Capital Wealth
Specialty · Markets · The Verdict File

Heard on the Street: Twelve Columns, One Verdict

We read every Heard on the Street column in last week’s Journal — twelve of them, Monday through the weekend — and graded our own investment ideas against the paper’s in-house analysts. The result: four validations, one warning, zero contradictions.

By Sean Anees Saifi · Capital Wealth · Published Sunday, August 9, 2026 · Source: The Wall Street Journal, Heard on the Street columns, August 3–8, 2026 editions
Key Points
12
columns read, Monday through the weekend
4–0
our ideas validated vs. contradicted
230%
stock market value vs. U.S. economy
3.9%
high-dividend index yearly return; plain index: 13.2%
Twelve columns from the Journal’s analysis desk, one week, one verdict — four themes validated, one warning taken seriously.
Twelve columns from the Journal’s analysis desk, one week, one verdict — four themes validated, one warning taken seriously.
In one line: The Journal’s own analysts spent a week testing ideas like ours, and our plan passed — with one warning about AI borrowing that we were already built for.

Heard on the Street is the Journal’s in-house team of analysts — writers paid to be skeptical of everything, including their own front page. Last week they published twelve columns. We read every one, twice. Then we lined them up against the ideas the Capital Wealth Growth Portfolio is built on.

One warning first: when your homework matches the answer key, check that you did not copy it. So for each idea we wrote down what would prove us wrong. That list is at the end.

The twelve columns

ColumnThe callOur matching idea
The cloud payoff · MonAI profit lives in the clouds — own one, or don’tOwn the cloud giants ✓
The AI bond flood · Tue$400B coming; stay short-term and high-qualityTreasury-bill ladder ✓
The Buffett Indicator · Tue230% of GDP; expect smaller long-run returnsNo chasing ✓
Fallen investing stars · WedMoney arrives after the fame, then regrets itBoring on purpose ✓
The yen’s only fix · TueOnly Japan’s central bank can stop the slideShort-term bonds only ✓
Texas grid U-turn · WedNext bottleneck: electricity and permitsCaterpillar, the workaround ✓
The Dividend Mind Trick · Thu3.9% vs 13.2% a year; affordable dividends are incomeChevron & Williams in; Pfizer review ✓
SpaceX’s rockets · Thu28x sales; everything rides on Starship workingWe own none ✓
Apple’s AI bill · ThuDoing AI on the cheap may be on borrowed timeCloud rule, again ✓
AstraZeneca’s discount · FriUnder 15x earnings — a bargain either wayPay for cash flow ✓
CME vs the perps · Fri100-to-1 bets coming for regular investorsNo gambling products ✓
Reformation’s IPO · WkndTwice sales in a 1.7% category; style is not a moatPay for cash flow ✓

The AI columns. The payoff is finally visible, and it lives in the cloud — the giant rent-a-computer businesses. Amazon’s (AMZN) cloud grew 37% with a 39% profit margin; Microsoft’s Azure grew 43%. The rule is blunt: own a cloud, or don’t bother. But the boom now runs on borrowed money — about $250 billion of tech-giant bonds this year, $400 billion expected next — so the columnists advised staying in short-term, high-quality bonds. Apple (AAPL) may be living on borrowed time now that Siri runs on Google’s (GOOGL) Gemini. And SpaceX, with five of thirteen test launches showing failures or concerns, is valued at 28 times its sales.

The electricity column. When Texas ordered an audit of every data-center grid connection, power producers Vistra (VST) fell 8% and NRG (NRG) 15% in one day. Caterpillar (CAT), which sells generators that skip the permit line, rose. AI’s next bottleneck is electricity and permission slips, not chips.

The discipline columns. The stock market’s total value now equals 230% of the U.S. economy — above the dot-com peak — which historically points to smaller future returns. Investors forgive famous money managers too fast; most of the money arrives after the fame, then regrets it. And offshore crypto bets using 100-to-1 borrowed money are coming for regular investors.

The income column was the most useful all week. Over ten years, the highest-dividend stocks returned 3.9% a year; the plain S&P 500 returned 13.2%. Pfizer (PFE), paying a 6.9% dividend that eats nearly all its spare cash, was exhibit A. A high dividend is not income. A dividend the company can afford is income.

The price columns. AstraZeneca (AZN), under 15 times next year’s earnings, looks like a bargain with or without a merger. Reformation’s IPO — twice yearly sales in a category averaging 1.7% over three years — does not. Style is not a moat, and a moat is what keeps competitors from stealing your customers.

The verdict

The Journal’s analysts spent a week writing our checklist back to us. Own the businesses that collect a fee when AI money moves — the clouds — not the promises. Keep bond money short and high-quality: that is our ladder of Treasury bills, the short-term government IOUs funding the next few years of client spending. And buy dividends a company can afford. That is why Chevron (CVX), whose record quarter covers its dividend about twice over, and pipeline operator Williams (WMB) joined our income holdings. It is also why Pfizer went under formal review instead of getting a pass for its yield.

The one warning we take seriously: the AI boom is borrowing hundreds of billions while running into electricity limits. That is exactly why we pair the cloud giants with Caterpillar and refuse long-term bonds. Gold’s 7.2% week was our defensive holdings doing their job in public.

What would make us wrong

If long-term AI bonds do well from here, our caution cost us. If the Texas audit resolves fast and the power producers bounce first, we watched too carefully. If Pfizer’s new drugs restore its spare cash, the review closes with a hold — the review is a question, not a verdict. We write these down now so that in December we grade ourselves against the list, not against our memory.

What It Means For Your Portfolio

No change - it confirms the plan

The Capital Wealth Growth Portfolio passed an outside exam this week, and nothing needs to change.

The moves the columns support were already made: Chevron and Williams joined our income holdings, Pfizer went under formal review, and we keep refusing long-term bonds while the AI boom borrows. Our defensive holdings — gold and Treasury bills — did their job in public, with gold up 7.2% for the week. We keep the checklist and grade ourselves against it in December.

Book a 15-Minute Review → Back to the August 8 Edition →