 SpaceX common share equivalents in an SPV with a total value of $147.4M.

The S-1 for SpaceX was recently filed May 20. Expected listing date: June 12, 2026, on Nasdaq under ticker SPCX. Reported target valuation: $1.75T to $2T, with $70–80B raise size, underwritten by a 23-bank syndicate led by Goldman.

The math people aren’t doing:

$NASA’s SpaceX position is currently marked at a ~$1.51T implied valuation
Midpoint IPO target is ~$1.87T — about 24% above the current mark
Synthetic pre-IPO contracts have quoted as high as ~$2.4T implied
Prediction markets are pricing 72% odds of >$2T close on day one

Because the SPV is held at transaction cost and doesn’t mark to public prices, the ETF’s SpaceX exposure should step up on IPO day, then again as the SPV converts to freely traded shares after the standard 90–180 day lockup (window opens roughly September to December 2026). Which means the ETF is sitting on a holding that’s almost certainly carried below market, with a forced re-rate event on the calendar.

Why I’m Bullish

1. The SpaceX re-rate. Even at the conservative end of the IPO range, the SPV alone steps up materially. If SpaceX prints at $2T+ and trades higher into year-end (which the prediction markets are pricing), that single position alone moves the NAV.

2. The "space race" trade gets a public face. Right now, retail can't easily express a SpaceX view. After June 12, every brokerage account in America can. The flows don't just go to SPCX — they spill into ASTS, RKLB, FLY, LUNR, and the rest of the basket because that's where the second-order plays live. $NASA owns all of them. I believe with the hype around SpaceX everyone may be racing to capture the broader space market.

3. The contracts are real. Rocket Lab is ramping Neutron. Firefly is winning DoD and NASA contracts. ASTS is moving from demos to commercial direct-to-cell. BlackSky and Planet are inside the intelligence-community spend cycle. The fundamentals are catching up to the multiples in a way they weren't in the 2021 SPAC bubble.

Potential Risks

One of the potential risks I see in the short-term for this trade is if SpaceX doesn’t have a successful IPO and plumets after becoming public. This could also take the wind out of the sails of all other space companies too. However, it feels like this is an obvious risk that everyone has thought about around SpaceX. The company is one of the leaders in innovation and exploration and I just can’t see them trading down.

My Take

I've held $NASA at a low-medium weight within my portfolio for a while now and the thesis hasn't changed: this is the cleanest public vehicle for the space economy trade, and it has a dated catalyst that most thematic ETFs don't get. June 12 isn't the end of the trade — it's the start of the re-rate. The lockup expiry in Q4 is the second leg. I don’t want to pick a single space stock and I feel this trade is a great way to take advantage of the broader space and satellite market.

Thanks for Reading!

If you made it this far, you already think differently about markets than most people. That’s worth something.
Subscribe now
If you’re not subscribed yet, it’s free to join and you’ll get every deep dive, breakdown, and market take delivered straight to your inbox.

Disclaimer: This newsletter is for informational and educational purposes only and does not constitute financial advice. Finance Flow is not a registered investment advisor. The portfolio and trade content shared with paid subscribers reflects personal investment activity and opinions only. Always conduct your own research and consult a qualified financial professional before making any investment decisions. Past performance is not indicative of future results.

TTEC Is Raiding Employee Retirement to Fund Its AI Pivot

The Finance Flow — Wed, 13 May 2026 16:15:42 GMT

Who is TTEC Holdings?

TTEC Holdings (NASDAQ: TTEC) a $2 billion revenue customer experience outsourcer, made headlines this week and has reportedly paused employer 401(k) matching contributions for its ~51,000 employees. This move, alongside a Q1 2026 earnings miss and a stock down 95%+ from pandemic highs, paints a picture of a company under real financial stress. TTEC runs call centers, customer service operations, and CX technology consulting for major brands across 22 countries. Enterprises pay TTEC per agent seat to handle their customer interactions. That model is now under direct threat from AI. This is why the company is pivoting to embracing AI.

Subscribe now

The Pivot

This isn't just a generic cost cut, TTEC told staff in an April 30 memo that it was pausing the company 401(k) match effective Q2 2026 through the rest of the year, and the company explicitly acknowledged it's redirecting those resources toward AI rather than simply citing general economic pressures. TTEC said the halt will help ensure the company's "long-term strength" by giving it flexibility to invest in the AI "tools, training, and capabilities that will define our future." The suspension affects approximately 16,000 U.S.-based employees. CEO Tuchman stated that management expects to achieve nearly 100% AI adoption with its current clients by the end of 2026, and said in the 2025 annual report that he wants the company to be "more agile and more profitable" by 2027, with the 401(k) pause being part of getting there.

Most companies cutting retirement benefits blame "cost pressures" or "macroeconomic conditions," but TTEC said the quiet part out loud and that honesty makes it both more alarming and more worth watching. Management expects to reassess the 401(k) policy in early 2027 based on the success of the AI integration strategy, making whether the match comes back the clearest signal of whether this bet is paying off.

Their Recent Earnings Miss

Revenue: $496M, down 7.1% YoY
Non-GAAP EPS: $0.15 vs. $0.25 consensus — a 40% miss
GAAP result: A net loss of $0.11/share
Adjusted EBITDA margin: 9.2%, down from 10.6%
Cash: $88.7M | Debt: $891.5M | Net debt: $803M
Debt-to-equity: ~8.8x

For every dollar of equity, TTEC owes nearly nine in debt. The company’s entire public market cap (~$130M) is a fraction of what it owes.

Some Potential Red Flags

1. The AI Threat Is Existential TTEC’s per-seat billing model is exactly what conversational AI is designed to replace. CEO Tuchman says “AI is our friend” — but clients are hesitant to spend, and the company is competing against pure-play tech firms without the cost structure of 51,000 employees across six continents.

2. The CEO Walked Away From His Own Buyout In September 2024, Tuchman (who owns 58% of TTEC) offered $6.85/share to take the company private — a 69% premium at the time. By August 2025, he withdrew, citing “market conditions.” The stock now trades around $2.50. When the person with the most inside knowledge of a company decides not to buy it, that tells you something.

3. Revenue Is in Structural Decline

2023: ~$2.46B
2024: ~$2.21B (-10.4%)
2025: ~$2.14B
2026 guidance: ~$2.03B (still declining)

Management is reaffirming full-year guidance, but Q1’s run rate ($496M × 4 = $1.98B) falls short of the $2.03B midpoint. They need to accelerate.

The company still generates positive free cash flow ($21M in Q1) and is attempting an AI pivot. But cutting employee retirement benefits to keep the lights on is not a sign that pivot is going well.

Thanks for Reading!

If you made it this far, you already think differently about markets than most people. That’s worth something.
If you’re not subscribed yet, it’s free to join and you’ll get every deep dive, breakdown, and market take delivered straight to your inbox.
Subscribe now

Disclaimer: This newsletter is for informational and educational purposes only and does not constitute financial advice. Finance Flow is not a registered investment advisor. The portfolio and trade content shared with paid subscribers reflects personal investment activity and opinions only. Always conduct your own research and consult a qualified financial professional before making any investment decisions. Past performance is not indicative of future results.

My New Price Target on Robinhood After Earnings

The Finance Flow — Wed, 29 Apr 2026 15:54:57 GMT

Robinhood’s Recent Earnings Miss

Robinhood reported Q1 2026 after the close on April 28. On the surface, the print looks fine revenue of $1.07B grew 15% YoY and diluted EPS of $0.38 was up 3%.

But the stock is down nearly 14% today because the bar was much higher: consensus expected $1.14B in revenue and $0.42 EPS, so the company missed on both lines.

The customer-engagement metrics are still excellent. The miss is on the monetization side, and the culprit is a single line item: crypto. The recent fall in crypto really put a squeeze on Non-Net Interest Income for Robinhood which dragged top-line revenue. Crypto revenue fell 47% YoY to $134M as retail digital-asset volumes slumped through Q1. This was the swing factor in the transaction-revenue line, which grew only 7% YoY despite strong showings elsewhere.

Subscribe now

The Bright Side: NII & Gold Subscribers

Net interest income grew 24% YoY to $359M, driven by interest-earning assets growth that more than offset the declining short-rate environment. Total assets jumped from $38.1B at year-end to $45.5B, and total loans (margin + the new Gold credit card) continue to climb.

Other revenue grew 57% to $85M, with Gold subscription revenue alone up 32% to $50M. Gold subscribers hit a record 4.3M (up 36% YoY) with a 16% attach rate to the customer base and a 40% attach rate for new customers. Robinhood Banking, launched late last year, has already crossed $2B in deposits from 125,000 customers — a 5x sequential expansion.

This is the “great wealth transfer” thesis that I keep hearing, and I love because it’s true. It’s working just not fast enough to offset crypto in any single quarter.

My Updated Outlook & Price Target

Management raised the 2026 adjusted OpEx + SBC guide to $2.7B–$2.825B, which now includes an additional $100M for the Trump Accounts program (Robinhood was tapped by Treasury to build the app for the new federal program targeting 60M young investors).

That added expense is real, but it's the kind of cost I'm happy to underwrite. Treasury handed Robinhood the keys to an app that will sit in front of 60 million young Americans — that's a customer acquisition channel most fintechs would pay anything to get, and Robinhood gets it as a paid contract. The market is treating the $100M as a drag on 2026 margins. I'm treating it as the cheapest user acquisition Robinhood will ever do.

I’m assuming net interest income grows 22% (in line with the Q1 actual), non-interest income grows 15% (which assumes crypto stays soft all year — probably too conservative given April is already trending better), and OpEx growth of 16% to match the company’s guide. I’m also assuming flat share count, because the $1.5B buyback authorization should fully offset stock-based compensation dilution.

These are base case numbers that assume Robinhood essentially keeps doing what it’s already doing.

At 30x exit P/E (right in line with where fintech peers trade) my 2028 EPS of $2.76 gets me to a future price target of $82.91, or roughly 17.6% upside from today's $70.50.

Three things have me leaning bullish on user and Gold subscriber growth specifically.

First, Gold subscribers are still scaling fast and getting stickier. 4.3M subscribers, up 36% YoY, with a 40% attach rate on new customers. That last number is the one I keep coming back to. Forty percent of new customers are signing up for the paid tier on day one. That’s not a finance app — that’s a consumer brand with genuine pricing power. As long as that attach rate holds, every new funded customer adds high-margin recurring revenue that’s invisible in the transaction-revenue line everyone is watching.

Second, funded customer growth is about to inflect. Robinhood added 1.7M funded customers YoY to reach 27.4M. That’s solid, but the Trump Accounts rollout puts the company in front of 60M young Americans through a Treasury-branded app. Even a 10% conversion to Robinhood’s main platform would be 6M new customers — more than 3x the YoY adds they just reported. Treasury did the marketing for them.

Third, the franchise is finally diversifying away from crypto. This quarter was painful precisely because it showed how reliant the headline numbers had become on crypto trading. But look at what grew: equities up 46%, options up 8%, prediction markets and futures hit record volumes, Gold up 32%, banking 5x’d to $2B in deposits, total platform assets up 39% to $307B. The next time crypto rolls over, those non-crypto lines will be twice the size they are today. The “Robinhood is just a crypto-beta” thesis dies in 2026.

Thanks for Reading!

If you made it this far, you already think differently about markets than most people. That’s worth something.
If you’re not subscribed yet, it’s free to join and you’ll get every deep dive, breakdown, and market take delivered straight to your inbox.

Disclaimer: This newsletter is for informational and educational purposes only and does not constitute financial advice. Finance Flow is not a registered investment advisor. The portfolio and trade content shared with paid subscribers reflects personal investment activity and opinions only. Always conduct your own research and consult a qualified financial professional before making any investment decisions. Past performance is not indicative of future results.

Is SoFi a Turnaround Play or a Value Trap?

The Finance Flow — Mon, 27 Apr 2026 17:36:09 GMT

Introduction

SoFi Technologies has had a rough few months. After touching $32.73 in late 2025, the stock shed more than half its value by today following a short-seller report that compared its accounting to Enron.

Today, shares sit around $19 and Wall Street is deeply divided, with price targets ranging from $3 (Wedbush) to $37 (Citigroup), because of this I think is worth a further look.

Who is SoFi?

Most investors may know SoFi as a student loan provider, however they have since expanded their model. The company now operates across three segments: a lending business (personal, student, home loans), a consumer financial services arm (SoFi Money, Invest, Credit Card), and a B2B technology platform (Galileo and Technisys) that powers other fintech companies' infrastructure.

The result is a vertically integrated financial ecosystem that's quietly become a real bank and emerging in the Fintech space with names like Robinhood. SoFi obtained its federal bank charter in early 2022, and the deposit base it's built since then is the core of the modern investment thesis. Deposits have grown from zero to $37.5 billion by the end of FY2025 — giving SoFi a low-cost funding source to underwrite its own loan book rather than relying on expensive securitization.

The financial trajectory here is genuinely striking. Net revenue grew from $566M in FY2020 to $3.6B in FY2025 — a five-year CAGR of roughly 46%. Net interest income alone went from $178M to $2.2B over that same stretch, as the deposit-funded balance sheet scaled. Total assets crossed $50 billion.

More importantly, the company is now profitable on a GAAP basis. After years of losses — including a ($301M) net loss in FY2023 — SoFi posted $499M in net income in FY2024, its first fully profitable year. The efficiency ratio tells the same story: from 156% in FY2020 (deeply in the red for every dollar of revenue generated) to 84.6% in FY2025, having broken the 100% break-even threshold in 2024. For context, most traditional banks run efficiency ratios in the 55-65% range — SoFi still has room to improve, but the direction is unambiguous.

Subscribe now

The Markets View

Investor’s are split on SoFi and reflects genuine structural disagreements about what this company actually is, and what risks come with owning it.

One of the main things the street is concerned about is their focus on crypto and stablecoins. SoFi isn't dabbling in crypto it's building an entire digital asset infrastructure stack. That includes SoFi Crypto trading directly from FDIC-insured accounts, SoFi Pay running on blockchain rails for international payments, and SoFi USD — a fully cash-backed stablecoin held in its Federal Reserve master account. The company is also pursuing secured lending against crypto collateral and institutional crypto trading. A federally chartered bank running a stablecoin and institutional crypto services faces a categorically different regulatory scrutiny than a pure-play fintech. The same charter that gives SoFi cheap deposit funding also makes it a more visible regulatory target. Any technical failure or governance lapse in the crypto stack doesn't just hurt that segment — it puts the entire "one-stop-shop" brand at risk, because SoFi's moat is built on trust and comprehensiveness.

The Muddy Waters short report from March 2026 alleged that SoFi’s reported personal loan charge-off rate of 2.89% was materially understated — with the actual rate closer to 6.1%. SoFi has pushed back, but the public filings tell their own story: net loan losses were $668M in FY2024 and $671M in FY2025, against a total loan book that hit $38 billion.

Management points to prime borrowers — average income around $158,000, average FICO of 746 — and net cumulative loss assumptions of 7–8% as evidence the portfolio is healthy. Recent vintages show losses tracking well below the prior higher-loss vintage, with 37% of principal still outstanding. On paper, that looks fine.

The concern is the trajectory. Personal loan annualized charge-offs are running in the mid-4% range, 90-day delinquencies have ticked up sequentially due to normal seasoning, and Q4 personal loan originations hit a record $7.5 billion — meaning the book is growing rapidly at the exact moment it’s maturing. Rapid growth, portfolio seasoning, and reliance on historical loss assumptions is a combination that tends to look manageable right up until it doesn’t. The fair value marks on personal loans currently sit around 105.7% — above par. Any deterioration in credit performance or capital markets appetite for SoFi-originated paper brings those marks down, and with them, a significant portion of reported earnings.

The Variant View

The market's concerns are legitimate, but the current price suggests investors are treating every risk as a certainty. At ~$19, the stock is implying roughly a 2% annualized EPS CAGR over the next three years. For a company that just grew net revenue 35% year-over-year and crossed into GAAP profitability for the first time, that's a remarkably pessimistic assumption and it's where the variant view begins.

Using a reverse multiple valuation at the current price with trailing EPS of $0.39 and a 46x exit P/E — consistent with fintech peer multiples — the market is implying a Year 3 EPS of just $0.41. My model gets to $0.55 by Year 3. That gap of $0.14 in EPS translates to a price target of $25.25, implying ~33% upside from today.

The key is that this isn't a heroic bull case. The model uses a 35% NII CAGR — well below SoFi's five-year historical NII growth average of 70.9%. Non-interest income grows at just 15% (vs. a 33.6% historical average). Expenses compound at 30%, roughly in line with history. Shares dilute at 8% annually — well below the distorted SPAC-era average, and credible now that SoFi is GAAP profitable and the heavy issuance phase is behind it.

The charge-off concern is real, but the efficiency ratio trajectory provides important context. SoFi's efficiency ratio has improved from 156% in FY2020 to 84.6% in FY2025. That improvement didn't happen despite rising loan losses; it happened alongside them. The loan book generated over $1.8 billion in adjusted net revenue in FY2025, with personal loan originations hitting $36.4 billion for the year — a scale that gives SoFi meaningful pricing and securitization leverage even if charge-offs drift higher. The crypto buildout is the most speculative element of SoFi's story, and the model treats it accordingly — it's not a primary driver of the EPS projection. The variant view is built on NII compounding as the deposit base scales and the loan book grows, not on stablecoin revenue or institutional crypto trading fees. Those are optionality, not load-bearing assumptions.

Summary

The deeper variant is structural. SoFi now serves 13.7 million members with over 20 million products, and 40% of new products are opened by existing members not new customer acquisition. That cross-sell dynamic is what drives the NII CAGR assumption. As members adopt banking, brokerage, and lending products, NII per member rises without proportional increases in acquisition cost. The blended acquisition cost falls. The efficiency ratio continues improving.

The deposit base reinforces the loop: $37.5 billion in deposits, 97% from direct-deposit primary relationships, funds the loan book at a cost well below what warehouse lines or securitization would require. Loan production hit $36.4 billion in FY2025, up from $23.2 billion in FY2024. That growth in origination volume — funded cheaply — is what makes the 35% NII CAGR assumption conservative rather than aggressive.

The market is currently pricing a company in structural decline. The numbers describe a company still in the early innings of monetizing a member base it spent years and billions of dollars building. At $19, you’re not paying for perfection — you’re paying roughly what the stagnation scenario implies. The variant view is simply that SoFi keeps doing what it’s been doing, at a fraction of its historical pace, and the stock has meaningful upside from here.

Thanks for Reading!

If you made it this far, you already think differently about markets than most people. That’s worth something.
If you’re not subscribed yet, it’s free to join and you’ll get every deep dive, breakdown, and market take delivered straight to your inbox.

Disclaimer: This newsletter is for informational and educational purposes only and does not constitute financial advice. Finance Flow is not a registered investment advisor. The portfolio and trade content shared with paid subscribers reflects personal investment activity and opinions only. Always conduct your own research and consult a qualified financial professional before making any investment decisions. Past performance is not indicative of future results.

The Finance Flow Stock Picks & Portfolio Update

The Finance Flow — Sun, 26 Apr 2026 17:27:32 GMT

Intro

Current Market:
We're heading into one of the most eventful stretches of the year. A new Fed chair announcement is working through the market, Middle East tensions remain a persistent geopolitical overhang, and this week kicks off the biggest earnings slate we've seen in 2026. There's a lot going on, but my stance stays the same: selectively bullish on a handful of names I have real conviction in. I'll walk you through exactly why below.

This Under-the-Radar Homebuilder Is a Buy Right Now

The Finance Flow — Wed, 22 Apr 2026 17:02:17 GMT

The Homebuilding Sector & Recommendation

In a housing market battered by rising mortgage rates, affordability concerns, and one of the worst sales environments in thirty years, identifying a homebuilder trading at a meaningful discount to its intrinsic value requires looking past the short-term cyclicality. There are only around 18 publicly traded homebuilders and after a thorough analysis of Green Brick Partners (NYSE: GRBK), a mid-cap homebuilder operating across Texas, Georgia, and Florida, myself and our team at Texas Tech University believes the market has underpriced a structurally differentiated company with durable competitive advantages and a clear path to continued earnings growth.

This article walks through our investment thesis, the key data underpinning our analysis, the risks we identified, and why we arrive at a BUY recommendation with a 12-month target price of $79.09 — representing 26% upside from the March 30th closing price of $62.75.

Company Overview

Green Brick Partners is a mid-cap homebuilder that closed nearly 4,000 homes in 2025, achieving a 12% compound annual growth rate in closings since 2020.

The company generates revenue through seven independent homebuilder subsidiaries, each operating under its own brand, with Trophy Signature Homes as the largest, accounting for over one-third of total revenue. The remainder comes from non-Trophy brands — CB Jeni, Normandy, Southgate, Centre Living, Providence, and GHO — which collectively represent 64% of revenue and provide meaningful geographic and product diversification.

GRBK’s five-year revenue CAGR of 17% has been driven primarily by volume growth rather than the post-pandemic price inflation that temporarily inflated peer revenues. This distinction matters enormously for forecasting the sustainability of future growth. It also signals that GRBK has built an operational engine, not merely benefitted from a macro tailwind.

The Land-Heavy Model Creates a Competitive Advantage

The first and most foundational reason to invest in GRBK is its land-heavy strategy. Unlike most large-cap peers who operate on a land-light or asset-light model — sourcing lots from third-party developers — Green Brick owns and develops 76% of its lots internally. This is not merely a stylistic preference; it is a deliberate structural decision that carries significant financial consequences.

In homebuilding, land costs flow directly through cost of goods sold (COGS) when a home is sold. By eliminating third-party land developers from the equation, GRBK captures the developer spread — an embedded cost savings that lowers its land basis by an estimated 20–25% relative to prevailing market prices. The financial result is striking: GRBK’s COGS as a percentage of revenue stands at 69.5%, compared to an industry average of 78.3% — nearly 9 percentage points lower than peers.

A fair critique of the land-heavy model is capital efficiency. Because GRBK owns and develops land rather than optioning it, capital remains tied up for longer periods. This is evident in the company’s cash conversion cycle of 493 days versus an industry median of 353 days. Many homebuilders prefer land-light models precisely to avoid this capital intensity.

However, GRBK’s returns contradict the concern. Despite the longer capital cycle, the company generates a return on equity (ROE) of 18.5% and a return on assets (ROA) of 13.1% — both well above the peer medians of 11.0% and 6.9%. The longer cycle reflects deliberate strategy and value capture, not inefficiency.

Trophy is the Fast-Cycle Growth Engine

The second pillar of our thesis centers on Trophy Signature Homes, GRBK’s flagship and largest subsidiary. Trophy is not merely GRBK’s biggest revenue contributor it is the operational proof-of-concept for the land-heavy model’s scalability.

Trophy consistently delivers below 100-day build cycles, among the fastest in the industry. The peer distribution for build times ranges from 90 to 150 days, with a median of 120 days, placing Trophy in the top quartile for construction speed, and because lots are pre-developed and controlled internally, Trophy can break ground rapidly without waiting for third-party lot delivery.

Speed that undermines quality would be a red flag. But Trophy’s cancellation rates tell a different story. At 7.5%, GRBK’s cancellation rate is approximately 50% below the industry average of 15.1%, and substantially better than small-cap peers averaging 26.5%. Low cancellations reflect buyer confidence, product consistency, and strong demand fundamentals.

The impact of the Trophy model on GRBK’s financials has been transformative. Since Trophy was founded, revenue grew 17% annually from 2020 to 2025, but earnings per share grew even faster more than tripling over the same period, outpacing revenue growth by approximately 10 percentage points. This operating leverage is a hallmark of a scalable, high-quality business model.

2025 was a difficult year, GRBK's EPS declined for the first time in eight years, reflecting what the chief economist of the National Association of Realtors described as one of the worst housing markets in three decades. But even in this environment, GRBK's revenue was essentially flat — declining only 0.02% — while peer revenues contracted an average of 5.6% and peer EPS fell an average of 35.8%. The company materially outperformed through a severe industry downturn, which we view as evidence of structural advantage, not cyclical luck.

Expansion of the Current Model

The third reason to invest is the replicability of GRBK’s operating model. Trophy’s success in the Dallas-Fort Worth market has demonstrated a template: identify high-growth Sunbelt submarkets, secure land early at below-market prices, build fast, and sell at competitive price points aligned to local median home prices.

The non-Trophy subsidiaries play an equally important role in the expansion story. They diversify the company’s exposure across geographies and price points, absorbing demand from rate-sensitive buyers (Trophy’s entry-level sweet spot) and rate-insensitive luxury buyers alike. In any market environment, this portfolio approach provides a degree of insulation that pure-play luxury or entry-level builders lack.

Expansion is only as credible as the balance sheet behind it. GRBK’s financial position is notably conservative for a growth-stage homebuilder. Its debt-to-assets ratio is among the lowest in the peer group, its interest coverage ratio is strong, and the company has consistently maintained high liquidity. This means the expansion is not predicated on aggressive leverage — a meaningful differentiator in a rising rate environment.

GRBK also uses Municipal Utility Districts (MUDs) and Public Improvement Districts (PIDs) — tax-increment financing structures common in Texas — to fund infrastructure without drawing on corporate debt. This off-balance-sheet financing mechanism is a key enabler of the land-heavy model and partially explains how GRBK maintains low leverage despite owning a large lot inventory.

Governance & Executive Team

GRBK’s governance profile is a material positive for long-term investors. The company is founder-controlled, with executive compensation below peer averages despite outperforming on returns. This alignment between founder incentives and shareholder outcomes helps explain the consistency of the land-heavy strategy — it is not a strategy subject to quarterly earnings pressure or activist-driven pivots.

On capital return, management has been disciplined and shareholder-friendly. GRBK has repurchased $327 million in stock between 2021 and 2025, steadily reducing the share count from 51.1 million to 43.6 million over that period. The company authorized a new $150 million buyback plan in December 2025, and the escalating size of successive programs — alongside a stock trading at a discount to peers. The company is still largely owned and controlled by both co-founders who align compensation with shareholder values.

Valuation & Conclusion

Our valuation blends two methodologies: a five-year discounted cash flow (DCF) model (weighted 60%) and a relative valuation using forward multiples against 15 peer homebuilders (weighted 40%).

GRBK screens as one of the strongest companies in its peer set on profitability and return metrics, yet trades at relatively low earnings-based multiples. Its P/E of 8.76x and EV/EBITDA of 6.9x both rank near the bottom of the peer group, despite its ROA of 13.1% placing it in the top tier. Our P/A regression — fitted across 15 peers — implies a fair Price-to-Assets multiple of 1.7x versus GRBK’s current 1.1x, representing a 45% discount to regression-implied fair value. Applying 75th percentile forward multiples to our earnings estimates yields a relative value of $86.52.

Blended Target Price

Blending intrinsic value ($74.14) at 60% weight and relative value ($86.52) at 40% weight, we arrive at a final 12-month price target of $79.09, representing 26% upside from the March 30th close of $62.75.

Sensitivity analysis across multiple DCF scenario weightings and relative valuation multiple mixes consistently produces target prices in the $74–$85 range, with no weighting scheme that supports the current market price as fair value given GRBK’s fundamentals.

Risks

Conclusion

Green Brick Partners is a structurally differentiated homebuilder that has translated a contrarian land strategy into consistently superior margins, returns, and earnings growth. Its core competitive advantages — a land-heavy model, a fast-cycle growth engine in Trophy, and a disciplined, founder-led management team — are durable, not cyclical.

The market appears to be pricing GRBK as though the 2025 industry downturn represents a permanent impairment of the business. Our analysis suggests the opposite: GRBK outperformed through one of the worst housing environments in thirty years, is expanding into high-growth markets with a proven playbook, and trades at a significant discount to every earnings- and asset-based fair value framework we applied.

We issue a BUY recommendation with a 12-month price target of $79.09 — 26% upside from current levels.

Thanks for Reading!

If you made it this far, you already think differently about markets than most people. That’s worth something.
If you’re not subscribed yet, it’s free to join and you’ll get every deep dive, breakdown, and market take delivered straight to your inbox.
Subscribe now

Disclaimer: This newsletter is for informational and educational purposes only and does not constitute financial advice. Finance Flow is not a registered investment advisor. The portfolio and trade content shared with paid subscribers reflects personal investment activity and opinions only. Always conduct your own research and consult a qualified financial professional before making any investment decisions. Past performance is not indicative of future results.

Why JPM is the Best Positioned Bank in Today's Market

The Finance Flow — Tue, 14 Apr 2026 20:51:05 GMT

The Financial Sector Today

The U.S. financial sector is navigating one of its more complex macro backdrops in recent memory. The Federal Reserve, having completed the most aggressive tightening cycle in four decades, is now expected to be easing and the market is trying to figure out which banks come out of that transition stronger and which ones were quietly living on borrowed time.

The concern is legitimate. Net interest income (NII) the spread banks earn between what they pay depositors and what they charge borrowers, expanded dramatically from 2022 to 2024 as rates climbed. Now that the Fed is cutting, that tailwind reverses. For institutions heavily concentrated in NII, the next two to three years represent a structural earnings headwind.

Layered on top of that is a geopolitical environment that has no clean historical parallel. The war in Ukraine, ongoing tensions in the Middle East, and a fracturing global trade order have introduced macro uncertainty that makes capital allocation decisions genuinely difficult. When businesses can’t price risk with confidence, M&A slows, capital markets activity contracts, and loan demand softens. That is the bear case for the sector in aggregate.

But within that environment, there is a meaningful difference between banks that built their earnings power on the rate cycle and banks that built it on something more durable. JPMorgan Chase sits firmly in the second category and the data bears that out in a way that wasn’t obvious until you separate the two revenue streams and watch what happens to each one as conditions change.

JPMorgan’s Two-Engine Revenue Structure

JPMorgan’s revenue splits into two broad streams (NII & Fee/Non-NII):

That near-50/50 split is deceptive in the best possible way. It looks like NII still dominates, but the trajectory tells the real story: fee income as a share of revenue has been climbing every year. In the forecast period, JPMorgan’s earnings trajectory implies fee income will exceed 50% of revenue before the end of the decade. NII is also faded by 0.75% each year to reflect macro-conditions and compression on their spread.

Crucially, while NII is rate-sensitive, fee income is not. It scales with markets, with AUM accumulation, with M&A volumes, and with the sheer dominance of JPMorgan’s Corporate & Investment Bank — the top-ranked franchise globally in debt capital markets, equity capital markets, and M&A advisory virtually every year.

The Fed is in a cutting cycle. Consensus expects 1–2 more cuts in 2025–2026, compressing NII industry-wide. For JPMorgan, this is a manageable headwind partially offset by fee tailwinds from the same macro environment that drives rate cuts: easing financial conditions typically accelerate M&A volumes, capital markets activity, and AUM appreciation.

Even under the variant view — where Fed cuts compress NII below consensus — JPMorgan still grows revenue every single year through 2028. Total revenue reaches $197 billion in FY2028 versus $182.5 billion in FY2025, an 8% compounded gain during a period most bank analysts would treat as structurally difficult.

EPS follows: from $20.65 in FY2025 to $22.97 in FY2028, supported by the continued ~1% annual reduction in shares outstanding as buybacks run at roughly $12–14 billion per year. That buyback program is itself only possible because JPMorgan’s CET1 capital ratio (15.7% versus Bank of America’s 11.9%) gives management enormous flexibility to return capital without regulatory friction.

AUM Growth

Between 2016 and 2020, a reasonable analyst looking at NII-to-revenue ratios would have concluded that JPMorgan and Bank of America were structurally similar institutions at different scales. Both hovered around 50% NII dependency. The spread appeared stable.

What the NII ratio obscured was the compounding going on inside JPMorgan’s Asset & Wealth Management segment. As JPM expanded both its branch network and its digital platform (the most-downloaded banking app in the US, supported by a $17 billion annual technology budget) it captured a disproportionate share of wealth management relationships. AUM grew to over $3 trillion, generating recurring fee revenue that scales automatically with market appreciation and requires no incremental balance sheet risk.

Bank of America’s Merrill Lynch franchise is not weak. But JPMorgan has been taking share. And the earnings compounding that flows from that difference is structural, not cyclical: every $100 billion in AUM growth generates roughly $500–$700 million in incremental annual fee income at prevailing management fee rates.

The operating margin divergence is the financial fingerprint of this dynamic. In 2016, JPMorgan and Bank of America had near-identical operating margins. By FY2024, JPMorgan’s operating margin reached 42.8% versus Bank of America’s 31.7%. That 11-point gap — on a $177 billion revenue base — represents roughly $19 billion in annual earnings that JPMorgan generates and Bank of America does not.

Leave a comment

Overview

Bull Case: Rates Hold, Fee Engines Accelerate

The bull case doesn’t require anything heroic from the macro environment — it just requires the status quo to persist. If the Fed pauses its cutting cycle and the federal funds rate stabilizes near current levels for the next two to three years, JPMorgan’s NII holds relatively steady while the fee engine continues compounding. Stable rates historically correlate with increased CEO confidence, which translates into M&A activity, IPO pipelines, and debt issuance — all of which flow directly into JPMorgan’s top-ranked CIB franchise. Add AUM appreciation on top of $3 trillion in managed assets, and fee income could grow at the high end of the 6–8% range rather than the base-case 6%. In this scenario, EPS tracks materially ahead of the $22.97 FY2028 estimate, and the current P/E multiple of ~15.5x looks conservative rather than stretched. A P/E multiple of 13.2x is more reasonable for banks vs the 12.56x 10Y average.

Base Case: Modest Cuts, Fee Income as the Hedge

The base case is the scenario already embedded in the forecast: the Fed cuts 1–2 more times, compressing NII by roughly 0.75% per year over the forecast period. JPMorgan absorbs that headwind because its fee income — growing at ~6% annually — more than offsets the NII decline in absolute dollar terms. This is the differentiation that matters relative to peers: most banks in a rate-cutting environment see total revenue decline as NII contracts and they have no offsetting growth engine. JPMorgan grows through it. Revenue reaches $197 billion by FY2028, EPS hits $22.97, and the buyback program continues reducing the share count by ~1% annually. Total IRR over the 3-year period — price appreciation plus adjusted shareholder yield — comes to approximately 4.9%. Not spectacular, but defensible and grounded in structural earnings power rather than macro luck.

Bear Case: Recession, Geopolitical Shock, and Spread Compression

The bear case is a genuine one, and it deserves to be taken seriously. A global recession, accelerated by the war in the middle east, a broader geopolitical fracture, or a trade-driven demand collapse would hit JPMorgan on multiple fronts simultaneously. NII spreads compress as the Fed cuts aggressively to stimulate growth. M&A and capital markets activity dry up, as they did in 2022 when JPMorgan’s fee income fell 17% in a single year. Credit losses rise, forcing higher provision builds that erode net income. AUM values decline with markets, pulling recurring wealth management fees lower.

In this scenario the earnings picture looks meaningfully weaker. NII declining faster than the base case, fee income contracting rather than growing, and the operating margin giving back some of the gains made over the prior decade. The multiple would likely compress from ~15.5x toward the 10–11x range that reflects trough earnings power for a money-center bank.

Even here, though, JPMorgan is not the worst place to be. Its CET1 ratio at 15.7% gives it a capital cushion that most peers lack. Its deposit franchise — built on the stickiest, lowest-cost funding in the industry — insulates it from liquidity stress. And its diversification means that while no engine is firing at full capacity in a recession, neither engine is completely shut off. The losses are real; the structural advantages survive.

Thanks for Reading!

If you made it this far, you already think differently about markets than most people. That’s worth something.
If you’re not subscribed yet, it’s free to join and you’ll get every deep dive, breakdown, and market take delivered straight to your inbox.
🔓 Free subscribers get:

Weekly market deep dives like this one
Macro & equity breakdowns on the companies and trends that matter

🔐 Paid subscribers ($10/month or $60/year) get:

Access to the Finance Flow Discord — where we talk markets in real time
Weekly personal trade breakdowns — every position I enter or exit, and why
A live look into my actual portfolio weightings and performance

Subscribe now

Disclaimer: This newsletter is for informational and educational purposes only and does not constitute financial advice. Finance Flow is not a registered investment advisor. The portfolio and trade content shared with paid subscribers reflects personal investment activity and opinions only. Always conduct your own research and consult a qualified financial professional before making any investment decisions. Past performance is not indicative of future results.

Inside the Portfolio & More

The Finance Flow — Sat, 28 Mar 2026 04:34:24 GMT

Welcome back and if you're newer to the paid side, welcome. This is the inner circle and backbone of The Finance Flow, and I'm glad you're here. Every paid post is my commitment to go deeper than the free content: real numbers, real positions, real reasoning, and real conversations with you all. Thank you for making that possible. Let's get into this week's breakdown.

The Market is Wrong on Amazon

The Finance Flow — Thu, 26 Mar 2026 01:44:12 GMT

When Amazon announced it would spend $200 billion in capital expenditures in 2026 ($50B more than expectations) the market did what markets do: it sold first and asked questions later. Shares dropped over 10% year-to-date. Bernstein trimmed its target. The word "unsustainable" started making the rounds.

We think the market is making a mistake, confusing the short-term cash consumption with long-term compression and destruction, but this is simply not the case. The $200 billion number feels scary in isolation. In context, it's one of the most strategically rational bets in corporate history.

So, let’s break down our main reasons we’re bullish on and why the CapEx panic is actually creating an entry point for investors.

Subscribe now

CapEx Isn’t Speculative. It’s Demand Driven

The bear case on Amazon’s spending boils down to one fear: “what if the AI bubble pops and all this infrastructure becomes stranded capital?”

It’s a fair question. But it’s the wrong question for Amazon specifically.

Here’s what gets buried in the headlines: Amazon isn’t building speculatively. CEO Andy Jassy was explicit on the Q4 2025 earnings call — “We are monetizing capacity as fast as we can install it.”

AWS’s cloud backlog stands at $244 billion, up 40% year-over-year and 22% sequentially. Amazon isn’t guessing that demand will arrive, it’s already there, waiting on infrastructure. The $200B in CapEx that they projected for 2026, is actually a small relative slow down in spendings growth compared to the past 2 years.

AWS posted $35.6 billion in Q4 2025 revenue, its fastest growth in 13 quarters, with an annualized run rate of $142 billion. (Yahoo Finance) Morgan Stanley reiterated an Overweight rating with a $300 price target, arguing AWS growth could exceed 30% in the coming quarters. The stock is currently trading around $210 today, and we at The Finance Flow give a 12 month rating closer to $280.

Yes, free cash flow took a hit — dropping to around $10 billion from $38.2 billion the prior years as CapEx consumed the majority of operating cash flow. But Amazon generated $139.5 billion in operating cash flow in 2025, providing a meaningful buffer. (Yahoo Finance) This is a company with the financial firepower to sustain this investment cycle. The margin pressure is real and temporary. The infrastructure being built is permanent and they are already seeing the demand.

"This isn't some sort of quixotic, top-line grab." — Andy Jassy, Amazon CEO, Q4 2025 Earnings Call.

Amazon built its fulfillment network over a decade through relentless capital deployment that made Wall Street uncomfortable every single year. Those who stayed patient were rewarded with a logistics moat that is now virtually impossible to replicate. The AI infrastructure buildout follows the same playbook.

A Moat Nobody Talks Enough About

Amazon isn’t just a cloud company or an e-commerce company. It’s a flywheel and each component reinforces every other component in ways that compound quietly.

AWS contributed approximately 60% of Amazon’s consolidated operating income in the first nine months of 2025, while representing just 18% of revenue. That’s the profit engine funding everything else.

On logistics: Amazon successfully regionalized its fulfillment network to the point where 76% of orders are fulfilled within their own local geography, drastically reducing the cost-to-serve and Amazon recently surpassed Walmart in total annual revenue.

On advertising: annualized ad revenue exceeded $60 billion in 2025, outpacing retail, subscriptions, and even AWS. The driver was simple: Amazon sits at the intersection of shopping intent, first-party data, and media consumption.

What makes this moat exceptional isn’t any single segment, it’s the integration. Google has search intent but not purchase data at this depth. Meta has targeting but not commerce. Amazon has all three, getting stronger with every Prime member, every purchase, every streamed episode. They are even expanding to autonomous ride-sharing too!

The logistics moat deserves special mention. By regionalizing fulfillment so that 76% of orders are handled within their local geography, Amazon has turned what was historically a margin drag into a structural competitive weapon. Since their beginning, Amazon has always leaned into innovation and expansion of their products & services.

Leave a comment

Quietly Winning AI

Everyone focuses on the ChatGPT blowup and OpenAI, the consumer-facing AI products grabbing headlines. Amazon has made a calculated decision to fight the AI war on different terrain: the infrastructure layer, the enterprise layer, and the silicon layer. They have also strategically partnered with both OpenAI and Anthropic (the two leaders in consumer facing AI).

Rather than competing for consumer-facing AI attention, AWS focused on powering the back end of AI adoption. Bedrock made it easier for companies to build generative AI applications using Amazon’s own models or those of partners. This deepened customer stickiness and expanded the long-term opportunity as more companies scale AI deployments. (The Motley Fool)

By designing its own AI chips (Trainium and Inferentia), Amazon has indicated potential cost and efficiency benefits relative to relying solely on third-party GPUs, which could help with both pricing and capacity flexibility over time. Trainium2 is already fully subscribed and a multi-billion dollar business. Trainium3 is in preview.

Jassy described the AI market as a “barbell,” with the AI labs on one side and enterprises on the other, looking to the technology as a “productivity and cost avoidance” tool and suggested that middle enterprise segment “very well may end up being the largest and most durable.” AWS is perfectly positioned to serve exactly that market.

Overall, Amazon is being punished for a capital allocation decision that mirrors every great investment the company has ever made. The fulfillment network, AWS itself, Prime — all of it looked like reckless spending in the moment and looked like genius in hindsight. The $200 billion AI infrastructure buildout is the same story, written larger.

The market is pricing in fear of stranded capital. We’re pricing in a $244 billion backlog, an unmatched moat across four business lines, a proprietary silicon strategy, and a management team that has delivered on exactly this kind of long-term bet before.

The panic is the entry point. Amazon isn’t broken. It’s building.

Thanks for Reading!

If you made it this far, you already think differently about markets than most people. That’s worth something.

If you’re not subscribed yet, it’s free to join and you’ll get every deep dive, breakdown, and market take delivered straight to your inbox.

🔓 Free subscribers get:

Weekly market deep dives like this one
Macro & equity breakdowns on the companies and trends that matter

🔐 Paid subscribers ($10/month or $60/year) get:

Access to the Finance Flow Discord — where we talk markets in real time
Weekly personal trade breakdowns — every position I enter or exit, and why
A live look into my actual portfolio weightings and performance

Subscribe now

Microsoft is off 20% YTD, Why?

The Finance Flow — Tue, 24 Feb 2026 21:36:05 GMT

Preview

With Microsoft ($MSFT) down over 20% YTD and 40% from their recent highs, I wanted to look at their most recent earnings call and dive deeper into their financials to better understand why investors have been selling off.

Subscribe now

Recent Earnings Report

Headline Numbers

Revenue: $81.3B (+17% constant currency)
EPS: $4.14 (+24% CC, excluding OpenAI accounting)
Operating Income: +21% CC
Operating Margin: 47%
Gross Margin: 68% (slightly down due to AI infra spend)
Operating Expenses: +5% CC

Despite $37.5B in quarterly CapEx, margins remain extremely high.

Microsoft 365 Copilot

15M paid seats
Daily active users up 10x YoY
Seat adds +160% YoY
450M+ commercial 365 seats = massive upsell opportunity

GitHub Copilot

4.7M paid subscribers (+75% YoY)
Strong enterprise adoption (e.g., Siemens 30K+ dev rollout)

Microsoft isn’t just selling AI infrastructure — it’s embedding AI inside its dominant SaaS base.

Microsoft delivered another strong quarter, proving that AI is already translating into real revenue, margin durability, and enterprise adoption. While CapEx remains elevated, demand continues to outpace supply, and much of the infrastructure investment is contract-backed and strategically allocated toward higher-margin first-party AI products. Overall, this quarter reinforces Microsoft’s position as the leading full-stack AI platform spanning infrastructure, orchestration, and applications.

The CapEx Story

One of the biggest investor concerns around Microsoft has been the surge in AI infrastructure spending, however after looking at the data I believe we see a different story.

It seems as though the heavy acceleration in CapEx spend has already happened, and that they’re actually looking now to maintain this growth rather than going “All in”. Microsoft front-loaded AI infrastructure investments in 2024. Now revenue is scaling on top of that installed base, with top line revenue growing by over 16% this last year.
If CapEx continues to moderate as it did last year, and revenue continues to compound, then this creates expanding cash flows.

Microsoft also made a deal with Anthropic and Nvidia earlier this year, where Anthropic (creators behind Claude) committed to over $30B in spend towards Azure compute over the next few years. The main economic driver for Microsoft is long‑term Azure consumption (tens of billions of high‑margin cloud revenue), plus having Claude models deeply integrated into Azure and Copilot products.

Revenue Growth

The growth in cloud revenue just over the last few years has been incredible, and I believe shows the shift into AI infrastructure for Microsoft. They’ve grown revenue in that segment by over 25% in only 2 years, and it’s showing no signs of slowing as mentioned before with their Anthropic deal.

I believe as they continue to moderate and slow CapEx spend, they will show exponential free cash flow growth. Top line revenue has also grown by over double digits YoY since 2023.

Potential Risks

Risks remain around the stock and their heavy investment into AI infrastructure. Herte are the top 3 risks most investors have:

CapEx Spend: Microsoft’s AI and data‑center build‑out has pushed quarterly capital expenditure to record levels, now far above its total annual capex just a couple of years ago. Investors worry that pricing for Copilot and other AI services may not fully cover the long‑term compute and infrastructure costs, pressuring margins and free cash flow if AI usage does not scale profitably.
Dependence on OpenAI and a few AI partners: A very large share of Microsoft’s future cloud revenue is now contractually tied to OpenAI; one analysis notes that roughly 45% of Microsoft’s commercial remaining performance obligations (RPO) are linked to OpenAI‑related commitments.
Regulation, antitrust, and competitive pressure: Microsoft’s dominant positions in PC OS, productivity software, and now enterprise AI put it squarely in regulators’ sights in the U.S. and Europe, with ongoing antitrust and digital‑markets scrutiny that could constrain deals or bundling strategies.

Thanks for Reading!

If today’s recap helped you, here’s how to stay plugged into everything we’re building:

1. Subscribe for Paid Weekly Deep-Dives

If you want more than daily headlines, upgrade to a paid subscription.

Subscribe now

You’ll get:

Weekly breakdowns of what really moved markets
My personal portfolio thoughts
My 3 stock picks & trades each week

2. Follow The Finance Flow on Twitter/X

I’m posting charts, quick takes, data drops, and real-time reactions throughout the day.
If you want to stay sharp between newsletter posts, follow here:

https://x.com/TheFinanceFlow

3. Listen to The Finance Flow Podcast

Every Saturday, I break down markets in a deeper, more conversational format, with interviews plus Q&A’s and segments like “3 Moves I’m Making.”

5 Stocks I'd Hold for the Next 2-5 Years

The Finance Flow — Mon, 23 Feb 2026 21:11:32 GMT

Preview

For today’s article I asked myself, “What businesses are so embedded in everyday life and corporate infrastructure that they’re nearly impossible to replace?”

I’ve broken these companies up into 3 separate categories:
Retail:
Walmart ($WMT)
Proctor & Gamble ($PG)

Corporate:

Microsoft ($MSFT)
Adobe ($ADBE)

Both:
Amazon ($AMZN)

Subscribe now

Retail

Walmart ($WMT)

Walmart wins in both strong and weak economies.

When times are good, it captures broad consumer demand. When times are tight, consumers trade down and Walmart gains share. People still need their basic needs, and a Walmart is about 10 minutes away from most American’s homes. Walmart leaned heavily into e-commerce this last year, growing that division by 24% last year, and overall operating income grew by over 10%. In Walmart U.S., 35% of store-fulfilled eCommerce orders were delivered in under 3 hours.

In the latest earnings call executives summarized their risks like customers making under $50k/year, and the increased popularity in GLP-1 drugs, and what kind of impacts those might have, “Households under $50,000 remain stretched, often living paycheck to paycheck, but still prioritize convenience nearly as much as price.” Also saying, “GLP-1 usage is having expected, but net-neutral, effects: Fresh is a big basket driver; puts and takes largely offset, so no major overall impact currently.”

Walmart is also looking to expand not just in e-commerce, but also in their overall use of AI and technology to save on labor and other costs. They trade on the NASDAQ which is usually for the top 100 tech companies. Walmart’s financials also look strong at a glance.

Proctor & Gamble ($PG)

Proctor & Gamble is another company that I feel strongly that most retail users can’t go without. They own brands like:

Tide
Pampers
Gillette
Bounty

These are repeat purchases with pricing power. PG works as a “necessity” stock because it combines staple products people buy in every macro environment with a long, reliable dividend and the ability to raise prices without losing too much demand.

Even when the stock doesn’t move, they still pay a pretty handsome dividend every quarter. The current yield is around 2.5-2.8%, which is paired with a higher safety and consistency compared to competitors.

Corporate

Microsoft ($MSFT)

Microsoft is enterprise oxygen.

Azure powers cloud infrastructure.
Office 365 runs internal workflows.
Teams handles communication.
Copilot is integrating AI into everything.

Switching costs are enormous. Companies don’t cut Microsoft subscriptions lightly. Microsoft’s products have become embedded in how modern enterprises operate, which makes its revenue base unusually durable and “stickier” than most tech vendors.

In cloud infrastructure, Microsoft Azure is the clear No. 2 globally, with around 20–21 percent market share versus AWS’s high‑20s, and far ahead of most others. Microsoft 365’s presence in millions of companies creates network effects (everyone knows the tools) and deep integration with identity (Entra ID/Azure AD), security, and device management.

Adobe ($ADBE)

Adobe has positioned itself as critical infrastructure for modern businesses by owning the tools that create, move, and measure digital content across marketing, documents, and commerce.

“Picks and shovels” of the digital economy

Adobe’s cloud pivot turned Photoshop, Illustrator, Premiere Pro, and related apps into subscription utilities that generated over 22 billion dollars in revenue in 2024, led by Creative Cloud and Document Cloud.
These tools are embedded in how companies produce ads, product images, UI designs, training materials, and videos—functions that exist in nearly every industry, not just “creative” sectors.

Adobe has evolved from a creative software vendor into a utility-like provider whose clouds—Creative, Document, and Experience—run the content, contracts, and customer interactions that keep many businesses operating day to day.

Both

Amazon ($AMZN)

Amazon is two businesses in one.

Retail Side:

Prime ecosystem
E-commerce dominance
Growing advertising business

Enterprise Side:

AWS cloud infrastructure
AI compute buildout
Data center scale moat

AWS YoY growth to highlight

Full‑year 2024 AWS revenue grew about 18–19% year over year, from roughly 90.8 billion to 107.6 billion dollars.
Full‑year 2025 AWS revenue then accelerated to about 20% growth, reaching roughly 128.7 billion dollars.
In Q4 2025 specifically, AWS revenue was 35.6 billion dollars, up 24% YoY, its fastest quarterly growth in 13 quarters (these lines up with the 24% point in your chart).

For the retail narrative, focus on online and physical stores plus the broader “Stores” segments:

Online Stores revenue grew about 7% in 2024, from 231.9 billion to 247.0 billion dollars, after a slower 2023.
North America segment sales (largely retail plus some services) grew 10% in 2025 to about 426.3 billion dollars, while international segment sales grew 13% to 161.9 billion dollars.
In Q4 2025 alone, North America sales rose 10% YoY to 127.1 billion dollars and international retail sales were up 17% YoY to 50.7 billion dollars, showing solid global consumer demand.
Physical Stores revenue also resumed modest growth in 2024, increasing about 6% from 20.0 billion to 21.2 billion dollars, helped by grocery and convenience formats.

Thanks for Reading!

If today’s recap helped you, here’s how to stay plugged into everything we’re building:

1. Subscribe for Paid Weekly Deep-Dives

If you want more than daily headlines, upgrade to a paid subscription.

Subscribe now

You’ll get:

Weekly breakdowns of what really moved markets
My personal portfolio thoughts
My 3 stock picks & trades each week

2. Follow The Finance Flow on Twitter/X

I’m posting charts, quick takes, data drops, and real-time reactions throughout the day.
If you want to stay sharp between newsletter posts, follow here:

https://x.com/TheFinanceFlow

3. Listen to The Finance Flow Podcast

Every Saturday, I break down markets in a deeper, more conversational format, with interviews plus Q&A’s and segments like “3 Moves I’m Making.”

Is Robinhood a Buy Now?

The Finance Flow — Wed, 11 Feb 2026 22:23:48 GMT

Why is the Stock Down?

Robinhood stock is down 50% from the highs following their Q3 earnings call and yesterday they fell over 10% following their Q4 earnings call. The company beat on EPS expectations of $0.63, reporting $0.66/share. However, they company missed on revenue expectations of $1.32B, reporting $1.28B for Q4.

Subscribe now

What Q4 Earnings Said about this Year

Robinhood had an impressive 2025 with the stock up over 200%, and here are their yearly numbers:

Revenue: $4.5B (+52% YoY)
Adjusted EBITDA: $2.5B (+76%)
56% EBITDA margin
EPS: $2.05
$68B in net deposits
$324B in platform assets (+70%)
4.2M Gold subscribers (+60%)
Nearly $1B in crypto revenue

Fastest-growing business ever in prediction markets

Management guided for an increase ~18% in overall expenses for next year focusing on:

Gold Card
Banking
Prediction markets
Cortex AI
Robinhood Social
Robinhood Ventures
Robinhood Chain
International expansion

They want to invest heavily into these as they reported impressive growth and outlooks in these categories for 2026.

Concerns from the Call

Monthly active users (MAUs) in Q4 2025 were about 13.0 million, down from roughly 14.9 million in the prior‑year quarter (a decline of about 1.9 million users), this has been a concern around the company for awhile as many investors believe that once younger users get enough money, they eventually switch platforms. Assets slipped about 3% from Q3 to Q4 2025, the first quarter‑over‑quarter decline in assets under custody (AUC)/platform assets since 2021.

Bullish Outlook

Although Robinhood seems to be losing users, before this last quarter they haven’t seen a drop in QoQ AUC since 2021, and I believe this drop has the tech sell-off and crypto crash to blame. More users lost money during the sell-off and the holiday season. I was impressed with the companies growth in gold membership users growing over 60% YoY to 4.2m users, as well as the growth in revenue from prediction markets growing their “other revenue” segment by over 700%.

Customers traded over 12B prediction contracts in 2025 and more than 4B in early 2026, with record volumes in Q4 and into January across sports and a growing set of non‑sports events. The Rothera joint venture, slated to launch around mid‑2026, will give Robinhood vertical control over the exchange leg of these contracts, improving economics and product control while remaining open to external flow.

Is the Stock a Buy Today?

Price Today (2/11): $77.97

P/E: 37.8x

Last time the stock traded at this P/E: June 1, 2025

At ~38x trailing earnings after a 50% pullback, I see Robinhood as a higher‑risk compounder rather than an obvious bargain. The core brokerage is throwing off cash, net deposits and Gold subscribers are growing briskly, and the prediction‑market and card/banking franchises are scaling much faster than the rest of the business. At the same time, MAUs are falling, assets just posted their first sequential decline since 2021, and management is leaning into a complex slate of new initiatives that will keep expenses growing faster than many value‑oriented investors would like. For investors who believe Robinhood can translate this product roadmap into durable high‑teens growth and maintain its edge with active traders, the current drawdown looks like a reasonable entry point; for anyone less convinced on that long‑term execution, the stock is probably still one to watch rather than chase.

Thanks for Reading!

If today’s recap helped you, here’s how to stay plugged into everything we’re building:

1. Subscribe for Paid Weekly Deep-Dives

If you want more than daily headlines, upgrade to a paid subscription.

Subscribe now

You’ll get:

Weekly breakdowns of what really moved markets
My personal portfolio thoughts
My 3 stock picks & trades each week

2. Follow The Finance Flow on Twitter/X

I’m posting charts, quick takes, data drops, and real-time reactions throughout the day.
If you want to stay sharp between newsletter posts, follow here:

https://x.com/TheFinanceFlow

3. Listen to The Finance Flow Podcast

Every Saturday, I break down markets in a deeper, more conversational format, with interviews plus Q&A’s and segments like “3 Moves I’m Making.”