Clearing the Fog: New Narratives Worth Watching

 

The Weekender is my bi-weekly take on macro shifts and emerging themes. It’s not investment advice — or even our firm’s official view. I aim simply to inform, challenge, and maybe entertain. If you’d like this in your inbox every other Saturday morning via Northern Trust, subscribe to The Weekender.

Iran headlines have sucked the oxygen out of the room and consumed our attention. Whether the pause holds or not, it creates space to draw breath, for new narratives to emerge, and for investors to identify what is likely to endure. This week, we focus on drivers that may have been obscured by the fog of war (Military Keynesianism, AI acceleration, growing scarcities, multiple compression, etc.) and attempt to sketch the headlines that might matter a few months from now.

Military Keynesianism and Ongoing Debasement

President Donald Trump’s budget request for fiscal year 2027 includes $1.5 trillion in defence spending, a $445 billion (42%) increase from FY2026 — the largest peacetime defence hike since WWII. If enacted, the incremental spend alone could add roughly 1.5% to gross domestic product.

This is Military Keynesianism in its purest form, and it will require funding — most likely via increased Treasury issuance. More debt implies further debasement, and with it, rising demand for supply-constrained (and perhaps sanction-free) real assets (perhaps explaining why China accelerated their gold purchases last month) Those same assets will likely find growing support from countries prioritising strategic autonomy and resilience over efficiency, particularly during periods of geopolitical stress — whether through stockpiling or resource nationalism.

This comes just as the largest construction project in history — AI — continues to accelerate. And yet, most public resource equities still trade at a fraction of the multiples awarded to sectors increasingly reliant on them for growth. This year alone, roughly $750 billion is being spent on AI capex versus an estimated $550 billion on finding new oil. One of these numbers feels too low — good news for companies with real picks, pipes and shovels.

AI Is Still Accelerating

The numbers this week were remarkable. OpenAI’s ChatGPT ads pilot in the United States crossed $100 million in annualized revenue within six weeks of launch. Anthropic announced it surpassed $30 billion in annualised run-rate revenue in April, up from ~$19 billion just a month earlier and ~$9 billion in December. These. Are. Astonishing. Especially considering we may still be early in the adoption phase given only ~3% of households (per Bank of America) currently pay for AI services, while the vast majority already own a mobile phone.

Even more striking: OpenAI raised $122 billion in a private placement — the largest capital raise in history — at an $852 billion post-money valuation. To put that in context, it exceeds what SpaceX hopes to raise in its record-breaking IPO (reportedly $50 billion–75 billion at a valuation north of $1.5 trillion), and 4x that of the previous record, Saudi Aramco ($29 billion in 2019).

These figures are both astonishing and a little unnerving. But they carry an important message: Capital is not the binding constraint for the AI roll-out. Open AI should be able to honour its spending commitments — and I can’t help wondering whether that reassurance has something to do with investment-grade and high-yield bond spreads tightening back to levels last seen before the Iran conflict began.

Peak Fear?

Despite our recent concerns around a data centre backlash and AI deployment becoming politicised, it nevertheless feels like we may be approaching peak AI doomerism.

From an environmental perspective, fracking was arguably more damaging — yet it navigated the PR well enough not only to survive, but to thrive. Their charm offensive worked. Perhaps AI could learn from that?

What’s notable is that many of the most vocal doomers are the very companies that stand to benefit most from AI adoption — OpenAI and Anthropic included. Fear, it turns out, is a powerful sales pitch. Bloomberg tech columnist Parmy Olson highlights research dating back to the 1940s showing that acknowledging an argument’s weaknesses actually makes it more persuasive. Signaling honesty reduces suspicion and builds trust. Doomerism, in other words, is marketing. Worth keeping in mind?

As with the war in Iran, we should be careful not to react to every headline, especially the negative ones. Look hard enough and you’ll find counter-narratives. For example, there is yet little evidence of AI-driven job destruction, even among younger workers. Recent MIT research finds zero observed AI adoption in tasks requiring empathy, presence, creativity or authoritative judgment. Demand for software engineers is up year-to-date, as is demand for accountants — another supposedly “at-risk” profession. For accountants, the more pressing challenge is a demographic cliff, with a large portion of the workforce nearing retirement. In other words, AI may not be arriving soon enough for your tax return. Demographics are an equally, if not more powerful force than technology. Something the doomers overlook.

Multiple Compression and Range-Bound Markets

One of the most debated topics in recent meetings has been the price to earnings (PE) cycle and the anatomy of range-bound markets. Two enduring references remain: Vitaliy Katsenelson’s Active Value Investing, Making Money in Range Bound Markets and Robert Hagstrom’s Who’s Afraid of a Sideways Market?

Market multiples are cyclical.

History shows extended periods where multiples compress as prior secular excesses are worked off. In such regimes, earnings growth, rather than multiple expansion, tends to drive returns. As the required margin of safety increases — particularly amid rising discount rates and inflation volatility — the value factor becomes more prominent (see the 1970s for clues).

Several forces are compounding this: capex is replacing buybacks; new issuance must be absorbed (SpaceX and others); AI disruption is eroding terminal values of large software companies, and a shift from asset-light to asset-heavy business models mechanically lowers multiples.

So, we are not surprised by recent multiple compression — but we are struck by its speed. Technology has all but surrendered its premium in only a few months, with several Magnificent 7 names now trading close to market multiples, despite strong earnings growth and ongoing analyst upgrades. Earnings season starts next week, which should help resolve whether earnings are too high — or prices too low.

Own the Rails, Control the Rewards

We perhaps shouldn’t become overly concerned that capex will replace buybacks and remove market support. Reinvestment ultimately drives productivity — the key source of earnings growth going forward — rather than the financially engineered growth of the past decade, powered by cheap capital and debt-funded buybacks.

And narratives change faster than we expect. Alphabet’s search revenues once looked imperilled; then came Gemini. Walmart was meant to be disrupted by Amazon; it now trades at a +50% premium. Companies adapt. The narrative today is “AI is eating SaaS.” Tomorrow it may be “SaaS is monetising AI.”

We may already be seeing early signs. Last week, a major software provider announced plans to deliver a multi-large language model (LLM) solution — at a higher price point — allowing LLMs to cross-check each other’s output to improve trust and reduce hallucinations. Smart: it increases model usage, drives compute demand and reminds markets the value isn’t just in compute, data or cash flows — it’s in orchestration and distribution. Those who own the rails may also own the rewards.

Remember, the greatest stock-market bubble ever blown was railroads in the early 1900s, which at one point comprised over 60% of index value. And as finance scholar Elroy Dimson reminds us in the most recent Global Investment Returns Yearbook, owning that sector from then until now would have outperformed the broader market. Could history rhyme? Read more in The Past is Prologue.

Real Chokepoints, Real Leverage?

Chokepoints have become quite popular. There’s even a book about them, see Edward Fishman’s Chokepoints: How Economic Warfare is Changing the World.

The Strait of Hormuz isn’t the only chokepoint worth watching.

Compute capacity — especially advanced chips — remains scarce. Nvidia has reportedly reserved the majority of TSMC’s leading-edge packaging capacity, a critical bottleneck for AI hardware, signalling expectations that demand will exceed supply for years. Now, unlike the Straits of Hormuz, there is no alternative supply.

In what has been, I think, an overlooked point,  is that roughly two-thirds of Hormuz flows have already been bypassed via pipelines and workarounds that now look semi-permanent. Markets adapt when properly incentivised — Europe’s early-2022 energy response is a case in point, where a ~40% shock to gas supply triggered a rapid liquefied natural gas infrastructure build-out and a surge in U.S. imports.

The implication: Unlike TSMC — a genuine chokepoint — Iran’s leverage decays with time. Hold out too long, and its leverage may fade entirely.

Maybe this truce will hold?

What Could Go Right?

With left-tail risks well-articulated, the underpriced right tail is a long-term Middle Eastern resolution: A non-nuclear Iran, no Hezbollah escalation, sanctions lifted gradually and reintegration into the global economy.

Put simply, four to six weeks of volatility could buy 50 years of stability across oil markets, supply chains and regional geopolitics. A half-century-old trade and infrastructure model is being redrawn in weeks, rerouting commerce away from single-point failures and de-risking global supply chains.

Badr Jafar, the UAE’s special envoy for business and philanthropy, captured this well in his opinion piece in the Financial Times:

“The world is watching what is being destroyed. It should pay equal attention to what is being built… The crisis is creating the conditions for genuine intraregional economic integration. States whose ties were strained only weeks ago are now finding common cause.”

This is the grand prize. And while it feels implausible today, Israel and Oman flirting with all-time highs — Oman is up ~40% YTD — offers a glimmer of hope.

While hope is rarely a strategy, I still assign it a higher probability than the left tail: Armageddon.

Why? Because I don’t believe President Trump ever contemplated that outcome.

In his 1962 book Thinking About the Unthinkable, futurist Herman Kahn argued that “appearing a little crazy” could be an effective deterrent.

I’m hoping you’d find that book on the president’s bed-side table. Just next to Art of the Deal.

Have a great weekend.

Gary