Future Download

A.I., Crypto & Tech Stocks

Not a jobs crash, but a capex wave: sectors to watch this quarter

Artificial intelligence is moving from pilot projects to physical infrastructure. The headline last week, as CNN reported, is that AI is not destroying U.S. jobs at scale. A new international study found that roughly a quarter of roles will change meaningfully, but most will evolve rather than disappear. For investors, the more urgent story is where the capital is going: into electricity, land near substations, and compute hardware that is already reshaping multiple sectors.

Jobs are stable. The spending surge is not.

The labor picture looks more evolutionary than disruptive right now. The CNN piece summarizes research showing task-level reshuffling rather than mass displacement. That context matters because capex is accelerating even without a jobs shock. Citigroup now estimates AI-related infrastructure spending by the largest platforms could surpass $2.8 trillion through 2029, up from its prior forecast. When you track the announcements, the money shows up in concrete, transformers, and cooling gear far more than in head-count cuts.

Electricity markets are already flashing the impact. In PJM, the largest U.S. grid region, data-center load was the primary reason capacity auction revenues jumped, according to the market monitor’s Oct 2 report. Another analysis published Friday underscored growing sticker shock for PJM ratepayers as utilities procure supply to meet expected data-center demand. That is the capital cycle investors can actually model.

Where the AI wave hits first

Utilities and grid equipment. Rising AI workloads are colliding with old transmission and tight transformer supply. Utilities in data-center corridors are writing explicit AI-driven load growth into planning documents and signaling more frequent rate cases. The regulatory shape of cost recovery will separate winners from merely busy operators. Watch jurisdictions that pair long-term commitments from tenants with mechanisms that protect existing customers from subsidizing hyperscalers. The dynamic is visible on the tape: when a hyperscaler commits to a region, local power names can move quickly.

Data-center real estate. Electricity, not dirt, is the bottleneck. Power-rich parcels with substation access now command premiums, and developers are reserving megawatts years ahead. Last week’s announcements give a clean example: Google’s new $4 billion campus in West Memphis, to be powered by Entergy Arkansas, puts tariff design and grid access at the center of value creation. Local reporting confirms the site scale and timelines, reinforcing how siting choices ripple into municipal finance and workforce programs (Arkansas Advocate).

Semiconductors and networking. On the compute side, the first-derivative winners remain accelerators, memory, packaging, and high-bandwidth networking. A timely lens on competitive dynamics comes from the Financial Times’ deep dive on Nvidia’s race against Chinese challengers, which highlights how software ecosystems and interconnects shape share over simple chip counts. For investors, the key takeaway is not a quarterly unit number but the durability of demand for training and inference infrastructure as architectures evolve.

Local proof points and the rate case pipeline. The PJM capacity outcome is one signal. Another is how utilities frame new-load contracts. Market monitors and state commissions are increasingly explicit about beneficiary-pays structures, minimum bills, and special-rate agreements that tie megawatt delivery to multi-year commitments. The Utility Dive and E&E News pieces are worth bookmarking as templates for what is likely to spread beyond Mid-Atlantic markets.

What to watch, quickly

• Signed power contracts that de-risk capex for utilities tied to new data-center load, plus the structure of any special rates.
  

• Interconnection queue timelines in key metros and how quickly substations can be expanded or duplicated.
  

• Liquid-cooling readiness across servers and switches as rack densities rise.
  

• Corporate disclosures that link AI services revenue to specific infrastructure outlays.
  

Where pressure is building

Back-office outsourcing and clerical work. Routine, text-heavy tasks such as support, claims triage, scheduling, and summarization are being automated piece by piece. Contracts are being rewritten around productivity and quality metrics rather than billable hours. The current labor data do not show a collapse, but buyers are already repricing service levels. That is more of a margin story than a head-count story, and it plays out quarter by quarter.

Media, marketing, and professional services. Generative tools compress drafting and analysis time, which reduces demand for junior labor even when total headcount holds steady. The CNN coverage underlines that the near-term risk is a change in the staffing pyramid, not broad unemployment. Firms that tie AI to client-visible outcomes retain pricing; those that do not will face fee pressure.

Security and operational risk. New AI-driven attack surfaces make security spend more resilient across cycles. Last week’s headlines were a reminder that enterprise software and infrastructure remain targets, with Oracle warning that extortion campaigns are hitting customers via known vulnerabilities. That combination—rapid AI adoption and elevated threat activity—supports steady demand for identity, endpoint, and cloud posture management even if broader IT budgets wobble.

How to use this

Barbell for long-only. On one side, hold structural beneficiaries of the AI capex cycle: quality utilities with credible load growth and constructive regulators, grid-equipment makers with backlog visibility, data-center landlords controlling power-first sites, and compute and networking vendors tied to large AI clusters. On the other, trim exposure to labor-intensive service providers that cannot re-bundle offerings around speed, accuracy, and measurable outcomes.

Income angle. Regulated utilities and select data-center REITs can pair yield with growth if commissions allow rate recovery on AI-linked investments and tenants sign long-dated power deals. Your checklist should include allowed returns, balance-sheet capacity, and the presence of beneficiary-pays provisions in special-rate agreements. The PJM capacity outcome is a useful reference point for how load can affect forward pricing.

Software selectivity. Buy the line items that prove value in a quarter. Security, developer productivity, and infrastructure automation continue to rank high with CIOs even as discretionary IT spend cools. The Citigroup capex forecast suggests the platforms will keep spending; vendors that ride alongside with clear ROI should benefit.

Bottom line. The real story of AI’s economic impact right now is not layoffs. It is load growth and the capital cycle required to power it. The CNN study coverage captures the labor side, while PJM and West Memphis show the infrastructure side in real time (Utility Dive, Barron’s). Treat AI less as a labor event and more as a decade-long build-out. Own the pipes and the power. Be cautious with commoditized services that cannot re-price around automation.

Partner perk. Want a fast way to match with investing tools that fit your goals? Tryhttps://quiz.moneypickle.com/?partnerid=morning-download to find brokers and advisory platforms that align with your preferences. It is free and takes just a few minutes.

Nothing here is financial advice. Do your own research.

👩🏽‍⚖️ Legal Stuff
Nothing in this newsletter is financial advice. Always do your own research and think for yourself.