“Today’s number is 80.” That’s the quip making the rounds—that 80% of Nvidia employees are now millionaires. Whether apocryphal or not, the spirit of the statistic captures the moment: we’re living inside a once-in-a-generation wealth transfer powered by AI infrastructure, where capital is flowing at unprecedented speed to the platforms, vendors, and talent closest to the compute stack.

However, the same tide that is minting millionaires is also exposing hard truths about valuation, capital intensity, and jobs. The latest earnings from the “Magnificent 7” tell a story of extraordinary performance and extraordinary spend. Meanwhile, the on-the-ground reality for workers is shifting: AI isn’t a distant threat; it’s a live variable in boardrooms and budgets. Layer on a fragile geopolitical détente over rare earths, and you have the contours of an economy that’s booming at the top while rewriting the rules for everyone else.

This is your PyUncut field guide to the moment—what just happened in big tech earnings, how AI is changing the work you do, why geopolitics can yank the steering wheel, and the practical playbook to protect your career and portfolio.

---

Quick Summary

• Big Tech’s quarter was excellent: Alphabet, Amazon, Microsoft, Apple, and Meta posted strong revenue, margin, and usage metrics, but the capex supercycle (AI data centers, chips, power) is the new center of gravity.
• Valuation dispersion matters: Amazon and Meta screens as relatively attractive on a growth-to-multiple basis; Apple remains quality but richly valued.
• AI is taking jobs—now: Short-term job destruction is real, even if long-term net creation may follow. Customer service, sales support, marketing analytics, and certain white-collar tasks are most exposed.
• Winners in the AI era: Hands-on technical trades, healthcare and human care, and in-person operational roles remain resilient.
• Career playbook: Become the AI-literate person in the room, get proximate to decision-makers, make your value legible, and embrace reinvention as a feature, not a bug.
• Geopolitics is a risk vector: A rare-earths “truce” is a ceasefire, not a peace treaty. Supply chains and defense-adjacent inputs remain pressure points—and investment opportunities.

---

Part I — Earnings: The Utilities of the Information Age

If you only glanced at the headlines, you saw green. If you read the footnotes, you saw gigawatts.

Alphabet (Google): The AI Cloud Narrative Becomes a P&L

Alphabet beat on the top and bottom line with revenue up ~16% to ~$102B, powered by resilient Search, surging YouTube, and especially a 34% jump in Cloud. The striking subplot: Google Cloud is positioning as the preferred platform for AI workloads. Over 70% of existing cloud customers using Google AI products and a backlog above $150B fortify a compelling thesis: as AI moves from prototype to production, the cloud that stitches models, data, and governance together wins durable, high-margin business.

Amazon: Profits, Power, and… Its Own AI Chips

Amazon also delivered a beat: revenue +13%, AWS +20%, ads +24%, and a share price pop to match. The sleeper headline is Trainium 2, Amazon’s in-house AI training chip, now a multi-billion-dollar business with triple-digit growth. Pair that with ~4 GW of new power capacity and the capex arc becomes clear: AWS is not just renting servers; it’s building an AI-first utility—compute, silicon, and energy—under one roof.

On valuation, Amazon trades near 31× earnings versus a 5-year average around 59×. If you believe in the operating leverage of AI-enhanced retail + high-margin AWS + structurally growing ads, the risk-adjusted entry point remains compelling.

Apple: Toyota’s Volumes, Ferrari’s Margins

Apple’s quarter beat, gross margin expanded to ~47%, and guidance for double-digit iPhone growth next quarter lit a spark. Still, iPhone +6% underwhelmed relative to the most exuberant expectations, and the stock’s ~36× multiple prices in a lot of perfection. Strategically, Apple continues to perform an economic magic trick: mass-market scale at luxury-grade margins. For investors, this is the quintessential quality compounder—but today it’s priced like one, too.

Microsoft: Azure Still Climbing, Capex Still Climbing Faster

Azure revenue +~39% kept the flywheel humming, but capex surprised to the upside (~$35B versus a $30B run-rate expectation). That’s the point: Microsoft is leaning into the AI infrastructure land-grab—models, GPUs, data centers, and the developer stack around Copilot. In the short term, the market flinched at the spend; over the medium term, installed-base leverage and SaaS bundling economics tend to reward category owners.

Meta: Reels Is a TV Network, The Datacenter Is a Power Plant

Meta’s revenue jumped ~26% YoY; time spent on Instagram +30%; Reels pacing toward a TV-scale ad business (tens of billions annually). The wobble? Capex at ~38% of revenue (up from ~20%) spooked investors conditioned by last year’s “year of efficiency.” But the story is consistent: Meta is funding AI ranking systems, recommendation engines, and a global compute grid. If Reels monetization keeps compounding and WhatsApp Business matures, ~29× earnings looks undemanding relative to growth.

The Capital-Expenditure Supercycle

Meta, Microsoft, and Alphabet are allocating a larger share of revenue to capex than the average global utility. Why? Because they are becoming utilities—of the information age. In classic utilities, a massive upfront investment buys decades of regulated returns. In tech, front-loaded AI capex—chips, land, power, cooling—buys moats: better models, faster inference, lower unit costs, and developer gravity. Yes, heavy capex can erode returns if growth stalls. But when scale begets quality, and quality begets network effects, spend today is monetization tomorrow.

Investor takeaway: This cycle rewards companies converting dollars into durable advantages (model performance, inference latency, developer lock-in). Evaluate capex not as a cost, but as future gross margin.

---

Part II — Valuation: Price Still Matters

Retail investors often stop at the narrative (“I like Company X”). Professionals add: “…at what price?”

• Amazon: structural growth across three engines (retail efficiency + AWS + ads), trading at ~31× versus its ~59× 5-year average.
• Meta: engagement up, monetization runway intact, and a sub-30× multiple.
• Alphabet: secular growth with Cloud/AI optionality; valuation sits in a “reasonable quality” zone.
• Apple: superb business, premium multiple; returns may track earnings growth, not re-rating.
• Microsoft: the safest AI platform bet; priced accordingly.

Positioning lens: If you want asymmetric upside, you typically want great businesses at reasonable prices, not perfect businesses at perfect prices. Today, Amazon and Meta tilt that way.

---

Part III — Jobs: AI Isn’t Coming for Work; It’s Already Here

For a year, the industry line went: AI won’t take your job—someone using AI will. The truth is less comforting. In the short run, AI is taking jobs—and not by magic, but by math. If a CEO can grow revenue without growing headcount, earnings expand. Boards notice. Budgets adjust.

Where the Cuts Are Landing

Functions that combine repeatable digital tasks with communication are most exposed:

• Sales & customer support (scripts, routing, triage, follow-ups)
• Marketing ops & research (audience segmentation, testing, briefs)
• Basic analytics and reporting
• Back-office admin and documentation

The near-term wave is white-collar and screen-bound. The next wave encroaches on physical operations as robotics and vision models mature (think warehouse automation and last-mile logistics). Major employers are already targeting automation at scale—the path from pilot to platform is shortening.

Where Work Remains Resilient

Jobs that demand hands, presence, and care remain far more defensible:

• Skilled trades (electricians, plumbers, industrial technicians)
• Healthcare and human-care roles (nursing assistants, phlebotomists, therapists)
• On-site operators (water treatment, industrial truck and tractor operators)
• Field construction and maintenance

Why? Because these roles blend physical dexterity, situational judgment, and human trust—hard for models to replace at scale and cost.

---

Part IV — Your Career Playbook: How to Be Layoff-Resistant

You can’t control the macro. You can control your legibility, leverage, and learning velocity. Here’s a pragmatic plan:

1) Become “the AI person” (in substance and brand)

• Substance: Master the tools your team uses—text generation, analytics copilots, prompt engineering, automations. Design a 10–20 slide deck that shows how AI cuts cycle time in your workflows (examples, metrics, a pilot).
• Brand: Share internal memos, lunch-and-learns, and periodic updates (“What we shipped,” “What we learned,” “What’s next”). Make your competence visible.

2) Get proximate to decision-makers

Reduction-in-force (RIF) lists are not fully scientific. Proximity matters. Book a monthly 20-minute with your VP/GM. Come with insight + initiative: a metric you improved, a risk you removed, a customer problem you solved.

3) Make your value countable

• Keep a living scorecard: cycle time reduced, dollars saved, pipeline added, defects removed.
• Package your wins in a narrative memo every quarter. Quiet output is invisible output.

4) Have an opinion (and own it)

AI is great at analysis, weak at conviction. Synthesize options, then pick a path. Organizations retain people who decide—and accept accountability.

5) Embrace reinvention as a feature, not a flaw

The elite reinvent all the time—a designer from a skater, a media mogul from a broker, a founder from a teacher, a governor from an actor. In a regime shift, identity agility is an asset. If the market says your old role is oversupplied, rebrand yourself to where demand is peaking.

6) Manage the psychology of setbacks

Layoffs are often indiscriminate. Don’t let a corporate event rewrite your self-story. Mourn briefly; move immediately. Network daily, ship small wins, get physically strong. Momentum in one domain spills into others.

---

Part V — Geopolitics: The Rare-Earths “Truce” Isn’t a Peace Treaty

Last week’s Trump–Xi meeting produced a trade truce: a one-year pause on certain rare-earth export controls from China and some tariff relief from the U.S. If that sounds tenuous, it is. In strategy, you must inventory your leverage and your counterpart’s. China controls a majority of rare-earth mining and most of the processing capacity—inputs that touch missiles, EVs, wind turbines, and chips. A calendar-limited relaxation is not a durable supply of security.

What could go right?

• The U.S. and allies’ onshore/refiner’s mining and processing capacity.
• Defense-adjacent funds invest in alternative sources and substitution technologies (magnet innovations, recycling).
• Corporates multi-source, builds buffer stocks, and designs around single-point failures.

Investor angles to watch (do your own diligence):

• Materials & processing companies outside China with credible scale-up plans.
• Power & cooling infrastructure suppliers for data centers (HVAC, switchgear, transformers).
• Silicon alternatives (inference-optimized ASICs) and AI server integrators.
• Logistics automation and machine-vision providers as the second wave of physical AI adoption spreads.

Bottom line: The AI supercycle is resource-constrained. Compute, power, land, water, and critical minerals are the new Four Horsemen of capacity planning. Policy risk isn’t a headline; it’s a cash-flow variable.

---

Part VI — Portfolio Construction in an AI Utility World

A few practical heuristics for PyUncut readers:

1. Favor platforms with three flywheels
   Compute + Distribution + Data. The more of those a company controls, the more pricing power and durability it has.
2. Separate “AI-adjacent hype” from “AI-essential picks”
   Ask: If AI adoption paused for 12 months, would this company’s revenue collapse or compound? Buy the latter.
3. Underwrite capex as a moat test
   Track return on invested capital (ROIC) over a multi-year window. Early ROIC compression is fine if unit economics improve and share of workload rises.
4. Use valuation as a throttle, not a blindfold
   High quality at a fair price usually beats fair quality at a high price. In this tape, Amazon and Meta skew to the former; Apple and Microsoft remain fantastic but fully loved.
5. Diversify your AI bet across the stack
   A core platform position + a power/cooling/real-estate beneficiary + a silicon/inference name + a workflow SaaS that monetizes AI usage can give you cycle balance.

---

Part VII — The Human Edge in the Age of Models

AI levels the playing field on knowledge retrieval and first drafts. What it does not replicate well—yet—is:

• Taste (what to build, which story to tell, what to omit)
• Conviction (which option to choose)
• Trust (who people prefer to work with and buy from)
• Embodiment (the lived, physical presence required in care, craft, and complex operations)

Build your moat around those. Use AI to eliminate drudgery; spend your human time on judgment, relationships, and craft.

---

Week Ahead: What to Watch

• Earnings cadence from AI-levered but still-debated names (software, semis, marketplaces) will show how broadly the AI tide is lifting all boats.
• Macro prints and power-grid headlines: data-center build-outs are running into electricity constraints—watch utilities’ interconnection queues and transformer backlogs.
• Policy chatter: any movement on export controls, rare-earths, or data-localization can swing input costs and deployment timelines.

---

Final Word: Play Offense

It’s tempting to treat this era like a storm to be waited out. That’s the wrong metaphor. AI isn’t weather; it’s climate. The baseline is shifting under our feet—how companies spend, how leaders hire, how value accrues.

For investors, the mandate is to underwrite capex-backed moats, pay attention to price vs. growth, and diversify across the AI stack and enablers. For operators and builders, the play is to become AI-fluent, visible to power, and relentlessly reinventable. And for everyone, the task is emotional as much as analytical: resist the paralysis of bad news, convert uncertainty into prototypes, and let forward motion do its quiet compounding.

The AI age will not treat all participants equally. But it will reward those who ship.