xAI’s 1.5-trillion-parameter model is now running against real engineering and manufacturing work at two operating companies, landing the same week SpaceX priced a $25 billion bond debut and Tesla staged roughly 80 Cybercabs in Houston.

By licensing Groq’s entire patent portfolio and hiring founder Jonathan Ross plus roughly 80% of its staff on Christmas Eve, the $5 trillion chip king is fortifying the one battlefield - inference - where in-house silicon from Google, Amazon, and Microsoft most threatens its 75% margins.

Two USPTO filings tie SpaceX’s compute ambitions to orbital data centers while Tesla’s Level 4 Cybercab moves from prototype imagery into operational documentation, a closed NHTSA steering probe, and a new Uruguay sales market.

Federal regulators opened rulemaking to let vehicles built without hand or foot brake controls prove compliance through alternative tests, just as Tesla ramps Giga Berlin to 7,500 vehicles a week and Starlink finishes wiring Alaska’s 91-jet regional fleet.

An early-termination notice on June 25th hands Musk the go-ahead to acquire a former-SpaceX-engineer startup building optical transceivers, even as X Money reaches its first paying users and Starlink weighs a direct retail phone service.

By pairing Starship’s freight-elevator payload with the higher-bandwidth Starlink V3 satellite, SpaceX drove the cost of deploying capacity to orbit from $6.55 to about 30 cents per megabit per second, the single number that decides whether the largest IPO in history is cheap or insane.

A planned 20% October output lift and 1,000 new hires at Berlin land the same week NHTSA moves to drop the manual brake-pedal rule for driver-optional vehicles and Starlink finishes wiring Alaska’s 91-jet regional fleet.

Days after the largest IPO in history crowned him the first trillionaire, Musk staked a 54x revenue leap not on rockets but on an audacious plan to run the world’s AI compute from orbit.

By aggregating hundreds of thousands of home batteries and eight million smart devices into 16.8 gigawatts of flexible capacity, Tesla is repositioning its energy business from hardware sales toward grid capacity sold as a service just as AI load strains the grid.

By dropping self-contained compute cabinets into Supercharger sites with spare capacity, Tesla can run distributed inference on infrastructure that’s already built, powered, and cooled—while SpaceX shares whipsaw from an all-time low of $147.11 back to $163.

A four-times-oversubscribed bond sale reframes SpaceX as investment-grade infrastructure while Tesla’s 25 GWh NatPower agreement, a Finnish FSD review, and the Starfall demo push the empire deeper into energy, autonomy, and orbital services.

As hyperscale equipment prices quadruple in eight months, the industry is racing toward 800-volt racks, model-specific silicon, and a co-designed home inference box that fuses compute, power, and connectivity into one appliance.

The nation’s densest cluster of compute is now cycling backup generators as primary power and colliding with a new governor’s moratorium, exposing how permitting fights over nuclear and battery storage are throttling the AI boom’s real bottleneck: electricity.

The best product on the planet still can’t get into most garages, and that same adoption friction is what gives the economy room to absorb AI - unless a slow bleed pushes unemployment to 7% and the ballot box turns.

A $150-million-per-month lease of NVIDIA GB300 capacity to Reflection AI, a unanimous Texas Supreme Court win over Boca Chica beach closures, and United’s first wide-body Starlink transatlantic flight together reframe SpaceX as a multi-market infrastructure company while Tesla fights to control the

A government database now holds the strongest real-world evidence for Tesla's self-driving in Europe. The figure that decides what happens next is the one Spain hasn't named yet.

For about two years, the thing blocking Tesla's supervised Full Self-Driving in Europe was never really the software. It was paperwork. The cars could do it; the regulators had not signed off. So the Tesla story I keep coming back to this week isn't a new FSD build or a flashy demo. It is a spreadsheet sitting inside a Spanish government agency. To me, that is the more important development.

Spain's national traffic authority - the Dirección General de Tráfico, or DGT - has published data showing Tesla's supervised test fleet has covered roughly 275,000 kilometers on public roads. Call it 171,000 miles, run by 30 cars since November 2025, with no serious incidents on file. The cars operate under a national permit that clears them for any road in the country and runs through November 2027.

Key Takeaways

• 275,000 km. That is the supervised distance Tesla's 30-car fleet has put on Spanish public roads since November 2025 with no serious incidents logged by the DGT - roughly 171,000 miles.

• The figure comes from a government registry, not a Tesla safety report. Third-party numbers held by a regulator are the kind of evidence a European approval case actually runs on.

• Tesla's distance dwarfs the rest of Spain's top testing tier - the pre-deployment phase, the one cleared for fleets of more than 10 vehicles. The next-biggest entry, a set of ALSA buses, has covered about 1,400 km. The only other Level 2 tester on file, Wayve, is running three cars.

• Every one of those miles had a human behind the wheel, fully responsible. This is supervised driving, not a robotaxi, and the DGT has never publicly defined what counts as a serious incident.

• One clean government record becomes a reference other regulators can cite. That second-order effect is how a slow European approval timeline could compress faster than almost anyone expects.

• The Netherlands logged 1.8 million km over 18 months in its own program. That hints at how much clean data Spain might want before it lets Tesla go customer-facing.

Fresh off the largest IPO in history and Musk’s own trillionaire milestone, SpaceX is projecting $1 trillion in revenue by 2030 - a 54x leap that only pencils out if the newly absorbed xAI turns orbit into the world’s data center.

A June 18 intent-to-use filing points Tesla’s energy and AI stacks at data center hardware, while SpaceX lines up a June 23 Falcon 9 test of a cargo capsule built to bring payloads home from orbit.

Two Palo Alto software engineering roles built around PyTorch and low-level GPU tooling point to Tesla owning the AI compute layer that sets how fast it can scale FSD, Optimus, and the CyberCab, even as viral crash narratives and safety scrutiny escalate alongside.

The June 12 IPO priced SpaceX as the world’s sixth-largest company because underneath the rockets sits a plan to turn never-setting orbital solar power into cognition, using the $20-a-year human brain as the floor and the sun as the ceiling.