Macro Context Brief — verified as of 2026-09-15

Macro Context Brief — verified as of 2026-09-15

Shared framing for all industry research. Every figure below was retrieved live on 2026-09-15 from the named source. Treat as context, not as findings to re-report.

Monetary / macro (US)

• FOMC held the federal funds target at 3.50–3.75% at the July 28–29, 2026 meeting, on a 9–3 vote, with Hammack, Kashkari and Logan dissenting in favour of a 25bp HIKE. Source: Federal Reserve, FOMC minutes, 2026-07-29. https://www.federalreserve.gov/monetarypolicy/fomcminutes20260729.htm
• Total PCE inflation 4.1% (May, 12-month), est. 3.7% (June). Core PCE 3.4% → est. 3.3%. Unemployment 4.2% (June). Wage growth ~3.5%.
• FOMC-identified inflation drivers: tariff pass-through, Middle East conflict disruption, and "the surge in demand related to the AI buildout."
• FOMC explicitly flagged financial-stability risk from high AI-firm equity valuations and increased leveraged financing of infrastructure buildout.
• Implication for all sectors: this is a rising-rate-risk, sticky-inflation environment, not the disinflationary easing cycle of 2024–25. Any trend thesis that assumes cheap capital must be stress-tested. The September 16, 2026 FOMC decision falls one day after this research date and is UNRESOLVED — do not assume an outcome.

Capital formation

Two authoritative trackers disagree; both are reported, neither is treated as truth.

• Divergence is a methodology difference (deal inclusion, corporate/PE rounds, announced vs closed), not an error. This is the canonical worked example for the source-quality framework.
• Both agree on direction: record half-year, collapsing deal count. KPMG: 8,440 deals in Q2 is the lowest count since Q3 2017.
• Extreme concentration. OpenAI + Anthropic alone took $217B = 43% of all H1 2026 funding (Crunchbase). Top 10 deals = $105B = 46% of Q2 (KPMG). >70% of Q2 capital went to AI-focused companies, up from ~50% a year earlier.
• Americas = 66% of Q2 funding. Asia $50.8B (18-quarter high, 5th consecutive quarter of growth). Europe $25.6B — the structurally weak leg.
• Sources: https://news.crunchbase.com/venture/global-startup-exits-ipo-ma-soar-ai-q2-h1-2026/ and https://kpmg.com/xx/en/media/press-releases/2026/07/vc-investment-already-at-five-year-high-of-billions.html

Exit window reopened, violently

• SpaceX IPO, 2026-06-12, Nasdaq: SPCX. Fixed price $135/share (no book-built range), raised $75B at a $1.5T valuation, closed day one at $161.11 (+19%) for ~$2.1T market cap. Largest IPO in history by an order of magnitude (prior record: Facebook, $104B, 2012). Source: https://news.crunchbase.com/public/spacex-record-breaking-ipo-spcx/
• Q2 2026 global exit value reached $1.9T, exceeding the previous annual record; US = $1.8T of it. 32 IPOs above $1B; 24 M&A deals at/above $1B; $113B quarterly M&A value.
• Anthropic and OpenAI public listings were reported as anticipated/concurrent. Status as of this research date is unconfirmed — verify before asserting.

Trade and export control

• 2026-01-14: US BIS moved advanced-compute export licensing for China/HK/Macau from presumption of denial to case-by-case review, covering H200 / MI325X-grade parts (ECCN 3A090; TPP <21,000, DRAM bandwidth <6,500 GB/s). Conditions: US-based third-party lab testing pre-shipment, end-use certification, and an aggregate cap at 50% of US shipment volumes. Enhanced verification for IaaS providers.
• 2026-01-14: A 25% Section 232 tariff applies to specified advanced semiconductor imports, with broad exemptions — data centres >100MW AI-dedicated load, startups and R&D, consumer electronics, industrial, US public sector, repairs.
• Source: Wilson Sonsini, https://www.wsgr.com/en/insights/a-mixed-bag-of-chips-significant-new-import-and-export-changes-for-advanced-semiconductors.html
• Implication: export policy loosened on the China demand side while import policy tightened on the cost side. Both directions are politically reversible — treat any supply-chain forecast built on either as low-persistence.

AI regulation (EU)

• The Digital Omnibus on AI entered into force 2026-07-27.
• AI Act transparency obligations took effect 2026-08-02 as scheduled.
• High-risk obligations for employment-context AI (recruitment, performance evaluation, worker monitoring, promotion, termination) were deferred from 2026-08-02 to 2027-12-02 — a ~16-month postponement pending harmonised standards.
• Source: DLA Piper, https://knowledge.dlapiper.com/dlapiperknowledge/globalemploymentlatestdevelopments/2026/The-Digital-AI-Omnibus-Proposed-deferral-of-high-risk-AI-obligations-under-the-AI-Act
• Implication: the single most-cited 2026 compliance deadline in vendor marketing (Aug 2, 2026 high-risk) did not bind on schedule. Any dossier claim that leans on it is stale. This is a live example of regulatory-deadline slippage as a trend-killer.

Standing cautions for every industry agent

1. Sticky 3–4% inflation + hawkish-lean Fed = capital cost is a live constraint.
2. AI funding concentration means sector funding totals are badly skewed by 2–3 companies. Report medians and ex-AI figures where you can find them.
3. Tariff and export-control regimes changed in Jan 2026 and remain politically volatile.
4. Middle East conflict is an active input to energy and food prices per the FOMC.
5. The Sept 16, 2026 FOMC decision is one day after this research date — unresolved.

ADDENDUM — verified by sector agents, added 2026-09-15

Material events surfaced during sector research 01–12. Later agents must build on these rather than rediscovering them. Each is traceable to a sector dossier's source list.

Trade law was rewritten by the courts, not by policy

• The Supreme Court voided the IEEPA-based tariffs. The tariff architecture has been rebuilt on sectoral authorities (Section 232 and similar) rather than emergency powers. Verified in sector 09.
• Refunds are flowing and they are material and non-recurring. Target booked $994mn — about 3.7pp of margin, ~$1.65 of FY EPS; Walmart's 96bp gross-margin gain was "primarily" refunds. 2026 retail and industrial margin expansion is partly a court-ordered one-off. Anyone extrapolating it into 2027 is wrong. Verified in 12.
• De minimis is closing on both sides of the Atlantic. The US Court of International Trade upheld the US suspension on 2026-08-13 on a "privilege revocation" theory that survived the IEEPA reversal. The EU imposed €3/item from 2026-07-01, covering 5.9bn items, 91% of them from China. Verified in 12.
• Effective-tariff-rate figures are routinely conflated in press coverage: Penn Wharton 6.7% "effective" vs Tax Foundation 7.2% effective / 11.8% "applied". Definitional, not a factual dispute. Verified in 09.

The Middle East conflict has a specific, measurable transmission channel

• The Strait of Hormuz closed. Hapag-Lloyd took a ~$600m Q2 hit with six vessels trapped; IATA cut its forecast. This is the concrete mechanism behind the energy and freight inflation the FOMC named. Verified in 09.
• It also invalidated the February 2026 container-shipping bear case (Maersk's first loss since 2017) — by event, not by modelling error.

US energy and climate policy reversed within the year

• OBBBA ended clean-energy tax credits on 2026-07-04, with a four-year placed-in-service limit on safe-harboured projects. Verified in 05.
• EPA rescinded vehicle GHG standards effective 2026-04-20, destroying the regulatory-credit profit pool that cross-subsidised several EV businesses. Verified in 10.
• The US EV tax credit expired; 2026 is the first year of declining annual US BEV sales and share (6% of Q2 light-duty vs 7%), with hybrids at a record 16%. Verified in 10.

An inflation measurement discrepancy worth noting

• Core CPI 2.4% (Aug 2026) against core PCE ~3.3%. That is an unusually wide gap and in the unusual direction. Do not treat "inflation" as one number; say which index. Flagged in 12, unresolved — record it as a contradiction if you rely on either.

SpaceX as a public company — new Tier A disclosure

First IR disclosure for what is now the sector's largest company (sector 08): Q2 2026 revenue $7.814bn, of which the Space segment is only $962m and is the only loss-making core segment (−$542m operating). Starlink at 12.0m subscribers with ARPU down from $85 to $66. SPCX closed below its IPO price on report day and is −30.3% from debut. The largest IPO in history is not trading like a success.

Capital markets: the paper/cash divergence

Sector 11 established the single most important correction to the macro framing above: record deployment and record exit value have not produced cash. 2021-vintage DPI is 0.05x; net LP cash flow since 2022 is −$202B; four companies account for 93.5% of 2026 exit value; seed→Series A graduation fell from 55%+ to 16% for the 2024 cohort. Aggregate dollars look euphoric; the median company's experience is the opposite. Never cite the headline VC totals without this.

Standing instruction for remaining agents

If WebSearch returns a budget-exhausted message, do not stop. Switch to WebFetch on primary-source URLs (regulators, statistics agencies, company IR, trade bodies — most are directly fetchable and do not consume the search budget), complete the deliverables, and record the reduced discovery capability explicitly in data_gaps and in dossier §13 so the sector can be prioritised for re-run.

ADDENDUM 2 — added 2026-09-15 after sectors 13–18

Further verified events

• Antitrust outcomes went the incumbents' way, and enforcement moved sub-federal. On 2026-09-02 the court declined to break up Google's ad stack — a behavioural-only remedy, with divestiture of AdX and DFP rejected outright (sector 16). Meanwhile state attorneys general have halted the DOJ-cleared $110bn Paramount/WBD merger (sector 14). Same pattern as the IEEPA reversal: courts and states, not federal policy, are setting the boundary conditions in 2026.
• App-store economics fragmented by jurisdiction inside nine months — the flat 30% commission was dismantled or altered across the US, EU, Japan, China, Korea, UK and Brazil (sectors 15, 17). Apple: EU 5% CTC from 2026-10-01; China 25%/12% from 2026-03-15. Any model of consumer app net revenue must now be jurisdiction-specific.
• Public-market disclosure is shrinking. EA completed its PIF/Silver Lake/Affinity take-private on 2026-08-04, removing a top-five publisher's quarterly reporting. Note this as a data-availability trend in its own right.
• Rare-earth export licensing is a robotics and industrial constraint, not just a defence one — NdFeB magnets are ~80% Chinese and sit in every servo motor; Japan received zero covered exports in July 2026 (sector 13).
• Data-centre construction is crowding out other construction through shared labour and switchgear, now confirmed from both the spending side (09) and the labour side (18): construction unemployment at a record-low 3.1% despite a soft market, with fewer than half of US metros adding construction jobs year over year.

Statistical traps confirmed across multiple sectors

1. US Census reports data-centre construction inside the "Office" category. This is why Census office construction is +21.3% y/y in the middle of an office-distress narrative. Any sector touching nonresidential construction data must decompose it.
2. Core CPI 2.4% (Aug 2026) vs core PCE ~3.3% — still unreconciled. Always name the index.
3. Date laundering in search results. Searching a current metric often surfaces a two-year-old article in the top results. Check the publication date of every result before using it, and prefer the primary issuer's own current page.
4. Rounded cumulative disclosures cannot be differenced. YouTube's "$100bn paid to creators over 4 years" was reported at the same $100bn a year earlier — the figure is rounded, so no annual growth series can be derived from it. Applies to any "cumulative paid out" claim.
5. Aggregate vs median. Established in sector 11 and true everywhere: totals are carried by a handful of outliers. Report medians and counts alongside sums.

Retrieval playbook (practical, learned the hard way)

• SEC EDGAR: browse-edgar is robots-disallowed and efts.sec.gov full-text search may 403 through the proxy. Use data.sec.gov/submissions/CIK##########.json to list filings, then fetch the document under /Archives/. This works.
• Corporate IR: index pages are usually JS-rendered and fail. Individual press release detail pages and PDF links fetch reliably. Go straight to the release URL.
• Government statistics: direct file paths work well — e.g. the Census e-commerce PDF at census.gov/retail/mrts/www/data/pdf/ec_current.pdf. Navigation pages often don't.
• Known-blocked at time of writing: congress.gov, FDIC QBP pages, NPCI (robots.txt), Banco Central do Brasil (JS), Box Office Mojo (malformed tables), CISA KEV JSON, cve.org. Record these in the source registry as licensed/alternative-access needs for Phase 2.
• High-value structured dataset not yet exercised: the EU DSA Transparency Database (transparency.dsa.ec.europa.eu) — daily bulk download, common schema across TikTok, Meta, Snap, Pinterest and X. Flagged by sector 17 as the best structured source in its sector. Phase 2 should wire it first.

Sectors flagged for re-run at full search budget

13, 14, 15, 16, 17 and 18 completed with 5–11 searches each instead of ~22 because the shared pool was exhausted. Their Tier-A verification is sound; their discovery is narrow. This is recorded in the QA register, not hidden.

More research from the library · Explore trend records