Visa, Mastercard and Ant want one trust layer for AI agents

Share
Visa, Mastercard and Ant want one trust layer for AI agents

Two infrastructure bets landed this morning: the payments industry agreed on how to tell a good shopping agent from a bad one, and the US Army bought navigation that doesn't need satellites.


Ant International, Visa and Mastercard said Thursday they will build a common "Know Your Agent" interoperability framework — one identity and behaviour check for AI agents that spend money, recognised across card networks, wallets, agent platforms and marketplaces. The three companies have each shipped their own agent-payment protocol in the past year — Visa's Trusted Agent Protocol, Mastercard's Verifiable Intent and Ant's Agentic Mobile Protocol — and this initiative is the glue between them: register an agent once with Ant and it doesn't need registering again with Visa or Mastercard, while each network keeps its own approval and risk rules. The work runs through BuildFin.ai, a platform convened by the Monetary Authority of Singapore.

The commercial logic is a number the companies put in their own release: McKinsey's projection that AI agents will handle $3 trillion to $5 trillion of global consumer commerce by 2030. That is why Ant's chief innovation officer Jiang-Ming Yang told CNBC the point is not capability but confidence — agents hallucinate, so "we need to make sure people feel safe." This is also a wallet story as much as a card story: more than 50 e-wallets already connect to Ant's Alipay+ system, and digital wallets carried 56% of global e-commerce value in 2025.

Our take: this is the least glamorous and most load-bearing piece of agentic commerce so far. Every lab is shipping agents that can browse and buy; nothing has answered the question a merchant actually asks — who is this agent, who stands behind it, and can I charge back when it goes wrong? A shared registry answers that in one lookup instead of fifty bilateral deals, and the side that owns the registry quietly sets the rules for who gets to transact. We tracked the Chinese end of this earlier — China's payment industry adopts a 'Know Your Agent' rulebook for AI that spends money — and the ambition now is clearly cross-border.


The US Army has awarded Austin-based Tern an $11.26 million contract for a positioning system that works without GPS, in a vehicle-mounted kit the startup calls "Google Maps for the battlefield." The firm-fixed-price Phase III SBIR award covers an "Independently Derived Positioning System" plus vehicle health and maintenance monitoring, with work in Austin and an estimated completion date of August 2027. The hardware is deliberately mundane: a device plugged into the vehicle's diagnostics port, wired to a tablet that shows the soldier position, navigation and routing.

Tern's pitch is that modern vehicles already emit enough sensor data to compute their own location on the edge, and CEO Shaun Moore says the stack passively listens to that data rather than injecting anything. Co-founder Brett Harrison used it on a 1,300-mile drive from Austin to Laguna Beach without losing position once, while conventional GPS dropped out dozens of times. Because the system is closed and self-contained, it is harder to jam or spoof than a satellite signal — which is the actual reason the Army is buying, given GPS interference is routine over the Middle East and around Ukraine.

What to watch: whether the Know-Your-Agent framework ships as a real cross-network spec or stalls as a memorandum, and whether Tern's Army award converts into a program of record rather than a one-off test.

Would you let an agent spend your money automatically if the network behind it could vouch for the agent — or is that a line you don't cross? Tell us in the comments.

Read more

Byteification turns Qwen3 and Llama 3 into byte-level models

Byteification turns Qwen3 and Llama 3 into byte-level models

The tokenizer is the last piece of the language-model pipeline that was never learned end to end — and a Nature paper out this week shows it can be swapped out for less than one percent of a pretraining budget. Researchers have retrofitted four open subword language models into true byte-level models using less than 1% of a typical pretraining budget, and the result — published 7 October in Nature — closes the performance gap that has kept byte-level models a niche curiosity for nearly a decad

OpenAI's dots go mobile and can start your Codex work

OpenAI's dots go mobile and can start your Codex work

One product update and two benchmark papers this hour, all circling the same question: how much of the work can you actually hand over, and what happens to the part you didn't? OpenAI's dots — the always-on agents that get their own cloud computer — can now be created and configured entirely from the ChatGPT app on iOS and Android, and a dot can start Codex work or pick up an existing Codex thread on your behalf. The October 9 release, billed as the first of a fresh run of updates for dots, mo

Coding agents as a hospital: Cockroach's five-month results

Coding agents as a hospital: Cockroach's five-month results

The agentic-coding debate keeps being framed as speed versus trust. Cockroach Labs just published the most detailed trust-first ledger yet — five months of numbers, including the parts that didn't work. Cockroach Labs ran its codebase like a teaching hospital for five months and says the pipeline merged well over a million lines of code with seven total reverts. In the setup the company calls MOLT Sinai, GitHub issues are patients, merging is discharge, and the humans in charge are Chiefs of M

Today in AI — October 10, 2026

Today in AI — October 10, 2026

A day where the boring machinery took center stage: proof checkers, order books, warehouses and the humble text message — plus fresh arguments about who gets to declare any of it safe. Models & Research * Mathematicians get a reliability primer for the Lean Theorem Prover. A guest post by Thomas Hales on Terence Tao's blog walks through what Lean actually guarantees — and devotes itself to the "Summer of Soundness Bugs," the string of kernel bugs found this summer that let false proofs thr