American Express said on Friday it had agreed to acquire Hyper, the agentic expense-management startup founded by Marc Baghadjian, for an undisclosed sum, with the deal expected to close by the end of the second quarter. [1] [2] The Forbes 30 Under 30 founder built Hyper on seed capital that included an investment from Sam Altman, pitched as the AI agent that would unbundle the corporate card. On Friday he sold it to the card.

The transaction follows Amex's 2025 acquisition of Center, another expense platform aimed at small and mid-size businesses. [3] Three days before the Hyper announcement, on April 14, the company unveiled Agent Purchase Protection, a new policy covering fraudulent and unauthorized charges initiated by AI agents acting on a cardholder's behalf. [4] The sequence reads like an engineering diagram: build the liability product, then buy the agent that triggers the liability.

Agentic commerce was supposed to route around the card networks. A software agent — the pitch went — would comparison-shop, negotiate, and transact without the cardholder ever picking a rail. Visa and Mastercard built working groups. OpenAI launched Operator. Stripe shipped developer tools. For a brief stretch in 2025, the category looked like the latest fintech vertical ready to disintermediate a trillion-dollar incumbent.

Then Amex did what Amex has always done. It bought the disruption, put it on the network, and billed the merchant. The Hyper deal follows the 2024 Center acquisition into a pattern: Amex now owns the card, the expense software that books the card's transactions, and — as of Tuesday's policy — the liability wrapper that covers whatever an agent does with the card. [5] The agent is no longer a threat vector. It is a new customer-acquisition surface.

Baghadjian's trajectory is the story compressed. A founder profiled as the future of unbundled enterprise spend is, by Q2 close, a line on the Amex org chart. The Altman seed ticket returns at whatever multiple Amex paid, which the companies declined to disclose. The founder's next deck will describe the acquirer's distribution as an accelerant. The deck will not be wrong.

What the transaction reveals is the shape of capture. Every revolution that gets bought by American Express stops being a revolution and becomes a product line. The interchange fee that the agents were supposed to eliminate survived, because the agents now run on the card. The credit card industry is not being disrupted by agentic commerce. It is acquiring it, one Forbes alum at a time.

The Hyper deal is not large by Amex standards. Center's 2025 purchase cost was also undisclosed; neither transaction moves the earnings dial a quarter-cent. [6] What they move is positioning. On the same week the company told cardholders that it would insure their AI agents' purchases, it bought one of the best-known AI agents doing the purchasing. The two announcements were not coincidence. They were a single product launch, sequenced to look like two.

The reader who finishes Baghadjian's Forbes profile on Monday and learns of the sale on Friday has watched a category life-cycle collapse into a single news week. That is how agentic commerce ends — not with a refusal, but with an offer sheet.

-- THEO KAPLAN, San Francisco