The announcement landed Wednesday from a company that did not exist two years ago. Wonderful, an Israeli-Dutch startup that builds software to coordinate artificial intelligence agents inside large companies, said it has raised $550 million in a Series C round led by Insight Partners. The financing values the company at $5 billion, more than double the $2 billion attached to its Series B in March. Salesforce joined as a new investor, alongside returning backers Index Ventures, IVP, Vine Ventures, 9Yards and Bessemer Venture Partners.
The markup is the story. Six months ago Wonderful raised $150 million at a $2 billion valuation. Today the same company, with the same product line and more customers, is worth two and a half times as much. Investors are paying for position in a category they believe is about to consolidate: the layer of software that decides which agents do which work inside an enterprise, and that keeps the agents honest.
Wonderful was founded in 2025 by Bar Winkler and Roey Lalazar, who met in Israel’s startup economy. Mr. Winkler previously built Approve.com, a payments platform acquired by Tipalti in 2021, and was an early employee at ironSource, the app monetization company. Mr. Lalazar, the chief technology officer, previously founded Kaps, an AI localization company. The company is headquartered in Amsterdam, keeps a development center in Tel Aviv, and says it now operates in more than 35 markets with about 650 employees. Including the new round, it has raised more than $800 million in under two years.
The product is called Wonderful AI OS, and the company describes it as a shared operating layer that coordinates agents, workflows, AI-native applications, enterprise context, integrations and what it calls governed execution. Stripped of the jargon, the pitch is this: companies have adopted scattered AI tools, and they now need a layer that routes work between agents and existing systems, supplies the context those agents need, and decides what software is allowed to do on its own.
The emphasis on governance is aimed at the people who actually buy enterprise software. Chief information officers will not let agents touch customer records, inventory or payroll without controls, and Wonderful treats compliance as a feature rather than an obstacle. Its argument is that a neutral operating layer, sitting above any single vendor’s tools, is easier to audit and easier to switch than a platform’s own agent stack.
The company also stations resident engineers at customer sites to wire AI into real business processes, a services-heavy approach that is expensive but converts early adopters into references. Funding will accelerate product development and expand those deployment teams globally, as enterprises move from isolated AI experiments to organization-wide change, the company said.
Insight Partners led the round for the third consecutive time, a vote of conviction that carries weight in a market where lead investors rotate. The New York firm has said it plans to help Wonderful triple the size of its Israel-based development center over the next year.
Salesforce’s presence is the round’s most interesting detail. The customer-relationship software giant has been building its own AI agents under the Agentforce brand, yet here it is backing a company that coordinates agents across vendors, including Salesforce’s competitors. The logic is a hedge: enterprise software is migrating toward agent-based operation, and no one knows which company will own the coordination layer. Investing in a plausible owner of that layer is cheaper than betting the migration entirely on one’s own platform.
The round is also a wager on the shape of enterprise AI. Point products proliferated first: assistants for writing, coding, customer service and analytics. Wonderful’s argument is that the next phase belongs to software that manages the assistants themselves, deciding which tasks run automatically, which need human approval, and how the agents share a single source of truth about the business.
The skepticism is easy to sketch. Wonderful does not disclose revenue, its embedded-engineer model carries the margin drag of a services business, and the category is crowded with well-funded rivals chasing the same chief information officers. Valuations in enterprise AI have moved faster than the contracts behind them, and a $5 billion price for an 18-month-old company assumes the coordination layer becomes a standard part of the corporate stack rather than a passing phase of the AI cycle.
What the round says about the market is clearer than what it says about Wonderful: capital is rushing to own the plumbing of agentic AI, and investors have concluded that the company that coordinates agents will be worth more than most of the agents it coordinates. Every major software vendor is trying to build the same layer, which is exactly why a neutral startup can be valuable and vulnerable at once.
For Mr. Winkler and Mr. Lalazar, the next test is the unglamorous one: winning the trust of customers who must let an outside operating layer run software that touches their core processes. The valuation says the market believes they can. The resident engineers say the founders know it will be won one account at a time.
