XAI AI data center economics are even better when you can start a project and finish it in 12 months and get it rented and paid off while competitors are still building. An active AI Gigawatt of data center in the hand is worth $20 billion per year while an under construction Gigawatt data center is bleeding $5-15 billion per year of cash. XAI is building them for about $30-40 billion per gigawatt while competitors need $50 billion and add on another $10 billion or more in extra financing costs for a multi-year project.
Data center project cancellations more than quadrupled to 25 in 2025 from six in 2024. Between a third and a half of all US data centers planned for 2026 are likely to be delayed or totally cancelled, Bloomberg reports, amid ongoing supply chain challenges and campus location concerns. The original plan might have been four years but now it could be longer or never.
There was an estimated 12-16GW of planned capacity in 2026. Only 5GW is currently under construction and many projects remain having been announced, but with no physical progress yet. Of the 5 GW, XAI is 1-2 GW of the under actual construction or completed.
OpenAI’s $500 billion Stargate Project has been stalled, downsized or delayed from the original 4 year timeline.
Amazon AI revenue run rate last quarter was $15 billion per year. XAI matched the $15 billion per year with the Anthropic deal. IF XAI were to rent out the rest of Colossus 2 that would be another $15-20 billion per year.
Cloud computing margins have been great and have been the envy of other businesses. Traditional cloud margins (AWS – Amazon Web Services) are stable and high at ~35% operating margins, Bitcoin mining margins are volatile/compressed (often breakeven or negative in 2025-2026), and AI cloud margins (CoreWeave, SpaceXAI, Amazon AI) offer the highest unit economics with ~75%+ gross and ~60% EBITDA margins despite heavy growth capex.
Competitors are still waiting for permits, poring concrete and waiting on transformers. XAI Capex pays itself off in 12 months instead of 3–5 years. AI multi-year demand stays insatiable. XAI capture it first. It’s the ultimate build faster, rent sooner, kick ass.
xAI recycles tens of billions into the next project 1 to 3 years earlier. XAI is compounding returns while others are still in risky construction.
Risk reduction is from shorter exposure to technology obsolescence, energy price volatility, regulatory delays, or demand shifts. A 4-year build window carries far more execution and market risk.
Tesla Building 67+ Megacharging Sites and Has Thousands of Supercharger sites
Tesla is building 67+ Megacharging Sites and has thousands of supercharger sites. One Megacharging site or ten supercharging sites can add the equivalent of one Coreweave size 10-20 megawatt AI data centers.
SpaceX SP500 will happen in 6 months. Passive S&P 500 funds WILL HAVE TO BUY roughly 19% of public SpaceX shares within 6 months under fast-tracking framework. Russell 1000 and Nasdaq 100 will buy 5.5% within 15 days of the IPO. Active benchmarked to those indices and you get to HALF of SpaceX shares. SpaceX IPO is over 10X oversubscribed and Blackrock has put in an order for $5-10 billion.
This is continuation of a 30+ year trend (internet era) and 50+ year trend (computing). The companies that built the internet, cloud, semiconductors, and now AI have compounded at extraordinary rates because they became infrastructure for everything else. Skeptics who said “this tech bubble will pop and we’ll go back to normal” in the 1990s, 2000s, or even post-2010 were wrong over entire investing lifetimes. The internet and computers did not reverse. They swallowed more of the economy every year.
AI has already swallowed large parts of the world economy — AI is in every major product roadmap, every hyperscaler’s capex, and increasingly in enterprise workflows. Demand for compute (per analysts like Dylan Patel, Semianalysis) has run 5x+ ahead of supply in key areas, and AI cloud margins are far higher than traditional cloud or Bitcoin mining.
Passive S&P 500 funds could have to buy roughly 19% of public SpaceX shares within 6mo under fast-tracking framework (it would enter the index at the est 6th spot), Russell 1000 and Nasdaq 100 may buy another 5.5% within weeks of the IPO. Thrown in active MFs benchmarked to those… pic.twitter.com/FA8MQE0LSa
— Eric Balchunas (@EricBalchunas) May 22, 2026
BlackRock is trying to buy 13% of the entire SpaceX IPO in one order.
– They want $5 to $10 billion of the $75 billion raise
– BlackRock manages $11.5 trillion in total assets
– They’re deploying from their conviction funds, not index funds
– The IPO is already 10x… pic.twitter.com/ms1YICpARv— cape (@capexbt) May 16, 2026
As an active manger myself who manages 3 bln (not sure of all the people commenting actually run money), I can tell you the risk is too high to not own it if the benchmarks fast track.
Part of what we do constantly is trying to game the passive index adds and deletes to gain…
— John Davi (@AstoriaAdvisors) May 23, 2026
Satya Nadella says AI is becoming the new “Lean” for knowledge work at $MSFT, helping companies automate processes and reduce waste.
Meanwhile, $NOW and $CRM are both well off their lows. Maybe AI isn’t killing software after all… maybe it’s making the best platforms even… pic.twitter.com/R68hH31Cyz
— Sam Badawi (@Sam_Badawi) May 24, 2026
AI Data Centers
CoreWeave (AI-native cloud) has 850 Megawatts across 43 data centers and about 250,000 GPUs. They have ~3.1 GW contracted power capacity (end-2025) and targeting over 1.7 GW active by end-2026 and adding ~5 GW more by 2030.
XAI with one deal has matched CoreWeave with about 800-900 Megawatts contracted to Anthropic. XAI will match or exceed the 1.7 GW active by end of 2026.
OpenAI talked about big data center plans but they are mostly completing in 2029 or are getting cancelled.
xAI is building a third massive data center. This time in Southaven, Mississippi, right next to its Colossus 2 facility in Memphis. Construction begins in 2026.
xAI is buying another $2.8 billion worth of turbines for its AI infrastructure over the next three years.
Brian Wang is a Futurist Thought Leader and a popular Science blogger with 1 million readers per month. His blog Nextbigfuture.com is ranked #1 Science News Blog. It covers many disruptive technology and trends including Space, Robotics, Artificial Intelligence, Medicine, Anti-aging Biotechnology, and Nanotechnology.
Known for identifying cutting edge technologies, he is currently a Co-Founder of a startup and fundraiser for high potential early-stage companies. He is the Head of Research for Allocations for deep technology investments and an Angel Investor at Space Angels.
A frequent speaker at corporations, he has been a TEDx speaker, a Singularity University speaker and guest at numerous interviews for radio and podcasts. He is open to public speaking and advising engagements.
