Beyond Rates: AI Infrastructure, New Rules, Commodity Power Plays & the Grid Bottleneck

In partnership with

Opening Note

There’s something a bit disorienting in the air: signs of economic slowdown are piling up, but markets are still acting like they believe in goldilocks. As a result, investors—especially those who earn well but aren’t yet comfortably wealthy—must pay attention: soft unemployment, stubborn inflation, political tensions, and changing Fed dynamics are all colliding. The weeks ahead may deliver clarity... or more confusion. Either way, being deliberate now matters.

7 Actionable Ways to Achieve a Comfortable Retirement

Your dream retirement isn’t going to fund itself—that’s what your portfolio is for.

When generating income for a comfortable retirement, there are countless options to weigh. Muni bonds, dividends, REITs, Master Limited Partnerships—each comes with risk and oppor-tunity.

The Definitive Guide to Retirement Income from Fisher investments shows you ways you can position your portfolio to help you maintain or improve your lifestyle in retirement.

It also highlights common mistakes, such as tax mistakes, that can make a substantial differ-ence as you plan your well-deserved future.

Download your free copy today.

The AI Compute Boom Just Hit the Power & Real-Estate Wall

The AI capex super-cycle is no longer just a chip story—it’s a power, land, cooling, and grid-hardware story. Primary North American data-center supply reached a record ~8,155 MW in H1’25 while vacancy fell to an all-time low of 1.6% according to CBRE. Tight markets are pushing leases into secondary metros and driving pre-leasing at scale, often where interconnection queues and equipment lead times are longest, with CBRE noting record-low vacancy.

There’s also a physical bottleneck: transformers. Wood Mackenzie estimates power-transformer supply deficits of ~30% in 2025, echoed by PV Magazine. Lead times have stretched from months to as long as three years for high-voltage units per Utility Dive and even three to five years in some documented cases via Bloomberg. Meanwhile, imports supply ~80% of U.S. power transformers according to Renewable Energy World. Manufacturers are scrambling to add capacity—Brazil’s WEG just earmarked $77M to expand U.S. output as AI demand bites (Reuters).

This is a multi-year picks-and-shovels story across colo REITs, power/cooling vendors, optical networking, and grid hardware—but it’s cyclical and exposed to overbuild once supply catches up.

AI Regulation Graduates From Talk to Law (in California)

On September 29, California enacted SB 53, the first state law requiring frontier-AI developers with $500M+ revenue to publish safety protocols for catastrophic risks, with penalties up to $1M per violation, per Reuters. Governor Newsom’s office framed it as advancing California’s leadership on AI (Governor’s release).

Why it matters to markets:

• Valuation sensitivity: Mandatory governance disclosures can compress multiples for vendors without credible safety programs.

• Procurement moat: Enterprises may prefer compliance-ready partners, advantaging scaled players who can shoulder reporting costs.

• Cost curves: Compliance and model-eval overhead widens the gap between hyperscalers and smaller AI firms.

China’s Commodity Leverage: the BHP Iron Ore Freeze & PMI Cross-Currents

China’s state buyer CMRG instructed mills and traders to halt purchases of BHP’s dollar-denominated iron ore cargoes, escalating a pricing dispute with one of the world’s largest miners, reported Reuters. Mining.com added that Beijing is willing to discipline suppliers in markets it dominates.

At the same time, the private manufacturing PMI rose to 51.2, the fastest since March, according to Reuters, even as the official PMI showed contraction for the sixth month (AP). Divergent PMIs point to selective export-led strength amid domestic weakness.

ESG’s Reset: Flows Fragment While Rules Stall

On September 12, the Eighth Circuit paused litigation over the SEC’s climate-disclosure rule after the agency stopped defending it, according to Reuters. Legal summaries from DLA Piper and JD Supra note corporates remain in limbo.

Meanwhile, Morningstar reported record Q1 outflows (~$8.6B) from global sustainable funds, led by the U.S., as covered by Reuters. Yet global sustainable assets still exceed $3.5T, per Sustainable Views.

Capital Is Still Chasing AI—Across Equities, Bonds, and Convertibles

Global equity funds swung back to net inflows late September on AI optimism and easing hopes, according to Reuters. Weekly data also showed a renewed pickup in U.S. equity flows (Channel NewsAsia).

At the financing level, global convertible issuance hit a five-year high, per Reuters, while Citi now projects >$2.8T in AI infrastructure capex by 2029, with ~55 GW of incremental power needed (Reuters).

Quiet but Important: EM Sovereign-Risk Plumbing Gets an AI Upgrade

The Global Emerging Markets Database (GEMs) run by development banks is deploying AI to improve sovereign-risk analytics, according to Reuters. Better transparency could compress risk premia for some issuers and expose others.

Strategy in Focus: Next-Gen Data Infrastructure

This week’s themes all point in one direction: the world is building the digital backbone of the AI era. Hyperscaler demand has pushed North American data-center vacancy to just 1.6%, a record low even as total supply hit new highs, according to CBRE. That scarcity is colliding with an infrastructure bottleneck: power transformers face a 30% global supply deficit in 2025, with lead times stretching to three years or more for high-voltage units, based on Wood Mackenzie and Utility Dive. Meanwhile, governments are reshaping the playing field—California’s new SB 53 law requires AI firms above $500M in revenue to publish risk disclosures, raising the compliance bar and favoring “infrastructure-ready” partners, as reported by Reuters. And at the capital level, the scale of investment is staggering: Citigroup now projects AI infrastructure spending will exceed $2.8 trillion by 2029, with massive knock-on demand for power, cooling, storage, and networking capacity, according to Reuters.

For investors, this isn’t about chasing the latest AI stock—it’s about systematically owning the “picks and shovels” that make the AI boom possible: the data centers, interconnect hardware, power and cooling systems, and storage technologies that keep the digital economy running. The Next-Gen Data Infrastructure strategy was designed for exactly this environment, offering diversified exposure across the stack and rebalancing every 30 days to avoid concentration risks or overbuild cycles. By leaning into the structural tailwinds of digital infrastructure while letting rules-based signals manage the inevitable volatility, this approach helps investors stay positioned for the long-term buildout without betting everything on a single company or narrative.

Closing Thoughts

This week’s story isn’t a dot plot—it’s the plumbing of the next decade: who has power, who can comply, who controls inputs, and where capital pre-funds tomorrow’s capacity. For HENRYs, the edge comes from systematically owning the infrastructure of change—not predicting every headline.

A rules-based sleeve like Next-Gen Data Infrastructure keeps you exposed to the trend while rebalance discipline manages risks. Pair it with a diversified core, and let real-economy constraints guide sizing.

— Surmount Markets