Summary of "3 Stocks You'll Wish You Bought in 2026 (And 2 You'll Regret Owning)"
Finance / Investing Themes (Energy + AI)
The video argues that the current AI “boom” is inseparable from a power/energy boom needed for data centers—specifically the electricity required for AI training and operations.
Grid planning and power demand projections
It cites U.S. grid planning expectations showing sharply rising power demand forecasts:
- 2022: ~24 GW by 2030 forecast (described as ~2.4% growth over 7 years)
- 2024: ~64 GW
- “Last year”: ~166 GW (described as 16% growth in 5 years)
Key risk factor: “nimbyism”
A major risk factor highlighted is “nimbyism”—regulatory and political pushback against building new data centers due to rising power prices and approval friction. Examples mentioned include:
- State-level friction
- A veto scenario in Maine
- Constraints and project walk-aways in Virginia
Key Framework / Methodology Mentioned
“Uniform accounting” (standardized comparison)
The video references “uniform accounting” to standardize comparisons across energy turbine/fuel-cell peers and reduce accounting noise, making margins more comparable.
Repeatable stock-selection idea
A repeatable approach is described:
- Look back across bull markets (including the 1990s/2000s/2010s, plus the last 5 years and 2020s).
- If a stock already doubled, the probability it doubles again is framed as roughly a coin flip.
- With uniform accounting, the odds are described as improving to about ~60%.
“Momentum confirmation tool”
The video also uses a momentum confirmation tool—market recognition of fundamentals—combined with uniform accounting, to identify stocks where the market is pricing growth but still underestimates it.
Macro / Capex Context Supporting the Thesis
The discussion centers on hyperscaler capex escalation for AI/data centers.
Hyperscaler capex escalation
Examples include:
- Meta data center capex guidance: $125B–$145B
- Total hyperscale capex: described as potentially ~$700B+ this year (as a likely overshoot versus earlier budgets)
“Winner-take-all / prisoner’s dilemma” framing
The argument: hyperscalers are in a winner-take-all / prisoner’s dilemma, where stopping capex risks losing the AI race (with AGI framing).
Tickers / Instruments Mentioned
“3 stocks to buy” (energy)
- GE Vernova (GEV)
- Bloom Energy (implied ticker “BE”; not explicitly stated in the text)
- Kodiak Gas Services (KGS) (explicitly “Kodiak gas”)
“2 stocks to avoid”
- NextEra Energy (NE)
- ACOM (explicitly “ACOM”; ticker spoken as “ACM” — likely referring to AECOM per context)
Other Companies / Tickers Referenced (Context or Examples)
Companies mentioned without clear tickers in the excerpt include:
- Meta, Microsoft, Amazon, Alphabet, Oracle
- SoftBank, OpenAI
- Intel (noted as up +130% “or more” in the last month; no ticker)
- Nvidia (referenced via “Nvidia chips”; no ticker)
- Micron
Additional explicit example:
- EQT (used as an example of low gas extraction cost; explicit ticker “EQT”)
Assets / Sectors / Instrument Themes
- Natural gas and LNG
- Wind turbines / generators
- Transformers / grid infrastructure
- Fuel cells
- Compressors / power generators
- Base-load electricity vs. intermittent solar/wind
- Battery backup (described as limited “extension,” not true 24/7 coverage)
Key Numbers & Explicit Performance Claims
GE Vernova (GEV)
Capacity / power sold:
- 2 years ago: 12–15 GW/year
- This year: ~20 GW/year
- Goal by 2030: 25–30 GW/year (framed as >doubling over ~5+ years)
Backlog:
- “Booked out to 2030” if they don’t expand capacity.
Profitability / margin via uniform accounting:
- Peer uniform earnings margin: ~20%
- GE Vernova “on spin-out” earnings margin: ~3%
- Bull case: margin expansion toward peer-like profitability and thus value creation.
Recommendation posture:
- Acknowledges the stock already ran (described as a phenomenal run over the last 2–3 years)
- Still argues for upside from both:
- capacity expansion
- margin re-rating
Bloom Energy
Stock performance (as stated):
- ~1100% since recommendation
- ~1500% in the last year (figures vary slightly across subtitle snippets; core takeaway is extremely strong recent rally).
Capacity numbers:
- Previously about 100 MW annual power capacity (subtitle suggests timing as “two years ago or last year”)
Plan / expansion:
- 2 GW/year on current campus
- By 2030: potentially 5 GW annual capacity (framed as selling out)
“Stargate Jupiter” deal:
- Jupiter described as 2.4 GW to be supported by Bloom fuel cells (agreement signed)
Cost and breakeven:
- Break-even cited as $50–$80 per megawatt-hour (MWh)
Earnings outlook via uniform accounting:
- Market priced around ~$3B “uniform earnings per year by 2030”
- Argument: Bloom could exceed that via recurring catalyst/reagent-related revenue and margin potential
- Comparison example:
- If GE Vernova at 12–15 GW had 20–25% margins, it might generate ~$7–8B profit (used to justify Bloom’s potential)
Kodiak Gas Services (KGS)
Business model:
- Owns compressor fleets for natural gas pipelines (compressors require power)
- Emphasis on “mobile power” / generators that can be brought to data centers/industrial sites
Expansion numbers:
- Caterpillar backlog mentioned for more generators (subtitle suggests several hundred megawatts)
Risk framing / why it may “stick around”:
- Not positioned as “top of the table” like the first two
- But it may persist due to:
- Double demand from increased natural gas throughput
- Plus “power on demand” sold to data centers at potentially better prices
U.S. natural gas competitive advantage:
- EQT extraction cost example cited as ~$1 per BTU
- Natural gas going rate cited as ~$5 (used to support the resilience/advantage thesis)
Explicit Recommendations / Cautions
Buy recommendations (thesis summary)
-
GE Vernova (GEV):
- Positioned to benefit from AI-driven power buildout
- Capacity scaling: 12–15 GW → 20 GW → 25–30 GW by 2030
- Backlog to 2030
- Margin expansion potential using uniform accounting
-
Bloom Energy:
- Fuel cells as a “behind-the-meter / BYOP (bring your own power)” solution to avoid grid and regulatory constraints
- Stargate Jupiter: 2.4 GW agreement
- Claimed profitability/margin upside with recurring catalyst-driven revenue
- Break-even: $50–$80/MWh
-
Kodiak Gas (KGS):
- Compressors require mobile generation
- Can sell power capacity directly to data centers (potentially better pricing)
- Also gains from increased natural gas pipeline/compression needs (“double dip”)
Avoid recommendations
-
NextEra Energy (NE):
- Data centers need base-load 24/7 power
- Wind/solar are described as intermittent
- Battery backup described as mainly extension, not true coverage
- Conclusion: limited relevance of solar/wind benefits; implied preference for firms with natural gas and nuclear exposure
-
ACOM / ACM (likely AECOM per context):
- Not positioned where spending is concentrated (not exposed to power/telecom/data center construction)
- Rising-tide idea argued not to apply because demand may not flow to its specific end markets
Disclosures / Disclaimers
- The excerpt provided does not include a verbatim “not financial advice” or legal disclaimer.
Presenters / Sources
- Bridget (host; name not provided)
- Rob Spivey, Alimemetry Research
- “Ultimatry/Alimemetry Research” referenced again for a “dark energy” special report (subtitle spelling is inconsistent between “Ultimatry” / “Ultimetry”).
Category
Finance
Share this summary
Is the summary off?
If you think the summary is inaccurate, you can reprocess it with the latest model.