Macro triggers — AI capex · rates · energy · breadth

Snapshot Fri, 16 Oct 2026 09:07:11 GMT · refreshed automatically by Cloudflare cron, 08:00 Sydney daily · data: Yahoo Finance & FRED, no API keys · manual fields last updated 16 Oct 2026 (Q2 FY26 earnings round)
The bear case is live 3 red · 0 orange · 2 amber · 0 green — of 5 triggers Refresh now
1 · AI capex funding (hyperscaler free cash flow) manual broken
-$5.1bn reported FCF (2 of 4 names)
The two names that have reported are already cash-flow negative — the buildout is funded by debt and vendor financing, not operations. Amazon reported negative too but unquantified; Microsoft is the swing name to verify.
Alphabet $-5.9bn (reported; buybacks suspended) · Amazon: reported negative in Q2; not quantified in sources I have · Meta $0.8bn (reported; from ~$13bn a year earlier) · Microsoft: not in sources I have; capex $30.9bn/quarter
Watch: Does FCF turn positive again in the Q3 round? That single line kills or confirms the bubble thesis.
2 · AI revenue vs break-even manual broken
$230bn short
Needs ~$300bn/yr of AI revenue to break even on the buildout; running ~$70bn above pre-AI trend.
Goldman: ~$300bn/yr needed to break even; cloud revenue annualising ~$70bn above pre-AI trend
Watch: The gap has to close by revenue arriving — or by capex being cut. Watch cloud backlogs (RPOs), not headlines.
3 · US 10-year Treasury yield watch
5.19%
Bands: under 4.75% bull-friendly · 4.75–5.50% neutral · 5.50–6.00% pressure · above 6.00% breaks the bull case.
FRED DGS10 · 2026-10-14
Watch: A sustained break above 6.00% is the trigger that forces the AI paper to reprice.
4 · Brent crude (Iran / Hormuz risk) watch
$102.54
Bands: under $85 benign · $85–115 chronic risk premium · $115–130 growth hit · above $130 the IMF downside case.
52-week range $58.72–$126.10
Watch: Trump's pre-midterm restraint on Iran expires after 3 Nov. A true Hormuz closure is the fastest route to $130+.
5 · Breadth (equal-weight vs cap-weight) broken
20.7 pctile · S&P +6.6% vs 200dma
RSP/SPY percentile over 12 months plus the S&P's distance from its 200-day. Narrow, cap-weighted rallies are the late-stage signature.
RSP/SPY = 0.27857 · S&P 200-day = 7268
Watch: Breadth improving while the index rises = healthy. Index up while breadth falls = the warning in my 2027–28 call.
Cross-checks
S&P 500
7,744
+6.6% vs 200dma
Nasdaq
26,713
concentration engine
ASX 200
8,807
52w high 9,297
Gold
$4,208
-25% from record
Brent
$102.54
52w $59-$126
US 10-year
5.19%
FRED DGS10 2026-10-14
VIX
16.03
complacence
AUD/USD
0.6963
52w 0.642-0.728
Thesis in one line: a debt-financed AI capex cycle decelerating from 100% to 54% to 12% growth (2026→2028), funded ~35% by borrowing and $3.1tn of off-balance-sheet commitments against a ~$300bn/yr revenue break-even — with a structurally higher inflation regime (2.5–3.5%) and energy as the binding constraint. Expect a 25–40% tech drawdown between mid-2027 and mid-2028, then a bull market led by compute and power.

Falsifiers: hyperscaler FCF turning positive again, the AI revenue gap closing below ~$75bn, the 10-year back under 4.75%, or Brent under $85 with Hormuz normal — any two of those and the bull case runs longer than I expect. Accelerants: a 10-year above 6.00%, or Hormuz closing (Brent > $130, which the IMF models as −0.5pp of 2027 growth and 5.2% inflation).

Not investment advice. raw JSON · source diagnostics