Bird’s $400m Dividend Recap: Owners Cash Out Before the AI Agent Product Proves Itself
The $400 million term loan is about 2.4 times 2025 EBITDA and goes to shareholders. The company’s communications platform throws off cash with 120 staff; its agent product has yet to show paid usage.
Bird.com has arranged $450 million of financing as it launches communications infrastructure for AI agents. The package includes a $400 million term loan and a $50 million revolving facility. Bird made $165 million of EBITDA last year, putting the term loan at about 2.4 times that figure. Proceeds go to current and former shareholders through a dividend recapitalisation. The recap monetises the legacy messaging business now, before the agent product has shown that anyone will pay for it. The company announcement gives the terms.
The leverage is moderate on the disclosed term loan. Federal banking guidance describes leveraged loans using thresholds above 3 times senior debt to EBITDA or 4 times total debt to EBITDA, and Bird’s 2.4 times sits below both. That is not the whole burden: EBITDA is before interest, working capital and capital spending, and Bird’s full debt stack, interest cost and covenants have not been disclosed. The $50 million revolver can bridge cash timing if drawn, but it adds debt. The Federal Reserve’s guidance sets out the benchmarks.
The cash cost is easier to see than the ratio. Bird has not disclosed the coupon. If the loan carries 9%, interest alone takes about $36 million a year, more than a fifth of 2025 EBITDA, before any principal repayment, tax or capital spending. Each percentage point on the coupon moves that bill by $4 million. That leaves real room, but it is room the legacy business must keep earning every year while the company also funds a new product.
The headcount explains why lenders will lend. Bird, formerly MessageBird, has cut its staff from a peak of more than 1,000 to about 120, according to Tech.eu. Spread across that workforce, $165 million of EBITDA is roughly $1.4 million per employee. That is an unusually cash-generative profile for a software company: a mature messaging platform, a largely fixed network and few people to pay. It gives lenders a stable cash stream to lend against, and 75% of revenue now comes from U.S.-headquartered customers, including large technology companies.
The same figure explains why the agent product is the risk. A company run for cash with 120 people has little spare engineering and support capacity. Building a new product line means hiring, and every new salary is paid from the cash flow that now also services $400 million of debt. The cost cuts that made the recap possible also narrow the room for a build.
The product gives software agents access to messaging, calls, email and eSIM lines. Bird already operates business communication services, so it can offer a network and APIs that agents may use to reach customers, suppliers or employees. It is a sensible adjacency: agents can plan a task but need an approved channel to complete it.
The commercial hurdle is customer control. An enterprise needs to specify which agent can contact which person, what messages it can send and when a human must approve or take over. Identity, permission and audit records have to fit existing risk rules. If each deployment needs custom integration and manual review, service costs rise with every customer, which a 120-person company cannot absorb for long. If controls can be reused, Bird can attach more traffic to a platform it already runs at high margin. Bird has not reported paid agent deployments, active usage or product revenue, so there is no adoption measure yet to set against the legacy platform’s $165 million.
The financing route says something about the owners’ view of value. Bird’s last widely reported private valuation was $3.8 billion in 2021, when MessageBird extended its funding round, according to Business Insider. Venture-backed companies normally return capital to investors through a sale or a listing. Returning cash through borrowed money instead suggests the owners see no exit available at a price they would accept, and the 2021 mark is the obvious reason. Selling or listing now would test that number in public. A recap lets holders take $400 million off the table without doing so.
The counter-argument is that the owners keep the upside. They have taken cash without selling shares, so if the agent business works, they still own all of it. That is true, and it is why a recap can be rational for owners who believe in the product. But the structure distributes risk unevenly. Shareholders already have their cash. Lenders hold a senior claim on the legacy cash flow. The company holds the debt and the product bill. If agent revenue is slow to arrive, the constraint lands on Bird’s operating budget, not on the investors who were paid.
Lenders are protected in a way shareholders chose not to be. Their claim ranks ahead of equity on a business with low staff costs and a customer base concentrated among large U.S. companies. If the agent product fails, the legacy platform still services the loan; if it succeeds, lenders collect only interest. The downside of the agent bet sits with the company’s budget.
The next test is disclosure. The desk expects Bird to report its first agent-product revenue or customer count before it seeks further financing. Until then, the recap reads as a bet that the legacy platform can carry both the debt and the build. The legacy platform clearly supports the debt. Whether it can also pay for the build will only show once the agent product earns recurring revenue.