$META reports its Q3 earnings tonight Here's what the Street expects: 1. Revenue: $63.34B, up ~24% y/y and at the TOP of its $61B to $64B guide 2. EPS: ~$6.53-$6.64 (some estimates sit near $3.56 because they include a ~$10B legal charge) 3. Q3 capex: ~$41.7B and 2027 capex: ~$193B The real swing factors are the Q4 revenue guide and 2027 capex, plus how many people are actually using Muse Meta is becoming a beast of an AI company across its new models, Muse and its impact on its advertising business

WOW! We just got guidance for Q1 2027 from the SpaceX CFO They are expecting $33B-$37B in total revenue in Q1 27', a 150% increase in the next 6 months! The bulk of this comes from reaching 3 GW of compute by EOY and least 2.5 GWs/quarter in 2027 The big advantage SpaceX has is their ability to energize GPUs. The bottleneck right now is the blades for turbines which take 12-18 months to deliver But Elon says SpaceX is doing advanced casting to build their own blades at scale much faster, which can get them to at least 13GW total by EOY 2027, but potentially upwards of 20GW And because SpaceX can deliver these GWs so fast, they payback period takes just months (rather than years). Giving them the ability to fund the Capex required to build this amount of compute through revenues, rather than only relying on debt

WOW, this is the number that matters for Tesla earnings! FSD subscriptions hits 1.74 MILLION, up 260k from last quarter…its biggest jump in FSD subs ever! Tesla says they achieved record FSD attach rates in North America in Q3 2026, with 75% of all new deliveries including FSD subscriptions, up from 55% in Q2 2026 FSD is driving demand for Tesla vehicles I own a Tesla myself and I can tell you first hand it’s a game changer. I use it 99.9% of the time and it’s changed what it means to drive for me I get work done while my car drives me and I never have to think about traffic or directions. Once you try it you will see why this is revolutionary technology This is what keeps me bullish on the stock even though its growth is not as strong as its valuation premium is. Tesla is ready for a truely breakout year in 2027 imo because of this technology, from car delivered to the robotaxi networK

$SPCX isn't just the fastest to build AI compute, it's also the most cost-effective! And after the week neoclouds just had, that really matters SpaceX installed 1.4 GW of compute at $29.6B/GW, with the latest build at $32.6B/GW Most estimates put a GW of compute at $40-50B That cost edge is why SpaceX's CFO said on the Q2 call that new compute deployments are paying back in LESS than 1 year (insane for infrastructure) SpaceX expects to end this year with over 2 GW and is guiding to "closer to 10 GW than 5 GW" by the end of 2027 (and Elon's hinted it could be even more) Adding ~8 GW in 2027 at their current $32.6B per GW is roughly $260B of capex That's a massive number, but I think they can fund it no problem SpaceX ended Q2 with $100B in cash They're reportedly raising another $40B in debt specifically to buy NVIDIA GPUs And they're guiding to a $100B annualized revenue run rate by December (I think they end the year closer to $125B) But here's the part most people miss With a sub-1-year payback, the GWs they turn on in early 2027 help pay for the GWs they build in late 2027. The buildout partially funds itself, and they can stay profitable while doing it My estimate is $350-450B of ARR by the end of 2027 There's also a second path to 10 GW (or more): buying distressed neoclouds CoreWeave's bonds are yielding over 15% (distressed territory) and the whole sector got hammered this week. If any of these smaller neoclouds mismanage their capital (or execution), they may soon be looking for a buyer... and no one is better positioned to step in than SpaceX Whether they build or buy, I think SpaceX will hit 10GW of compute by EOY 2027 for that reason, $SPCX is a core asset I hold in my Milk Road PRO portfolio. You can track my entire real-time portfolio here: link.milkroad.com/t5837z