Yup Reply to @andreaesposito: Talked with someone in the know about SpaceX / xAI construction roadmap last week. I am now convinced they will single-handedly accelerate 10x the US infrastructure building capabilities, through aggressive vertical integration + applying Elon's algorithm. It just makes sense: > data centers > rockets and satellites factories > launch towers > LNG/LOX production plants and pipelines > space data centers factories > eventually, lunar and martian infrastructure Construction will become one of their biggest competitive advantages, and they'll become the best construction company in the world. They'll also deploy compute faster and more efficiently than anyone else, and rent out inference to the model makers. I don't know if this means they will also win on models, but Elon will definitely try. Him who controls the weights controls the culture. Interesting times. Lfg 🇺🇸 Quoted @Object_Zero_: Capital Expenditure (capex) The global economy is $128 trillion Global capital expenditure (capex) is $28.5 trillion By sector, global capex looks like this… 1. $3.9 trillion - Manufacturing, all industries 2. $3.6 trillion - Residential property 3. $2.6 trillion - Non-residential property 4. $1.5 trillion - Electricity infrastructure 5. $1.4 trillion - Transport infrastructure 6. $1.1 trillion - Oil and gas 7. $950 billion - *DATA CENTERS & COMPUTE* 8. $750 billion - Defense 9. $620 billion - Healthcare and education 10. $300 billion - Telecoms 11. $250 billion - Water infrastructure 12. $200 billion - *SEMICONDUCTOR FABS* […] The capital structure that underpins the above capex shortlist varies materially by sector, because the assets are fundamentally different in financial terms. What matters is their; • Asset life / duration, it can vary from 3yrs to 50yrs • Cash flow forecastability, oil & gas is volatile, rent less so. • Residual value, highly specialized assets are harder to resell Capex is capitalised with a combination of debt and equity, what matters for pricing the capital is the combination of these things (there’s other stuff but I’m simplifying). Global capex is about 50% funded by loans and debt instruments, and 50% equity. What is happening with data centers has nothing to do with global liquidity, which is frankly vast and would like a lot more big projects to originate. The issue is that a lot of data center projects are delayed or overrunning. Execution is what matters, you must hit the cash requirement milestones. Many are not. The build out is made up of myriad construction projects that compete for the same land and suppliers. Much of the debt financing comes from the same few pools and those pools have reached the point where the marginal loan is net negative to the value of the pool. It seems that if the economy has to build more than $800bn of data centers in US, then projects start tripping over each other. This is a major problem, because capex was forecast to hit $2 trillion by 2028 but the construction industry seemingly struggles to deliver that level of site development. What lenders want and compute operators need is a non-competing supply chain that can deploy data center capacity. A supply chain that is not competing to buy the same powered land, plant equipment and diggers, something that isn’t built out in concrete and grid connections. The available investment capital and lending is actually greater than ever, and increasing, it’s just looking for another deployment channel, because powered land is stuffed. Capital is readily available but bankable delivery capacity is what is scarce. An alternative industrial supply chain (alternative to powered land) can massively expand the set of financeable projects, reduce delivery bottlenecks and make completion more predictable. Such opportunities exist, and they are able to absorb the stranded $1 trillion capex, and have the delivery capacity to scale well beyond powered land. What is required… is to make those new channels bankable. Asset financing precedent and mature liability and underwriting contract law all help here, along with actually building and qualifying the delivery capacity channel. But this is what 2027 is going to be all about without people actually describing it as this. Construction industry prices will likely rise, but this is somewhat a case of ‘nine women cannot have a baby in a month’. So there’s likely some market dislocations coming as major capital pools move to stand up the new bankable deployment capacity and get funds deployed into the most lucrative part of the economy. Interesting times!