← All research中文
01H1 2026 review · H2 outlook

One causal chain for US equities in the second half of 2026

The index is at a high, but what was holding it up has lost its footing. Seven pages build a chain; five more show two places on that chain where you can act.

Macro · eight steps · two stocks · about 20 minutes

Disclaimer · This material is for personal research and educational discussion only. It is compiled from public information and does not constitute investment advice, an offer, or a recommendation, nor a basis for buying or selling any security. The presenter may hold positions in the securities mentioned. Investing involves risk; make your own judgement and bear the consequences of your own decisions. Data as of 26 August 2026.

02Start with one observation

Same market. Why did tech crack first?

QQQ · drawdown from high
−4.1%
Nasdaq 100
SPY · drawdown from high
−1.1%
S&P 500
The conclusion first. The index is still high, but the assumption that pushed it there, "the Fed will cut", no longer holds. What the committee is now debating is whether to hike.

Same market, same period, and the two drawdowns differ by almost four times. This is not a broad sell-off; it is money changing legs. The next seven pages answer why.

Source: IBKR weekly closes to 2026-08-26

03The map

One chain, eight steps

Each step is explained by the one before it; the order is the causation. The three orange steps are the key nodes.

1War in the Middle EastIran war, 2/28
Hormuz blockaded
2Energy prices upThroughput 21.6M → 4.9M bpd
5.5M bpd of capacity offline
3CPI splitsHeadline 3.4%
Core only 2.5%
4DeadlockHikes can't fix supply
but cuts are off the table
5The Fed is stuckJuly FOMC 9:3
three votes to hike
6Bonds reprice2-year 4.24%
61bp above the policy rate
7Discount rate rises10-year 4.70%
fair multiple 20× → 17.5×
8Long duration gets hit firstQQQ −4.1%
SPY −1.1%
04Steps 1 to 4

This inflation is imported, not overheated

Headline CPI is pushed to 3.4% by energy while core is only 2.5%. The gap between them is food and energy.

Rate hikes can make Americans buy one fewer fridge. They cannot make the Middle East pump one more barrel.

3.4% = 79% × 2.5% + 21% × X
→ food + energy ≈ +6.8%

Approximation, backed out from published weights.

But 3.4% is the number on the tape, and inflation expectations form on the headline. The Fed has no tool for this and no room to let go. That is step 4, the deadlock.

Sources: July 2026 FOMC statement · BLS July 2026 CPI report

05Steps 5 to 6

Not "the cutting guidance is gone". Hikes are on the table.

FOMC, 28–29 July: rates unchanged, but three members voted to hike.

Hold
9 votes
3.50%–3.75%, fifth hold of the year
Hike +25bp
3 votes
Hammack · Kashkari · Logan
"Many participants judged that if inflation did not come down, tightening might become necessary."
July FOMC minutes

The market's next expected cut is now in early 2028.
RateLevelRelative
Effective fed funds3.63%
2-year Treasury4.24%61bp above
10-year Treasury4.70%term spread +46bp

The 2-year is roughly the average expected policy rate over the next two years. It is above today's rate, so the bond market's priced next move is a hike. Bonds have already caught up.

Sources: July 2026 FOMC statement and minutes · Federal Reserve H.15, 2026-08-24

06Step 7

The discount rate: the ruler every stock is measured with has changed

The Fed didn't hike, yet the long end kept rising, because the cut was removed and the market now demands more term premium.

Discount rateFair multiple
Start of year9.0%20.0×
Now9.7%17.5×
Change+0.7pp−12.3%

Assumes perpetual earnings growth of 4%; multiple = 1 ÷ (discount rate − growth).

Not a dollar less in profit, and fair value already fell 12%.

Federal Reserve H.15 puts the 10-year at 4.70% on 8/24; that figure is used for the valuation arithmetic.

07Step 8

Duration: the later the cash arrives, the harder rates hurt it

This is the step that turns an abstract rate into gains and losses you can see in your account.

Cash flow in yearRate 9.0%Rate 9.7%Change
10.9170.912−0.6%
50.6500.629−3.2%
100.4220.396−6.3%
200.1780.157−12.2%
300.0750.062−17.6%

Present value of $1 = 1 ÷ (1 + r)^n. Standard discounting arithmetic, my own calculation.

A company that sells cola
Earns this year and next; cash flows sit at the near end, like the "year 1" row.
An AI company
Valued on what happens after 2030, like the "year 20 / 30" rows.
Back to the opening question. QQQ −4.1% and SPY −1.1% is not a coincidence; it is this table showing up in the real market.
08Where the chain goes next

No index targets. One question: does the chain break, stall, or keep pushing?

ScenarioHow the chain movesProbabilityMarket
A · breaks at step 2Conflict eases, energy falls → headline CPI converges on core → cuts back on the table45%Discount rate falls, growth rebounds
B · stalls (base case)Inflation stuck around 3%, Fed on hold, 10-year between 4.5 and 5.040%Index sideways, violent rotation underneath
C · keeps pushingAnother energy shock → an actual hike15%Multiples compress 10–15%
Note that A + B add up to 85%. I am not bearish. What I am saying is that the money from multiple expansion is done and we are back to earning the money from earnings. That is not a crash. The right move is to change the structure of the portfolio, not to sell it.
09One chain, two entry points

The stocks aren't a new topic. They are two places on this chain where you can act.

Attached to step 2, the cause · the hedge
EOG Resources
Rather than hide downstream, own the cause. Energy keeps rising → the chain accelerates → growth gets hit → but this one makes money.
Attached to step 8, the consequence · the offence
Alphabet (GOOGL)
High rates punish long duration. The question is whether, in 2026, you can avoid it at all.
You can't. That is the real question of 2026. Microsoft, Meta, Amazon and Google are expected to spend more than $700B of capex in 2026 combined. AI capex has deliberately lengthened the duration of the entire large-cap tech sector. Step 8 says to pick the shortest-duration growth stock, but that filter now returns almost nothing.

So the real question becomes: if you have to carry this duration, whose odds are best? My answer is GOOGL, because it is the only one that can show evidence of a return.
10GOOGL · the bad news first

It genuinely fails a test I set myself

Alphabet is issuing stock to fund a ten-year infrastructure bet. Of my three criteria, it clearly fails the second.

Free cash flow turned negative
−$5.86B
Q2 2026, first time in years
Buybacks to zero
$0
Two straight quarters · $28.3B in H1 last year
Equity raised
$89B
Incl. $10B Berkshire private placement · ~2% dilution
Off-balance-sheet commitments
$707B
Up $474B in a single quarter

I pick it anyway; the reason is on the next page. These come first because you can look them up. Sources: Alphabet Q2 2026 earnings and 10-Q (SEC) · June 2026 equity offering press release

11GOOGL · the evidence

But it is the only one with evidence of a return

Burning cash is not the problem. Burning it and producing nothing is.

Cloud revenue
+81.8%
Year on year
Cloud operating margin
35.6%
20.7% a year ago
Cloud backlog
$514B
Remaining performance obligations
Search advertising
+16.8%
No sign in the filings of AI eroding search

TPUs are already being sold to external customers including Anthropic and Meta; revenue is recognised mostly from 2027.

Don't read the P/E straight off the report. It shows net income +298% and a trailing P/E of only 17.4×. That is an illusion: $77.1B of it is a one-off mark-to-market gain on the SpaceX IPO in June. Strip it out and the P/E is about 25.1×.

Sources: Alphabet Q2 2026 earnings and call · valuation data from stockanalysis.com, 2026-08-26, same basis

12EOG

This is not a bet on oil. It is position insurance.

If my macro call is wrong in the worse direction, I need something that is making money.

Break-even oil price
< $50
WTI · CFO's words
Net debt / EBITDA
0.4×
Credit rating
A− / A3
Shareholder return pledge
≥ 70%
of free cash flow
Two costs that have to be said. ① It is already up 40%; you are buying the hedge after the rally. ② The EIA's official forecast has Brent back at $69 in 2027. The supply shock is unwinding, and in that world this position loses money.

Sources: EOG Q2 2026 earnings and call · EIA Short-Term Energy Outlook, August 2026

13The two together

Their value is not which one is better. It is that something makes money whichever scenario arrives.

ScenarioGOOGL · step 8EOG · step 2
A · conflict eases (45%)Big gainLoss
B · high-level stalemate (40%)Earns the earningsSmall gain
C · an actual hike (15%)Gets hitGain
1
Not a bear market. The end of multiple expansion.
2
The core tension is energy inflation plus a Fed that won't backstop.
3
The portfolio moves from "buy duration" to "buy cash flow and buy inflation".

The right move is to change the structure, not to sell everything.

All data as of 26 August 2026, from the Federal Reserve, BLS, EIA, SEC filings and IBKR real-time quotes. Personal research, not investment advice. Investing involves risk; make your own decisions.