Behind the video
The Start of the AI Boom (Or Depression) [How Cooked are We?!?!]
Overall Thesis
The host believes the AI infrastructure boom is still in its early stages and that falling token costs will benefit hardware suppliers and software platforms. He sees current hardware valuations and selling pressure as buying opportunities and recommends diversified exposure across selected AI stocks.
Narratives
AMDAdvanced Micro DevicesThe host praises AMD's earnings, margin expansion and accelerating data-center growth. He recommends buying below $520 and expects the shares to remain above that level over the long term, while acknowledging a richer valuation than Nvidia.
Key Arguments
- Guidance implies accelerating sequential revenue growth.
- Gross margins increased and data-center revenue approximately doubled.
- AMD participates in both enterprise AI and frontier infrastructure.
Risks acknowledged
- Its PEG valuation is roughly twice Nvidia's.
- Free cash flow is limited and debt has increased.
- Client and gaming revenue declined.
NVDANvidiaThe host calls Nvidia a bargain and recommends exposure within a diversified AI portfolio. Rising GPU rental prices and expanding infrastructure commitments underpin his positive investment view.
Key Arguments
- Approximately 0.67 PEG valuation
- Strong cash flow
- Increasing GPU rental prices and infrastructure demand
Risks acknowledged
- Circular financing and large funding commitments create systemic risks.
MUMicron TechnologyThe host describes Micron as a bargain following volatility in memory stocks. He believes forced selling has created attractive hardware valuations despite strong AI spending.
Key Arguments
- Approximately 0.33 PEG valuation
- Leverage liquidations have pressured hardware stocks
- AI demand supports memory spending
Risks acknowledged
- He expects additional memory supply to come online.
Predictions (1)
The host considers SK Hynix dirt cheap after volatility in memory stocks. He also highlights storage demand as a prospective next phase of the hardware cycle.
Key Arguments
- Strong undervaluation assessment
- AI-related memory and storage demand
Risks acknowledged
- Leverage liquidations and additional memory capacity can affect the sector.
Predictions (1)
MSFTMicrosoftThe host recommends Microsoft within a diversified AI portfolio. He also questions whether lease accounting makes reported free cash flow look stronger than the underlying investment burden.
Key Arguments
- Existing cash-generating businesses fund AI investment
- Exposure across AI platforms and infrastructure
Risks acknowledged
- Lease spending can obscure the effective capital burden
- Borrowing costs have increased
Predictions (1)
AMZNAmazonThe host endorses Amazon as part of a diversified AI investment basket. He highlights its rapidly expanding purchase commitments as evidence that the infrastructure boom remains early.
Key Arguments
- AI purchase orders reportedly doubled since August
- Broad participation in AI infrastructure
Risks acknowledged
- Large spending commitments require funding and pressure cash flow.
Predictions (1)
TSLATeslaThe host includes Tesla in a diversified AI investment basket he expects to perform well over time. He provides no standalone Tesla valuation or price target.
Key Arguments
- Endorsed participation in a diversified AI portfolio
Risks acknowledged
- Concentrated single-stock exposure can produce severe volatility.
Predictions (1)
The host expects a very strong backlog and growth report as demand for chipmaking equipment expands. This is an operating forecast without an explicit share-price or investment-action prediction.
Key Arguments
- More companies want chip manufacturing machines
- He anticipates a strong January operating report
The host cites Super Micro's earnings and increased rack-installation forecast as evidence of strong hardware demand. He does not make a separate stock-direction prediction.
Key Arguments
- Reported earnings beat and raised rack forecast
- Hardware demand supports the broader AI buildout thesis
METAMeta PlatformsThe host discusses Meta's local-model strategy and customer acquisition through generous initial access. He expects platform providers to monetize users after attracting them, without giving a Meta stock forecast.
Key Arguments
- Local model deployment supports hardware demand
- Free access can lead to paid subscriptions
Risks acknowledged
- Frontier-model development requires substantial spending
The host briefly describes a referenced acquisition involving Lumentum as a great deal. He provides no associated prediction about where Lumentum shares will trade.
Key Arguments
- Positive assessment of the referenced transaction
NBISNebius GroupThe host says increasing GPU rental prices are helpful for underwriting Nebius debt. He distinguishes its lower debt burden from CoreWeave without making a share-price prediction.
Key Arguments
- Rising GPU rental rates
- Less debt than CoreWeave
Risks acknowledged
- The broader buildout still requires substantial financing
The host uses Cisco's historical post-collapse valuation as a comparison for current technology valuations. He makes no prediction about Cisco's future returns.
Key Arguments
- Historical post-bubble valuation comparison
AAPLAppleThe host views Apple as an AI hardware beneficiary and recommends portfolio exposure. He expects operating growth above Wall Street estimates, but acknowledges that the shares are relatively expensive.
Key Arguments
- Product and services margins expanded despite higher memory costs.
- Local model deployment could support Mac demand.
- Strong balance sheet and cash generation support the investment thesis.
Risks acknowledged
- Apple is not among the cheapest AI opportunities.
Predictions (1)
AVGOBroadcomThe host identifies Broadcom as a valuation bargain and recommends owning it within the AI stack. Custom chip demand supports his thesis.
Key Arguments
- Approximately 0.63 PEG valuation
- Participation in custom AI chip design
Risks acknowledged
- AI financing and spending sustainability remain broader risks.
Predictions (1)
SNDKSanDiskThe host calls SanDisk dirt cheap after selling pressure in memory-related hardware. He considers NAND a potentially important next area of demand.
Key Arguments
- Current valuation described as very cheap
- NAND capacity has not expanded like memory capacity
Risks acknowledged
- Hardware shares have experienced leverage-driven volatility.
Predictions (1)
MRVLMarvell TechnologyThe host recommends Marvell as part of diversified exposure to the AI hardware stack. He highlights custom chip designers as beneficiaries of expanding local and enterprise AI usage.
Key Arguments
- Custom AI chip design exposure
- Diversification across the AI supply chain
Risks acknowledged
- Single-stock concentration can produce severe volatility.
Predictions (1)
GOOGLAlphabetThe host explicitly excludes Google from his recommended AI portfolio. He does not give a detailed company-specific explanation for this avoidance stance.
Key Arguments
- Explicit current refusal to invest in Google
Predictions (1)
SPCXSpaceXThe host endorses SpaceX exposure within his AI portfolio. He also anticipates enormous borrowing needs to fund its infrastructure ambitions.
Key Arguments
- AI infrastructure and space-chip spending exposure
- Included in an endorsed long-term investment basket
Risks acknowledged
- Potential borrowing needs of approximately one trillion dollars
Predictions (1)
PLTRPalantir TechnologiesThe host praises Palantir's profitability, cash flow and customer expansion but expects stronger platform competition. He criticizes the earnings-call questioning and notes that its cash-flow yield is modest, without making an explicit stock-direction call.
Key Arguments
- Existing customers reportedly increased spending by about 95%
- Forward-deployed engineers help integrate the platform
- Profitable software-platform model
Risks acknowledged
- Competition from other AI platforms is coming
- Cash-flow yield is around 1.3%
- Earnings-call questioning lacked scrutiny
INTCIntelThe host attributes Intel's disappointing earnings to weak PCs rather than collapsing AI demand. He also identifies Intel as a potential manufacturer within the expanding AI hardware stack.
Key Arguments
- PC weakness also affected AMD
- AI demand can drive memory costs higher and pressure consumer PCs
- Potential chip manufacturing role
Risks acknowledged
- Disappointing earnings and PC demand
The host identifies Taiwan Semiconductor as a manufacturer supporting the AI hardware stack. He does not state a specific investment action or share-price expectation.
Key Arguments
- Manufacturing role in expanding AI chip demand
The host names Seagate while discussing storage as a possible next phase of AI demand. The remarks express sector and operating optimism without a clear standalone stock-return call.
Key Arguments
- Storage buildout may lag other memory capacity
- AI demand could support storage investment
CRWVCoreWeaveThe host monitors CoreWeave's credit default swaps as an indicator of AI financing stress. He notes recent increases but takes some comfort that they have not reached earlier highs.
Key Arguments
- GPU rental pricing supports debt underwriting
- Credit default swaps remain below prior highs
Risks acknowledged
- High debt and financing sensitivity
- Recent credit-spread increases
ORCLOracleThe host uses Oracle's long-term bond yield to illustrate rising AI funding costs. This is a financing-risk observation rather than an explicit bearish stock call.
Key Arguments
- Referenced 40-year debt yield of 8.43%
- Higher yields make debt-funded expansion more expensive
The host notes that more than 60% of S&P 500 constituents are over 20% below their all-time highs. He discusses market breadth and attributes additional downside estimates to others without adopting a specific S&P 500 return forecast.
Key Arguments
- Substantial constituent drawdowns indicate weak market breadth
- Some outside researchers view the repricing as healthy
Risks acknowledged
- Geopolitical uncertainty could prolong choppiness
- Additional downside estimates are attributed to Deutsche Bank
Hedges & Caveats
- Labor-market deterioration or an AI downturn could destabilize the economy.
- Large financing commitments, weaker free cash flow and rising borrowing costs create funding risks.
- The AI boom could eventually become a bubble.
- Election uncertainty and Iran could prolong volatility.
- Concentrating in one stock or one neocloud can expose investors to severe volatility.
- Apple and AMD are not the cheapest opportunities.
- The host repeatedly says his comments are not personalized financial advice.