Jacobus Enslin @Enslinjaco: "𝘼𝙥𝙧𝙞𝙡 𝟮𝟬𝟮𝟲 𝙪𝙥𝙙𝙖𝙩𝙚 𝙋𝙚𝙧𝙛𝙤𝙧𝙢𝙖𝙣𝙘𝙚 • The portfolio..."

United Kingdom

𝘼𝙥𝙧𝙞𝙡 𝟮𝟬𝟮𝟲 𝙪𝙥𝙙𝙖𝙩𝙚 𝙋𝙚𝙧𝙛𝙤𝙧𝙢𝙖𝙣𝙘𝙚 • The portfolio returned 4.62% in April vs 10.42% for the SP500. • YTD the portfolio returned -2.96% vs the SP500 return of 5.31%. • The annualized return since inception of the portfolio is 19.93% vs 12.22% for the SP500. $GOOG (Alphabet) reported impressive earnings last week and has grown into the second largest portfolio holding. The market continues to drive higher led by the AI and semi-conductor trade. I struggle to see how this changes abruptly, at least in the short-term. The main counterpoint from the people critical of the AI buildout is that not enough profit is being generated to justify/fund the large amount of capex being spent on the buildout. They point to the deteriorating FCF of hyperscalers as case in point. This however ignores the fact that this capex gets depreciated over a weighted avg of about 10 years (which incorporates a GPU depreciation schedule of 5 years). When looking at the industry wide revenue required to justify the cumulative capex spent on AI since 2024, I recon the industry needs to get to about $260bn in annual (AI) revenue. This is about double the run-rate depreciation on the AI capex spent since 2024. Revenue at 2x depreciation ensures an acceptable ROI. Currently industry wide AI revenues are around $140bn run-rate, which means revenues needs to double for the capex spent to date to make financial sense. At this stage it looks likely that AI revenues will double by year-end and hence provide an acceptable ROI on the AI capex spent to date. The other factor supporting the AI trade is that the AI labs are raising 100’s of billions which will provide continued financing for the buildout. OpenAI, Anthropic and XAi have raised a combined $172 billion just in the past 4 months. Anthropic is in discussions for raising another round of $40 billion. The IPO’s of OpenAI and Anthropic could raise an additional $200 billion later this year. That is a combined additional ~$400 billion ultimately being used to buy more compute, with the majority flowing to hyperscalers as revenue. It seems on paper that the AI business model might not be as broken as some report. However the problem I have is that valuations, especially across the semiconductor segment, seem very high. Valuation multiples are being assigned to peak earnings for semi’s as if the industry has stopped being cyclical. While I do not see any reason for the AI/semiconductor trade to unwind in the short-term, I do not have conviction that the upside justifies the risk with buying at current valuations. It has been a mistake on my part to have such little exposure to the AI trade. I am looking for downstream beneficiaries of all this AI capex that are trading at cheap valuations. I am currently focussing on opportunities in the electricity generation space. If all this funding translates into more compute we could easily see electricity demand in the US increase by 100GW by 2030. Some estimates are for up to 170GW. Most electricity generation methods require turbines including coal, gas, nuclear, geothermal and wind. With turbine manufacturing capacity fully booked out to 2029, incremental increases in generation from these sources becomes impossible. This provides new solar projects, which do not require turbines, a catalyst to increase its share of new production capacity. US solar+battery manufacturers look cheap and the OBBBA act in the US provides insulation against cheap chinese imports. $CSIQ (Canadian Solar Inc.) looks interesting from this perspective and I am currently working on a full report on them. The Iran war is not any closer to a resolution than a month ago and it looks probable that the war will escalate again as the ceasefire lapses. It does seem like the global inventories of the commodities affected by the Strait's closure is enough to supply the global demand for another maybe 3 months. However should the blockade of the Strait last longer than this the consequences on all economic activity will become extremely severe. If there is no resolution in the next couple of weeks I think it would warrant reducing the equity exposure of the portfolio. At this stage I'm not going to make any major changes. Best regards Jacob

Not investment advice. The author may have financial interests in the mentioned instruments.

2 Mentioned

GOOG

Alphabet

CSIQ

Canadian Solar Inc.

Strictly Necessary Cookies

Performance Cookies

Functional Cookies

Targeting Cookies