AI investment spreads across three deployment paths
New rounds for Lovable, Thrive Holdings and Blacksmith show capital spreading across app creation, the rebuilding of traditional-company workflows and testing the growing volume of AI-assisted code.
Artificial Intelligence··Morning
Three rounds at three different scales
Three funding rounds announced on the same day show capital reaching AI businesses with different sizes and operating models. Swedish app-building company Lovable confirmed a 400 million dollar Series C led by Menlo Ventures and the Scaleup Europe Fund at a valuation of 13.3 billion dollars. In December 2025, it had raised 330 million dollars at a valuation of 6.6 billion dollars. Thrive Holdings, spun out of Thrive Capital, raised 2 billion dollars led by SoftBank, D1 Capital Partners and Altimeter Capital at a 12 billion dollar valuation; OpenAI had taken an ownership stake in December 2025. Code-testing startup Blacksmith closed a 45 million dollar Series B led by Peak XV Partners at a valuation of 550 million dollars. Its Series A less than a year earlier was 10 million dollars at a valuation of 60 million dollars, and total funding now stands at 58.5 million dollars. These figures do not describe companies at the same scale. Lovable is growing around a direct app-building product, Thrive Holdings around acquired businesses, and Blacksmith around software-validation infrastructure. Placing the rounds side by side shows that the funding is not confined to one product category; each source describes a different route to deployment and revenue.[1], [2], [3]
Growth signals run from products to workflows
The operating indicators reported for Lovable point to a broad end-user surface. TechCrunch puts its annualised run rate at 500 million dollars as of June 2026, with 60 million hosted projects and 900 million monthly visitors. The company signed a multiyear Google Cloud agreement for a fivefold increase in usage and also offers an AI model trained in house. Blacksmith's growth indicators sit in a narrower part of software development. Founded in 2024, it sells continuous-integration services and Codesmith, a coding agent that automatically fixes failed code checks. Its customer count rose from more than 700 to more than 5,000 in under a year, while headcount increased from 10 to about 30. TechCrunch reports an annualised run rate in the tens of millions of dollars and says some customers spend more than 1 million dollars a year. Thrive Holdings follows a route different from selling one product to outside customers. It acquires traditional businesses and rebuilds their workflows around AI. These three arrangements place use across a range that runs from an app-building surface close to end users, through internal code validation, to transformation carried out through ownership of operating businesses.[1], [2], [3]
Capital is not going only to model development
Thrive Holdings' two current platforms provide concrete examples of how AI is being applied to established operations. Its accounting platform, Current, covers more than 50 firms and over 2,000 professionals, and a tool called TaxAI processed more than 7,000 tax returns at 98 percent accuracy. Its information-technology platform, Shield, covers about 20 companies, and TechCrunch reports a 36-fold improvement in help-desk resolution times. OpenAI staff work with portfolio companies on adoption under the partnership, and a third platform for regulatory services is planned. At Blacksmith, chief executive Aditya Jayaprakash says code validation remains a growing bottleneck as people write more code. Lovable operates at the direct production surface through app building and hosting scale. Read together, the sources place investment against three different constraints: creating an application, rebuilding an established workflow and testing an expanding volume of code. This is not a performance ranking. It is a view of the operating layers where the companies seek revenue and growth. The new rounds show capital in the AI economy reaching not only foundation-model developers but also businesses that connect models to products, operating companies and validation processes.[1], [2], [3]
