VideoBravos ResearchBravos Research

The Final Phase Has Now Begun.

The video argues hyperscaler AI capex—projected at roughly $700 billion this year and up to ~$800 billion by 2028—is reshaping markets and forcing trade-offs between investment and cost cutting, even as some firms report ~20% earnings growth. It warns of a telecom-era analogue where massive buildouts led to severe overcapacity, but contrasts that with a PMI rebound that historically signals continued earnings strength. The investment implication is a near-term bias toward semiconductors, followed by a rotation into power and base metals, while the presenters highlight a set of undervalued, high‑conviction stock ideas they expect to outperform soon.

Artificial IntelligenceSemiconductorsBase MetalsCorporate EarningsEconomic GrowthInvestment Strategy

Preview · 3 more points in the app

Read the Full Summary

Key Takeaways

Hyperscaler AI capex — roughly $700 billion this year and guided toward ~$800 billion by 2028 — is a dominant driver of recent equity momentum and is forcing trade‑offs (e.g., layoffs) even as many firms report ~20% annual earnings growth; that scale of spending should disproportionately lift semiconductor earnings first.

The late‑1990s telecom buildout (~$500 billion spent on fiber) is offered as a cautionary analogue: if AI infrastructure is built far ahead of sustainable demand, the market risks large excess capacity and a severe drawdown.

A rebound in US manufacturing PMI back above 50 makes the current environment look more like 1998 than 2000, implying S&P 500 earnings are unlikely to roll over for at least another year and supporting the near‑term continuation of the capex‑driven rally.

AI-SPENDING

Hyperscaler Capex — The five hyperscalers are expected to spend roughly $700 billion this year on AI infrastructure—more than the GDP of Sweden or Singapore—and management teams are guiding capex higher, with projections up to ~$800 billion by 2028. That scale is powering recent equity market momentum but forcing trade-offs: firms are combining heavy capex with cost actions such as layoffs even as some report roughly 20% annual earnings growth.

“$700 billion. That's how much companies like Microsoft, Amazon, Meta, Alphabet, and Oracle are expected to spend this year alone on AI infrastructure…”—

BUSINESS-CYCLE

Telecom Analogue — Between 1996 and 2001 telecom firms spent roughly $500 billion building fiber while the stock market nearly tripled, but demand never caught up and the industry ended up with on the order of 100x excess capacity, a dynamic that preceded a severe collapse. The video presents that historical buildout as a cautionary analogue for today's AI‑infrastructure surge.

“Between 1996 and 2001, companies like World.com, Global Crossing, and the entire telecom industry spent roughly $500 billion laying fiber optic cable across the United States to power the internet.”—

SECTOR-ROTATION

Sector Rotation — Semiconductors are presented as the primary way to play rising AI infrastructure demand, with the sector showing record earnings growth. Over time capital should rotate into the energy and raw-materials industries needed to build infrastructure—notably nuclear power and base metals like copper and aluminum—which are already starting to benefit.

“Now, the best way to play this is by far semiconductor stocks that are seeing absolutely record earnings growth as a result.”—

STOCK-PICKS

Undervalued Growers — Presenters highlight companies delivering record earnings growth while still trading cheaply, suggesting meaningful upside if markets re-rate them as fundamentals are recognized.

“These companies are already seeing record earnings growth and putting them on an accelerated path upwards…”—

Related Episodes

Important Disclaimers

Not Investment Advice: For education only — not financial, tax or legal advice.

AI-Generated: Made by AI from source transcripts and may contain errors.

No Guarantees: Past performance does not predict future results; investing carries risk.

Our Commentary · Request Removal