Capital Market Laboratories

$GOOGL That earnings report, while quite bad for GOOGL, was an AI HYPER bullish outcome for the AI trade. In fact, other than for GOOGL shareholders, I don't think it could have better news for AI.

Recent power infrastructure IPO: Spoke w former CTO of $GEV then spoke with an exec from one of the largest electrical distributors in North America and did find a new Top Pick Trial for retail CML Pro is just $10: pro.cmlviz.com/exclusive-pos…

$DOCN @digitalocean The chart below is revenue estimates through time for DOCN. I think analysts are low for 2028 given the recent cash raise, large customer wins, rising NDR, and hyperscaler wins. I'm closer to $1.85B than the current mean of $1.78B with back to back 30%+ growth years for '26 and '27.

$JMIA Just off with CEO and CFO; there is some small confusion w covering analysts (me too)/investors. The quarter was considerably more bullish: This is a $20 stock. • All-time high customers • ... with all time high repurchase rates • ... as margins structurally increase • ... with all-time low cash burn Full note out to CML Pro / Bloomberg/Facset/Refinitiv soon.

The Great Question of Our Time; And Perhaps Our Lives Wall Street’s mood right now is basically whiplash. On one side, there’s a fast-forming “software apocalypse” narrative, catalyzed by Anthropic’s latest agentic/coding step-function that’s pressuring SaaS multiples and triggering “get me out” style selling across software. On the other side, there’s a parallel disbelief about the scale of hyperscaler AI CapEx. The market is increasingly treating this as an ROI gamble that either (a) turns into capital destruction or (b) forces a valuation regime-change where these companies trade more like infrastructure operators than “tech,” because the spend profile starts to look like an arms race with uncertain payback. Now hold that pessimism in your left hand, and the actual quarter in your right hand... because it's never that easy. Amazon just printed a Q4 that makes the “AI build-out is slowing” narrative hard to defend on fundamentals: • AWS grew 24% YoY to $35.6B, Amazon’s fastest AWS growth in 13 quarters, and it was the largest AWS revenue quarter ever. • AWS operating income was $12.5B (also a record quarter), with AWS operating margin at 35.0%. • Company-wide Q4 net sales were $213.4B (+14% YoY) and operating income was $25.0B even after special charges; without those charges, operating income would have been $27.4B. And then comes the part that is smashing sentiment: capital intensity. Amazon explicitly said it expects to invest about $200B of CapEx in 2026. That sits on top of a 2025 reality where trailing-twelve-month free cash flow fell to $11.2B, driven primarily by a $50.7B YoY increase in property and equipment spend tied to AI investment. So the “CapEx is eating the P&L and FCF” crowd is not hallucinating. It’s visible in black and white. Now zoom out, because this is where the framing matters. With the four hyperscalers now having reported, the next 12 months look like an AI infrastructure regime, not a cycle: • Alphabet up 100%+ to ~$180B • Meta up ~80% to ~$125B • Amazon up 50%+ to ~$200B • Microsoft up ~35% to ~$135B (h/t @munster_gene ) That’s over $600B of CapEx from four companies in a single year. The majority of that money is flowing into physical reality: data-center land, power, cooling, racks, networking, semiconductors, and custom silicon. That creates the great question: Is this the beginning of the largest infrastructure supercycle in modern technology, or the most expensive misallocation of capital since telecom? Here’s the answer: It can be both, depending on where you sit in the stack. • If the spend is wasteful, the hyperscalers can absolutely get punished. The market is already signaling that fear. “Prove the ROI” is becoming the gating question, and the penalty for not proving it is multiple compression. • But even in the “wasteful” scenario, the waste doesn’t disappear. It still gets spent into concrete, copper (Top Pick), switchgear, transformers, cooling loops (Top Pick), GPUs (Top Pick AMD) and custom silicon (Top Picks). In other words: a software apocalypse, if it comes, still sits upstream of an infrastructure boom. And the bull case is just as coherent. This quarter itself is evidence that demand is not theoretical. AWS reaccelerated to 24% growth and delivered its largest revenue and operating profit quarter ever. Amazon’s custom-silicon strategy is already operating at massive scale and is being extended across multiple future generations. So the real question is not “is AI real.” The real question is: who captures the economics when AI turns into a trillion-dollar physical build-out, and can hyperscalers ultimately earn software-like returns on infrastructure-like capital? That’s the knife edge the market is standing on right now.

Market too bearish on ai infra and security. Demand (for more) is surging. Spend surging. Usage surging. Use cases surging. Hyperscaler monetization rising. If ai eats software that's a software problem (also over done). All needs security. Those stocks will be up on software apocalypse not down.

Nvidia is not the company to watch for an AI "bubble." It is and has been OpenAI. If you're a market bull on AI *anything,* you're a speculator on OpenAI.

yesterday our model predicted a 5 (ie strongest bullish signal). See the video below this post. Join the webinar Saturday, or at least register so you get the video and access: us02web.zoom.us/webinar/regi… See the strategies that triggeed at the bottom of the screenshot.

Tomorrow (Saturday) at 11a PT/ 2p ET: Quarterly CML Pro webinar. We identify three of the ten pillars of genAI that have risen to a new tier of opportunity (higher stock prices), while other pillars have shown the downside risk. Link below: 1/2

$VRT $AMZN incorrect take by market, imho. AWS’s solution lives at the server/rack level (chip-to-rack cooling); Vertiv’s core business operates at the facility level (chillers, room-air handlers, building heat rejection). The two systems are complementary, not identical. ofc if the AWS thing is like amazingly good it culd reduce demand for VRT indirectly, but this is not a replacement product. Vertiv also supplies the facility-scale chillers, air handlers, UPS, and services that AWS still relies on, it’s not a one-to-one replacement product.

This is good news today: » Initial Jobless Claims 227K vs 235K consensus and 232K prior. » Continuing Jobless Claims 1965K vs 1980K consensus and 1955K prior.

Trump says China will supply rare earths in ‘done’ deal - @CNBC

Trump Gives U.S. Negotiators Room to Lift Export Controls on China - @wsj

$GOOGL Waymo, Toyota strike partnership to bring self-driving tech to personal vehicles Alphabet-owned Waymo and Toyota on Tuesday announced a preliminary partnership to explore bringing robotaxi tech to personally-owned vehicles. via @cnbc