Research that does the work.

The Week Ahead · Edition 33: Super Wednesday

AI now has to pay for itself.

Wall Street had a second losing week: the S&P 500 fell 0.5% and the Nasdaq 0.7%, while the Dow rose 0.4%. The defining move came on Thursday, when Meta reported revenue up 26% but its first-ever negative quarterly free cash flow after $34.9bn of capex. It fell 7% and the Nasdaq dropped 2.6%, its worst day since early June. Treasuries caught the flight to safety and the 10-year briefly closed below 5%.

In the UK, CPI rose to 3.5% and the flash PMI hit an eight-month high, so a November Bank of England hike is close to certain; the FTSE 100 gave back 1.8%. French 10-year yields passed 5%, the highest since 2002. This week is as heavy as the calendar gets: the Fed, the UK Autumn Budget and Microsoft and Alphabet results all fall on Wednesday, with US GDP and PCE, the ECB and Amazon on Thursday and the Bank of Japan on Friday. The full brief ranks the eight events and sets out the desk’s three views.

Practitioner-led market commentary for information only. Not investment advice, issuer-sponsored research or a personal recommendation. Prices are closes on Friday 23 October 2026. Lundy may hold interests in subjects covered.

EM Week Ahead · Edition 2: $116 oil, a coin-toss vote

Emerging markets fell for a second week: 1.3% in US dollars, 7.2% over two weeks.

Brent rose 6.1% to $116.52, its highest since April, as Houthi strikes on Saudi airports and oilfields spread and the US ordered some diplomats out of Riyadh. That hit oil importers (India −2.4%, Pakistan −2.4%, a record low for the Philippine peso) and the Gulf, with Saudi Arabia down 4.1% for an eighth weekly loss.

Doubts about AI spending returned on Thursday when the Nasdaq fell 2.6%, but Korea held up better (−2.1%) as SK hynix completed its buyback. Central banks pulled in opposite directions: the Philippines hiked 50bp, Korea held with a hawkish tone, and Turkey cut 100bp, after which the BIST 100 fell 5.2% on the week. Jakarta (+4.9%) and Vietnam (+4.0%) rebounded. The full edition covers 28 markets, the scoring of last week’s calls and the week ahead: Brazil’s runoff result, the Fed, and SK hynix and Samsung results.

Practitioner-led market commentary for information only. Not investment advice, issuer-sponsored research or a personal recommendation. Prices are closes on Friday 23 October 2026. Lundy may hold interests in subjects covered.

Weekly Commodity Report · Edition 228

Attacks in Saudi Arabia lifted oil, oil lifted rate-hike bets, and those bets hit gold.

Attacks in Saudi Arabia’s Eastern Province forced Aramco to evacuate its Dhahran headquarters on 22 October. Brent rose 6.2% to $116.72, its highest close since late April and up 94% this year. $116 oil revived Fed hike bets, with futures pricing about a 76–78% chance of a December move, and gold fell 4.1% to below $4,000 for the first time since July; silver fell 5.9%.

Warm forecasts sank US natural gas 15.3% to a two-month low. Corn fell to its lowest close since July as the harvest reached 44% and exports lagged, but China bought 2.18 Mt of US soybeans in one week, the most in nearly three years. Six markets rose. The full edition covers each market, with the week ahead: the Fed, US GDP and PCE, and OPEC+ on 1 November.

Practitioner-led market commentary for information only. Not investment advice, issuer-sponsored research or a personal recommendation. Prices are closes on Friday 23 October 2026. Lundy may hold interests in subjects covered.

Weekly Cocoa & Coffee Report · Edition 16

Three days of buying on the Ivorian arrivals, then one day of selling on the exchange stocks.

Ivory Coast port arrivals for 1–18 October were 78,212 t, down 41% on a year earlier, as heavy rain cut roads and the farmers’ strike continued. Cocoa rose three days running, New York to $5,903 and London to £4,395. Then ICE US stocks, up every day of the week to 3.68m bags, the highest since May 2024, took over: on Thursday New York fell $261 and London £164, wiping out the week’s gains in one session.

New York closed at $5,592 and London at £4,193. In coffee, robusta fell 5.4% to $3,228, a 52-week low close, as Vietnam’s harvest picked up and ICE robusta stocks rose for a sixth week; arabica held a small gain. The full edition covers arrivals, exchange stocks, positioning and the week ahead, including the Fed and Hershey and Starbucks results.

Practitioner-led market commentary for information only. Not investment advice, issuer-sponsored research or a personal recommendation. Prices are closes on Friday 23 October 2026. Lundy may hold interests in subjects covered.

Nigeria Resource Research · Paper 1: The Neglected Endowment

“The endowment does not need inflation to be significant; it needs definition to be financeable.”

Nigeria’s resource endowment is real and unusually broad. Few countries hold a top-tier position across so many tables at once: the world’s largest producer of cassava and yam, roughly 40% of the global shea crop, the second sesame exporter by value, a top-three gum arabic producer, top-five in cocoa and oil palm, and the largest barite reserve among 44 documented solid minerals. Yet much of this is recorded as geological occurrence rather than compliant, bankable reserve, and the headline valuations attached to it are frequently promotional. A high place in the production table beside a low one in the value table is the rule across the endowment, not the exception.

The paper builds the global benchmark against which any producing country should be read: the world market for each resource group, and the conditions that separate countries which capture resource value from those which export it raw. The central finding holds in every resource examined: the value lies downstream of extraction. And the history sharpens the point. Nigeria produced 45% of the world’s palm oil in 1961 and was Africa’s largest rubber producer in 1966; each industry was built, then lost. Withdrawn, not incapable. With the currency regime making dollar generation an imperative, diversification is now forced rather than optional; what stands between the endowment and its execution is definition, capital and capability.

Neutral market and resource analysis. Not investment advice. Lundy may hold interests in subjects covered.

The Great AI Listing: anatomy of a $220 billion IPO supercycle

Five listings. US$220 billion. A capital-absorption event without modern precedent.

Executive summary · full 10-page note in the PDF

Between June and October 2026, US equity markets face a capital-absorption event without modern precedent: five companies preparing IPOs targeting roughly $220 billion in primary raises at a combined implied valuation above $3.4 trillion: nearly five times the entire 2025 US IPO market, concentrated in five listings. The note maps the cycle against the dot-com era and places today’s market at roughly late-1997: the narrative established, the manic final phase not yet arrived, with one structural difference, the earnings floor (NASDAQ-100 at ~30× earnings versus ~200× in 2000), and one uncomfortable similarity: the listing companies themselves are deeply loss-making at trillion-dollar valuations.

It then works the liquidity mechanics: total incremental capital demand approaching $1.5 trillion in six months once Treasury and corporate issuance are counted. It identifies the likely casualties of reallocation (small and mid-caps, non-AI tech, emerging markets, long duration, speculative assets), and sizes the underwriting fee pool for the lead banks. The house view is probabilistic:

Five actionable signals close the note: SpaceX first-week trading (holding above $135 and appreciating accelerates the pipeline; trading below issue within two weeks recalibrates the whole calendar); the 30-year Treasury yield (a sustained move above 5.25% compresses the growth-equity risk premium); the quarterly IPO deal count (any quarter below 40 deals echoes the dot-com crack); the lock-up expiration calendar (cascading insider supply in H1 2027 is the highest-probability correction catalyst); and the AI revenue-to-capex ratio (when it begins to improve, the fundamental case strengthens materially).

Practitioner-led market commentary. Not issuer-sponsored research or personalised advice. Lundy may hold interests in subjects covered.

The Arc of AI: a five-phase investment framework

We do not treat AI as a technology theme to be indexed, but as a sequence of capital cycles and sectoral realignments, each with distinct beneficiaries and casualties.

The report sets out a long-range investment map for the Arc of AI: the staged transformation of global business, labour, capital flows and market structure under frontier artificial intelligence. It builds on, and meaningfully sharpens, Aschenbrenner’s 2024 thesis, absorbing that roadmap but tilting from timelines to balance sheets: tracing the capital flows and bottlenecks through which the transition materialises. The macro numbers are large ($15.7 trillion of estimated GDP contribution by 2030; over 60% of large enterprises already testing agent tools) but the central finding is that distribution matters more than magnitude. The result is not AI uplift everywhere; it is bifurcation.

Five phases, each with a discrete locus of impact: semiconductors and hyperscalers (the opening bid, now largely played out and fully valued); data infrastructure and energy (the most capital-intensive and most durable phase (land, electrons and copper cannot be obsoleted by software); enterprise adoption, where efficiency turns into earnings and the story is margin, not revenue; AI employees, where software stops supporting work and becomes the work; and real-world AI, as autonomy leaves the screen and reprices manufacturing, logistics and defence. Leadership rotates as each phase matures; the investor’s job is to anticipate the sequence and position ahead of it.

Practitioner-led market commentary. Not issuer-sponsored research or personalised advice. Lundy may hold interests in subjects covered.