Artificial Intelligence Bubble Has Only Just Begun, Says Citrini7, World's Hottest A.I. Stockpicker

Episode Summary

Executive Summary: Investor Citrini7 describes a cross-asset, valuation-driven style that looks for mispriced opportunities across equities, rates, FX, and credit. He argues AI is a real megatrend centered on data-center and semiconductor buildout, favors a barbell of macro caution and tech upside, and discusses views on banks, gold, the dollar, ECB/Fed policy, oil, and Japan.

Main Topics: Cross-asset investing philosophy (Priority: 5/5): Citrini7 says he invests across asset classes, using macro signals from bonds, equities, and international flows to find assets trading below intrinsic value. AI as a real megatrend (Priority: 5/5): He argues AI/ML has reached a tipping point, with demand flowing first to infrastructure such as GPUs, servers, data centers, and supporting components rather than only to headline software names. Bubbles, sentiment, and historical parallels (Priority: 4/5): He frames AI through the lens of bubble research, especially Soros and the dot-com era, distinguishing between early-stage bubbles with real fundamentals and late-stage speculative excess. Valuation and sector winners/losers (Priority: 5/5): He prefers beneficiaries with attractive valuation and tangible businesses, citing Super Micro, Vertiv, and certain Taiwan suppliers, while warning that some software and education/outsourcing models may be disrupted. Macro views on rates, FX, gold, and central banks (Priority: 4/5): He is bullish on the dollar, expects real rates and Treasury yields higher, thinks gold should weaken, and sees the Fed/ECB/BoE/BoJ as constrained by market pricing and policy inertia. Banking, credit, and opportunistic trades (Priority: 4/5): He discusses regional banks as a selective opportunity, preferring preferreds and debt over common equity in some cases, and emphasizes risk/reward rather than a single macro thesis. Portfolio construction via thematic baskets (Priority: 4/5): He repeatedly says he builds baskets to capture themes, similar to creating a personal ETF, citing cosmetics for the 'lipstick effect' and AI beneficiaries for current positioning.

Key Arguments: His process is cross-asset and generalist: he looks for assets priced well below fair value and uses signals from rates, FX, and international capital flows. AI is different from prior hype cycles because there is real, fast-growing demand for compute, especially GPUs and data-center infrastructure. Even if AI is a bubble, early participation in a real bubble can be highly profitable; the key is recognizing when fundamentals and sentiment still support the trade. The most attractive AI exposure is not necessarily the obvious software winners, but the infrastructure stack: servers, cooling, connectivity, and data-center bottlenecks. Super Micro is an example of a cheap, tangible beneficiary: a server company trading at a low multiple before AI enthusiasm re-rated it. Some software-enabled business models may be threatened by AI, especially marketplaces and tools that automate low-value tasks such as Upwork/Fiverr-like work and certain legal/admin services. Long-only managers are effectively short what they do not own; many remain underweight technology because of prior drawdowns and valuation fears. Regional banks are best approached selectively, often through preferreds or debt rather than common stock, because funding and margin pressure remain. Gold is likely to fall if real rates rise; the dollar remains favored because de-dollarization narratives are overblown. Central banks may continue hiking or stay tighter for longer than markets expect, especially if inflation and labor data remain sticky. Japan remains a bullish equity story because policy remains easier than elsewhere, while the yen still has downside if rate differentials persist.

Data Points: Selling business / deal mix: 90% cash, 10% stock - He sold his business near the top of a frothy sector and watched the stock triple his net worth before halving. March 2020 drawdown: No drawdown during March 2020 - He says this helped attract interest from his network in running money. Silicon Valley Bank short: Shorted in August - He says he found SVB while seeking a hedge against a portfolio heavily exposed to no-recession outcomes. NVIDIA year-to-date move: Up about 180% - Used to illustrate the scale of the AI rally and bubble debate. ARK Innovation decline example: Bottom-to-low took about 2 years - He cites May 2020 to March 2022 to show how long bubble-like assets can take to unwind. OpenAI growth claim: 2000% year over year (annualized) - He uses this as evidence of explosive AI adoption and compute demand. Super Micro valuation: 2-3x free cash flow - He bought it when it was contrarian and very cheap relative to fundamentals. Super Micro current move: About 100% up - He says the position had risen roughly this amount after he bought it. AI basket performance: Up 40-50% year-to-date - He says the basket of AI beneficiaries moved strongly, mostly in the past three weeks. AI basket timing: Mostly from the past three weeks - He attributes much of the basket's gain to the recent surge in AI enthusiasm. NVIDIA valuation example: P/E around 200 - He compares this to Treasury-like returns to explain how richly valued the stock is. NVIDIA prior valuation: P/E around 20 before COVID - He references historical re-rating over the cycle. NVIDIA late-2020 valuation: P/E around 100 - He notes the stock already looked expensive before rising further. NVIDIA October valuation: P/E around 37 to 40 - Used to show the progression of multiple expansion. Treasury yield example: 5% - He compares this to a 20x earnings multiple in a valuation analogy. SVB/first-tier bank trade: Comerica straight bonds yielding about 11% YTM - He says he preferred bank debt over equity in parts of the regional banking selloff. First Horizon preferred purchase: About $16 vs $25 par - He bought preferred stock after the deal broke, citing the discount. ECB deposit rate: 3.25% - Mentioned while discussing whether the ECB will keep hiking. ECB market pricing: 3.8% terminal rate - He references market expectations for additional hikes. S&P 500 level: Above 4,200 - He says he is currently bullish on equities at this level. Potential S&P pullback: Around 4,150 - He thinks a modest pullback would be natural but not a thesis-breaker. Oil price: Around $70 per barrel - He describes oil as washed out and takes a short-term long position. Bank of Japan history: Kuroda era and YCC changes in December/September - He refers to prior yield-curve-control adjustments and expectations for slow normalization. Japan inflation: Positive for the first time in decades - Used to explain why the BoJ may move cautiously.

Pivotal Quotes: "If I find a dollar selling for 50 cents, I'm going to pick it up." — Citrini7: He explains his valuation-first philosophy and willingness to own many asset classes. "Being long in the beginning of a bubble is the best financial decision you can make in your entire life." — Citrini7: He argues that early-stage bubbles can be highly profitable if one understands timing and exits. "I don't ignore a trend when it's staring me right in the face." — Citrini7: He summarizes his approach to AI and other thematic trades despite valuation concerns.

Implications: Listeners should expect a pragmatic, valuation-aware but trend-friendly framework: own real beneficiaries early, avoid obvious losers, and let macro/regime shifts guide risk. The biggest opportunities may sit in infrastructure, not just headline AI/software names.