Today I’m excited to share that Optivolt is unstealthing with a $32M Series A. Solar growth continues to outpace all other forms of energy production, but the granularity of control we get from deployed systems hasn’t changed in decades. So we’ve added a brain inside solar panels, opening the door for a new unit of optimization and control.

Elon Musk: "It is absolutely insane to be deploying, providing the power and all of the electrical equipment, the cooling, the batteries, the networking, and everything else that's needed to turn on 220,000 GB300s a month. But that is actually what we're doing right now." SpaceX is "highly confident" of 13 GW of compute by end of 2027, and 23 GW is possible. It also casts its own turbine blades to skip a 12 to 18 month wait. $SPCX

There's a lot of speculation around an incoming SpaceX and Tesla merger Here's 8 ways both companies BENEFIT IF it happens: 1. Investment-grade debt instead of private credit $TSLA has $43.5B of cash and did $4.7B of operating cash flow last quarter. $SPCX is seeking $40B of debt just for NVIDIA chips and Morgan Stanley estimates it needs ~$84B of outside capital per YEAR through 2034 A combined company funds this with cash flow and investment-grade debt instead of private credit. Saving even 2-3 points of interest on $40B is $1B+ a year, and that compounds across the entire buildout 2. Starlink replaces cellular in every Tesla Today, carriers collect an estimated $10/month per connected car, a large % of the total Tesla charges. Moving this in-house to Starlink not only means 100% of revenues moves in-house, but also a better product with (eventually) no dead zones 3. Starlink for Robotaxi and Humanoid Robots Cybercabs are already getting built with Starlink inside them, a merger would mean this can come with no additional cost/sale required between the companies 4. Compute at cost and better optionality Tesla has its own training compute today, which doubled in the first half of 2026, but FSD, Robotaxi and Optimus are only just getting started and they will need a LOT more compute SpaceX is already building this out at size with 10GW expected by EOY 2027. Rather than leasing this to Tesla at market prices ($30M/MW ish right now), Tesla's extra demand can simply run on SpaceX compute at cost SpaceX also now has more optionally with their compute. To move between leasing, using for Grok or now using for Tesla products 4. Energy at cost Tesla deployed 13.7 GWh of energy storage last quarter and xAI/SpaceX is one of Tesla's biggest customers ($430M of mostly Megapack revenue in 2025). Merged, the data centers get Megapacks at cost with guaranteed supply, and yes, that intercompany revenue disappears in consolidation, but it's a wash since it's saved as an expense on the other side 5. Terafab gets built FASTER Terafab today is a joint venture between two separately owned companies, so every major decision crosses two boards, two capital committees, and negotiated cost-sharing agreements. Merged, one owner makes the call and capital moves the same day. Construction starts in December, and that speed advantage compounds through every phase after the first $16.8B. Elon's whole edge is iteration speed, and a JV structure is the opposite of that 6. Killing the duplicate overhead Two public companies means two of everything: two legal departments, two HR orgs, two finance and audit functions, two IR teams, two procurement groups buying from the same suppliers. Combined SG&A is about $11.2B. Cutting 15-25% of the ~$3.4B corporate slice of that is about $0.5-0.9B a year (est). Not the headline reason to merge, but still useful 7. A potential tax arbitrage SpaceX's AI buildout generated $3.7B of losses in the first half alone Merged, those losses offset Tesla's taxable profit at the 21% corporate rate, worth roughly $0.3-1.0B a year Though looks like SpaceX is soon going to generate a profit, so this may be irrelevant after next earnings 8. Combined Data for Training Tesla's fleet generates real-world data, X and Grok have additional data and SpaceX has its own data. Currently Tesla and SpaceX data are separate and have corporate boundaries and conflict disclosure issues Once merged, this can all be combined for more data, all in one place, trained across the same GPU clusters And of course big picture, there is plenty of overlap across all of SpaceX and Tesla products, so having it under one roof means they can build for a more seamless future (see my post below for more details on this) What else am I missing? Also, I remain bullish on both of these companies and both have a spot in my Milk Road PRO portfolio. You can see my entire portoflio here: link.milkroad.com/t5837z

What it’s like trying to rent AI/SI compute

DAMN: #Bears star QB Caleb Williams took a shot at ESPN’s Get Up crew after they all picked the #Falcons to beat Chicago. “I saw one thing where it was Falcons, Falcons, Falcons, Falcons. Look up at the scoreboard and it’s Bears” Caleb kept the receipts.