Newsletter
This Week in Sovereign: September 27, 2026
The largest US bank failure of 2026, UK remortgages settled in tokenised sterling, and Google's AI chips head for orbit
27 September 2026
Good day! This week: the biggest US bank failure of the year so far, the first UK remortgages paid in tokenised sterling, Google putting AI chips into orbit, and the largest AI labs naming their own regulator.
This Week's Spotlights
- Steve Farrell Left Silicon Valley for Humanity's Team and Its Case for Interconnectedness
- Allstate Agency Owner Pedro Hernandez Grows His Mesa Agency to a $5.4 Million Book of Business in Three Years
- Michael J. Gugliotti Brings Security Governance Into the Boardroom
- Rick Tabor's Sci-Fi Political Thriller The Blink of Humanity Explores the Dire Cost of Political Corruption
This Week's Digest
Regulators Shut Nano Banc, the Largest US Bank to Fail in 2026 So Far
California regulators closed the single-branch Irvine lender on Friday, 25 September. Nano Banc held $736 million in assets at the end of June, more than double the next-largest failure this year, and the FDIC puts the cost to its insurance fund at about $114 million.
The FDIC sold substantially all of the deposits and about $476 million of assets to Sunwest Bank of Sandy, Utah, so customers keep access to their money and the branch reopens under Sunwest. Neither regulator has given a reason for the failure.
Lloyds, NatWest and Barclays Complete First Remortgages Paid in Tokenised Sterling Deposits
Two remortgages have been settled in tokenised sterling deposits, which UK Finance says are the first live customer transactions of their kind. The funds were locked and then released automatically when each remortgage completed.
The payments ran on a single platform shared by seven banks, the form of digital bank money the Bank of England has said it would rather see than stablecoins. Earlier bank blockchain projects stayed inside one lender at a time.
Google Sends Its First AI Chips to Orbit Next Week to Test Data Centers in Space
Google will fly a handful of its own AI chips on a prototype satellite built with Planet, to find out whether they survive launch, radiation and the temperature swings of orbit. It is the first flight of Project Suncatcher.
The appeal is power and siting: solar panels in low Earth orbit can produce up to eight times what the same panels make on the ground, and a satellite needs none of the grid connections, cooling water or land that new data centres wait years for.
Google, OpenAI and Anthropic Named Their Own AI Regulator and Floated a Federal-Rules Skeptic as Its Chief
The three labs are closer to standing up a body to set frontier AI standards among themselves, tentatively called the Standards Authority for Frontier AI. Names floated to run it include Sriram Krishnan, who left the White House after publicly rejecting a licensing regime for AI.
It fills a gap Washington has left open, with no federal standard-setter for frontier models. What the body could not do is compel anyone outside the group, or inside it, to follow the standards it writes.
Also This Week
- Stripe Is Becoming the Wallet Behind Competing AI Shopping Agents
- Your Phone's Lifespan Is Now Written Into European Law
- Anthropic's Claude Surfaced an Enzyme System Biologists Had Overlooked
- AI Inference Startups Fal and Fireworks Weigh New Funding at $15 Billion and $30 Billion
- Tokenised Assets Can Now Settle in Central-Bank Money Across the Euro Area
- AI Startups Are Staying Small by Hiring Fewer Junior Staff
- The 2026 M&A Boom Is Concentrated at the Top of the Market
- Federal Molten-Salt Coolant Arrives for a Texas Reactor Under Construction
- Satellites Can Now Spot Wildfires From Orbit, and in 2026 the Technology Went Operational
- Drones Are Delivering Plasma to Wounded Troops in Ukraine at the Point of Injury
That's all for this week. See you next Sunday.