Behind the video
The AI Energy Bubble: Biggest Opportunity or Biggest Trap?
Overall Thesis
The hosts view AI demand and its electricity requirements as durable investment opportunities, favoring Bloom Energy, a basket of power suppliers, and Amazon. They also see potential upside in Reddit's data licensing while warning that short-term market reactions, high valuations, and AI disruption of established businesses complicate stock selection.
Narratives
BEBloom EnergyChris presents Bloom as his preferred investment in the AI power shortage, emphasizing technology-like scaling and improving economics. The hosts argue that its on-site generation can bypass grid constraints, although valuation and execution remain significant risks.
Key Arguments
- On-site fuel cells can supply data centers without drawing electricity from the local grid.
- Lower noise and fewer municipal objections could help deployments proceed.
- Technology improvements could narrow Bloom's cost premium over conventional power.
- Successful hyperscaler deployments could validate the technology and unlock additional customers.
- Chris rejects recurring short-report claims about scandium constraints.
Risks acknowledged
- The valuation is high and cash flow is negative.
- Bloom still relies on natural gas.
- Its electricity currently costs more than some alternatives.
- Major deployment failures or slower scaling would cause Chris to reconsider his position.
- Data-center power will come from multiple technologies and suppliers.
Predictions (1)
AMZNAmazonChris maintains a bullish long-term Amazon investment view and plans to buy more if an earnings reaction creates a selloff. He distinguishes that conviction from his uncertainty about the immediate market reaction to potentially higher capital spending.
Key Arguments
- AWS should benefit from sustained AI infrastructure demand.
- Agentic shopping concerns appear manageable and likely to develop over many years.
- Amazon can develop its own shopping agents and eventually integrate third-party agents.
- Chris has increased his leveraged long position and retains some earnings options.
Risks acknowledged
- Higher capital expenditure could trigger an immediate selloff.
- Agentic shopping threatens advertising economics and raises fraud, chargeback, and customer-control issues.
- Higher diesel and memory costs could pressure retail and device economics.
- Genuinely adverse company information would lead Chris to reduce exposure.
- Chris sells some options during the discussion because near-term market reactions are difficult to predict.
GOOGLAlphabetChris argues that AI search adoption threatens Google's traditional search monetization and explains why he prefers other AI infrastructure investments. He does not affirm an absolute stock-decline forecast when directly asked, acknowledging that cloud and TPU growth could offset the search headwind.
Key Arguments
- AI answers can satisfy users without the advertising exposure associated with traditional search.
- Chris believes advertisers are being squeezed as Google seeks to preserve search revenue.
- Other investments offer AI infrastructure exposure without the same legacy search risk.
Risks acknowledged
- Cloud computing and TPU growth could counterbalance lost search economics.
- Improved AI-mode monetization would change Chris's opinion.
- His previous short position was exited because timing the thesis was difficult.
- The hosts cite strong Google Cloud growth.
Jordan favors Constellation as part of a diversified investment approach to rising AI electricity demand. Long-term hyperscaler power agreements and nuclear capacity support the thesis, while the discussion emphasizes gradual deployment rather than a quantified stock-return forecast.
Key Arguments
- Hyperscalers are seeking long-term electricity supply commitments.
- Nuclear generation offers a cleaner source of data-center power.
- Jordan owns Constellation alongside Bloom and Vistra.
Risks acknowledged
- Nuclear projects and restarts take time.
- The power shortage will be addressed by a mix of technologies.
- The hosts provide no specific share-price target.
VSTVistraJordan sees investment value in Vistra's combination of conventional generation and nuclear exposure. He frames it as a comparatively measured energy investment supported by anticipated earnings growth and long-term power demand.
Key Arguments
- Traditional generation provides an existing business alongside nuclear diversification.
- Long-term power purchase agreements support demand visibility.
- Jordan cites expected earnings growth of roughly 20–30% for 2027.
Risks acknowledged
- The cited earnings growth is less explosive than Bloom's theoretical growth potential.
- Nuclear expansion takes time.
- Operational earnings estimates do not establish a specific stock-return forecast.
The hosts cite GE Vernova's orders and backlog as evidence that the AI power buildout is real and extends over many years. Its negative stock reaction despite strong reported business metrics illustrates their concern about unpredictable market responses.
Key Arguments
- The introduction cites $29 billion in orders and a $197 billion backlog.
- Turbines are part of the broader mix needed to supply data-center electricity.
- The discussion describes capacity commitments extending to 2032.
Risks acknowledged
- The stock declined despite raised guidance and strong orders.
- Jordan describes trading in and out rather than expressing a new directional recommendation.
TTWOTake-Two InteractiveChris has reduced his Take-Two position substantially because AI-assisted game creation could undermine the long-term economics and valuation of large publishers. He still owns a smaller position, while the hosts distinguish an expected GTA VI hit from sales that materially exceed expectations.
Key Arguments
- AI could lower development barriers and increase competition from inexpensive games.
- Questions about publishers' long-term relevance could pressure valuation multiples.
- Chris says he sold approximately 75% of his position after the narrative changed.
- GTA VI could still generate substantial monetization if it exceeds market expectations.
Risks acknowledged
- Chris retains roughly 20–25% of his original position.
- The hosts expect GTA VI to be successful.
- The timing and capabilities of AI game development remain uncertain.
- The stock-upside discussion about an extraordinary runaway hit is hypothetical.
TSLATeslaDave enthusiastically describes a Tesla FSD demonstration and says it led him to purchase two Model Ys. Chris remains interested in Optimus information and considers a Tesla for road trips, but neither makes a directional Tesla investment prediction.
Key Arguments
- Dave says the FSD demonstration exceeded his expectations and reduced driving stress.
- His positive product experience led to two vehicle purchases.
- Chris sees potential practical value in a Tesla for frequent highway trips.
Risks acknowledged
- Dave reports delivery delays.
- The hosts express uncertainty about driver-monitoring requirements.
- Chris says he is still waiting for substantive Optimus information.
- Product enthusiasm does not establish a stock-return claim.
Chris sees Reddit as an attractive investment because its data could command substantially better AI licensing economics. The hosts acknowledge that AI answers are reducing referral traffic and that uncertain deal timing makes the position difficult to trade.
Key Arguments
- Reddit contains valuable, current human discussions used by AI systems.
- Chris considers existing licensing agreements substantially underpriced.
- Usage-linked licensing could let Reddit benefit as AI queries increase.
- An announced deal could cause investors to anticipate similarly improved agreements with other model providers.
- Chris also sees strategic acquisition value.
Risks acknowledged
- AI-generated answers reduce the need for users to visit Reddit and see advertisements.
- The timing of a licensing announcement or leak is unknown.
- The hosts give conflicting expiration dates, citing both 2027 and February/May 2026.
- Reddit contains misinformation, and upvotes do not reliably establish accuracy.
- The near-term earnings outlook is viewed cautiously.
Predictions (1)
CRMSalesforceChris uses Salesforce to illustrate how AI may weaken legacy software growth and valuation economics without eliminating the business. Established customers may remain, while new companies could choose inexpensive AI-built alternatives.
Key Arguments
- Large incumbents benefit from Salesforce's established records and years of fixes.
- Startups may adopt lower-cost competitors built with AI.
- Long-term growth uncertainty helps explain pressure on software valuation multiples.
Risks acknowledged
- Chris explicitly says Salesforce is not going away.
- Existing large customers are likely to continue using established CRM systems.
- The discussion does not establish a separate absolute stock-price forecast.
Hedges & Caveats
- AI investment will produce winners, losers, and substantial volatility; real technological value does not guarantee every investment succeeds.
- Amazon could sell off after earnings if investors dislike higher capital spending, so Chris reduces options exposure during the episode.
- Bloom's valuation, negative cash flow, natural-gas dependence, and ability to scale deployments are acknowledged risks.
- Reddit's licensing catalyst has uncertain timing, and the hosts give inconsistent contract expiration dates.
- Alphabet's AI infrastructure growth or improved AI search monetization could offset search-business pressure.
- Chris retains a smaller Take-Two position despite concerns about AI-driven competition.