🕶️ SoftBank's $60B IOU to OpenAI

When the companies that build the roads start reserving all the lanes for themselves, everyone else learns to navigate differently.

SoftBank is shuffling debt the size of small nations to keep its OpenAI bet alive, while SpaceX turned Falcon 9 into a private shuttle and told the rest of the industry to wait until 2029.

The pattern is consolidation dressed as prioritization, and the second-order question is starting to land: what happens when your growth depends on infrastructure controlled by someone who competes with you?

Most workplace training teaches people what to do. Very little lets them practice it.

Synthesia Roleplay Sessions lets employees rehearse real conversations with AI video agents, from sales calls to difficult management discussions, and get instant feedback before the real thing.

It's a smarter way to turn knowledge into skills.

In Masayoshi Son’s world, the balance sheet reads like a science-fiction prologue. This quarter’s near-¥148bn profit is flattered by mark-ups on OpenAI, while SoftBank quietly shoulders more than $60bn in AI commitments and a wall of debt coming due.

By year-end, SoftBank must fund a ~$22.5bn tranche to OpenAI, on top of roughly $30bn in obligations in late 2026, including bridge and margin loans pinned to Arm. A stalled plan to borrow against its OpenAI stake and a higher-than-admitted loan-to-value ratio turn the AI thesis into a leverage puzzle. That’s the backdrop as its funding plans face reckoning.

The earnings call is now a financing test: can Son turn mounting OpenAI commitments and volatile paper gains into hard cash before markets demand proof, not projections?

⏰ Your laptop doesn't actually know what time it is.

Neither does your phone. Or your smartwatch. Or even the atomic clock sitting in a military bunker.

Every device you own is constantly asking someone else for the correct time, and the entire internet depends on everyone agreeing on the same answer.

SpaceX doesn’t have a launch shortage so much as a preference stack, and Starlink sits at the top. Starlink’s share of Falcon 9 launches has climbed from 54% in 2020 to roughly 79% today, and multiple satellite operators are now told the rocket is sold out until 2028–2029, according to reporting on the squeeze.

The math explains the behavior: Starlink generated about $11.4B in 2025 vs. $4.1B from launch, so every external payload displaces a higher-margin internal one. SpaceX has already acknowledged it “may prioritize our own launch payloads,” a statement that turns Falcon 9 into a quasi-closed platform and deepens the “valley of death” between building a satellite and orbit.

This bottleneck makes room for alternatives; from Rocket Lab’s vertical integration to Europe’s stumbling $3.1B constellation push, while inviting regulators to ask when preferential scheduling becomes essential-facilities abuse.

Scale your growth: Join names like HubSpot, Ahrefs, and Deel. Put your product in front of 85,000+ startup founders. Let's talk.