A new analysis reveals that AI’s next wave extends far beyond OpenAI and Anthropic, with infrastructure, coding, data, and enterprise startups already generating billions in annual revenue.
For years, the AI conversation has centered on OpenAI, Anthropic, Google DeepMind, and xAI. That focus has overshadowed a much larger shift happening across the private market. A new class of AI startups has quietly grown into multibillion-dollar businesses, generating real revenue at a pace few expected just a few years ago.
A new analysis shared by Deedy Das, an investor, market observer, and partner at Menlo Ventures, shows that 19 private AI startups, excluding the three largest AI labs, are generating a combined $18.3 billion in annualized revenue. Together, the companies carry an estimated combined valuation of about $319 billion, offering one of the clearest snapshots yet of AI’s emergence as a large-scale commercial industry.
The list spans AI coding, infrastructure, data, voice, video generation, cloud computing, and developer tools. It paints a picture of an industry where value is spreading far beyond foundation model developers.
Cursor Leads the AI Startups Revenue Race With a $4 Billion Run Rate
At the top of the list is Cursor, the AI-powered coding platform reportedly being acquired by SpaceX. Cursor leads the group with a reported annual revenue run rate of $4 billion and an acquisition price of $60 billion, making it the highest-revenue company among the startups included in the analysis.
Data infrastructure companies occupy many of the top spots. Mercor, which connects AI companies with experts for training data, is reportedly generating a $2 billion run rate and carries a rumored $20 billion valuation. Scale AI, backed in part by Meta, follows with an estimated annual revenue run rate of $2 billion and a valuation of $29 billion.
Training data has become one of the AI industry’s most valuable assets. Companies building advanced models continue to spend billions collecting, labeling, and validating data needed to improve model performance. That demand has created an entirely new category of startups serving the AI ecosystem.
These 19 AI Startups Are Generating $18.3 Billion in Annual Revenue and Are Worth a Combined $319 Billion
Forget Chatbots: AI Infrastructure Startups Are Quietly Building Multi-Billion-Dollar Businesses
Infrastructure companies make another strong showing. Together AI, Fireworks AI, Lambda, Baseten, and Crusoe have each built businesses around GPU clouds, model deployment, and inference infrastructure. Their presence near the top of the rankings reflects a simple reality. Running AI models at scale requires enormous computing resources, and demand for that infrastructure continues to climb.
Developer tools are another standout category. Beyond Cursor, companies such as Replit and Lovable have reached estimated annual revenue run rates of about $525 million and $500 million, respectively. The growth of AI-assisted software development has made coding platforms among the fastest-growing businesses in the sector.
Consumer AI has not been left behind. Perplexity, the AI search startup, has reached an estimated $500 million annual run rate alongside a reported valuation of $22.6 billion. AI voice company ElevenLabs, video generation startup Kling AI, image generation platform Midjourney, and video creation company Higgsfield have each crossed or approached the same revenue milestone.
The rankings highlight another trend. Several of the highest-revenue companies are not foundation model developers. They are building products and services around AI rather than creating the underlying models themselves. That shift suggests the biggest commercial opportunities may increasingly come from applying AI to solve practical business problems instead of training the next frontier model.
These AI Startups Are Growing So Fast They’re Already Generating Billions in Revenue
The combined valuation of roughly $319 billion reflects strong investor confidence, though many of the figures are private-market estimates or reported funding discussions rather than public-market prices. Revenue estimates likewise vary in certainty. Some were reported directly by the companies, others are industry estimates, and a few are based on market rumors.
Even with those caveats, the broader takeaway is difficult to ignore. Less than four years after generative AI entered the mainstream, nearly twenty private AI startups have already grown into businesses producing at least $500 million in annualized revenue. Several have crossed the billion-dollar mark, a level of commercial scale that usually takes technology companies many years to reach.
The AI economy is no longer defined by a handful of famous labs. A broader ecosystem of infrastructure providers, developer platforms, enterprise software companies, and specialized AI applications is now generating billions in annual revenue, suggesting the industry’s next chapter will be written by far more companies than many expected.
