AI software
AI software is the layer of frontier model labs, consumer assistants, vertical workflow products and enterprise applications built on top of large language models, image generators, video generators and agentic reasoning systems. ChatGPT crossed 300M weekly active users in late 2024 and Anthropic's revenue run-rate moved from roughly $1B at the start of 2024 to a reported $8B by late 2025, with the category now anchored by four model labs commanding a combined valuation north of $700B. The space covers everything from horizontal consumer chatbots to vertical workflow products in legal, sales, customer support, software engineering and knowledge management.
The sector spans foundation model labs, consumer AI assistants, enterprise search and knowledge AI, vertical AI applications, AI coding assistants, AI for customer support and sales, AI agents and autonomous workflows, and AI safety and evaluation tools.
Revenue comes from API token-based pricing on foundation models, consumer subscription tiers for assistants and creator tools, per-seat enterprise SaaS contracts on vertical applications, hyperscaler revenue-share deals that route inference through Azure, AWS Bedrock, Google Cloud Vertex and Oracle Cloud, and a growing line of usage-based pricing on agentic workflows priced per task or per outcome.
AI software is part of AI & ML.
$122B
Global market size
128
Public companies
Key VC investors
Key strategic buyers
How AI software companies monetize?
AI software companies monetize through API token pricing, per-seat enterprise SaaS and usage-based agent contracts.
API token-based pricing
Pay-per-token pricing on input and output tokens through model APIs. The reference model for OpenAI, Anthropic, Google and Mistral; pricing has dropped roughly 10x for equivalent capability since GPT-4 launched.
Consumer subscriptions
Monthly tiered subscriptions for consumer AI assistants and creator tools. ChatGPT Plus at $20, ChatGPT Pro at $200, Claude Pro at $20 and Perplexity Pro at $20 anchor the category.
Per-seat enterprise SaaS
Annual per-user contracts for vertical AI applications. Harvey, Glean, Hebbia and Cresta use this model; ACVs run from $30-$150 per user per month at scale.
Hyperscaler revenue share
Model labs route inference through Azure (OpenAI), AWS Bedrock (Anthropic), Google Cloud (Anthropic and Mistral) and Oracle Cloud (xAI) in exchange for compute commitments and revenue share. Re-routes most of the venture capital raised back to GPU capex.
Usage-based agent pricing
Per-task, per-resolution or per-outcome pricing on AI agents replacing labour. Sierra, Decagon, Cognition Devin and 11x.ai price closer to the cost of the human role being replaced than to traditional SaaS.
Custom model and fine-tuning services
One-off contracts for custom training, fine-tuning and dedicated deployments. Anthropic, OpenAI, Cohere and Mistral all run sizeable contract revenue tied to government, financial services, defence and healthcare customers.
AI software valuations in May 2026
Public AI software comps trade at 5.3x EV/Revenue. Median revenue multiple across AI software M&A deals was 4.8x in the last 12 months. Median revenue multiple across AI software VC rounds was 23x in the last 12 months.
5.3x
Median EV/Revenue as of May 2026 for public AI software companies
72x
Palantir is the highest valued public AI software company based on EV/Revenue (excluding outliers)
4.8x
Median EV/Revenue across AI software M&A deals in the last 12 months
23x
Median EV/Revenue across AI software VC rounds in the last 12 months
AI software market segments
AI software spans foundation model labs, vertical AI applications and AI coding assistants.
Foundation model labs
Labs training frontier general-purpose models and serving them via API and first-party apps. OpenAI raised $6.6B in October 2024 at a $157B post and a further round at $300B+ in 2025; Anthropic closed Series E at $61.5B in late 2024 and was reported to be raising at $170B+ in 2025; xAI raised $6B in 2024 at a $50B post; Mistral closed a $640M Series B in mid-2024 at a $6B valuation.
Consumer AI assistants
General-purpose chat and answer products sold direct to consumers. ChatGPT (OpenAI), Claude (Anthropic), Gemini (Google) and Perplexity define the category; Perplexity raised at $9B in late 2024 and again at $14B in early 2025.
Enterprise search and knowledge AI
Retrieval-augmented enterprise search across SaaS apps and internal documents. Glean reached over $100M ARR and raised at $7.25B in mid-2025; Hebbia, Vectara, Sana and Aisera compete in adjacent territory.
AI coding assistants
Tools generating and editing code inside IDEs and agents writing pull requests autonomously. GitHub Copilot leads incumbents; Cursor (Anysphere) reportedly hit $500M ARR in 2025 and raised at $9.5B; Cognition Labs (Devin), Codeium (Windsurf), Magic.dev and Replit Agent compete on autonomous coding agents; Codeium and Sourcegraph compete on enterprise deployments.
Vertical AI applications
Workflow products targeting specific functions and industries. Harvey leads legal and raised at $5B in 2025; Cresta and Sierra (Bret Taylor) lead customer support; Abridge and Ambience lead clinical documentation; EvenUp leads personal injury.
AI for sales, marketing, customer support and revenue operations
Agents and copilots automating outbound, support, marketing and revops workflows. Sierra (founded by Bret Taylor) raised at $4.5B in 2024; Decagon, 11x.ai, Clay and Jasper compete on agent-led outbound and support; Cresta and ASAPP compete in sales coaching; incumbents (HubSpot, Salesforce Agentforce, ServiceNow Now Assist, Microsoft Copilot) bundle aggressively.
AI agents and autonomous workflows
Agentic frameworks executing multi-step tasks across browsers, apps, APIs and operating systems. Adept (acquired by Amazon in 2024), Cognition Labs, MultiOn, Imbue and Reka AI compete; Anthropic's Claude Computer Use and OpenAI's Operator have reset the competitive frame.
AI safety, evaluation, observability and red-teaming
Evaluation, red-teaming and observability tooling for LLM applications. Scale AI dominates enterprise evals (Meta acquired a 49% stake at a $29B valuation in mid-2025); Arize, LangSmith (LangChain), Patronus AI and Braintrust lead developer-side tooling.
Fractional CFO, financial modelling and deal advice for AI software companies
See how Flow helps AI software founders.
We speak founders' language and have great operational understanding of AI software businesses.
Book an intro call - we'll look under the hood and recommend concrete next steps.
Fractional CFO
For founders who want to improve their FP&A functions, build an investor-ready financial model, and prepare for the next VC round.
Capital raising
For bootstrapped and already-VC-backed entrepreneurs who are looking to raise late stage venture or growth capital.
M&A
For category-leading technology companies who are exploring exit alternatives with either financial or strategic acquirers.
Key AI software KPIs to track
ARR, ACV, net revenue retention, gross margin and weekly active users are the metrics investors track in AI software.
| KPI | Definition |
|---|---|
| ARR | Annual recurring revenue. Standard for enterprise AI SaaS; harder to interpret for token-priced API businesses where consumption is the cleaner signal. |
| ACV | Average contract value. Frontier vertical AI (Harvey, Hebbia, Glean and Cresta) often runs $250K-$2M ACVs at enterprise; SMB AI products sit at $5K-$30K. |
| Net revenue retention | Existing-customer revenue 12 months later including expansion. Leading enterprise AI products are reporting NRR well above 130% as seat counts and per-seat consumption both grow. |
| Gross margin | Software gross margin net of inference cost. The single most-watched AI economics metric. Token-led businesses run at 40-60% gross margin; vertical AI apps with caching and prompt optimisation can reach 70-80%. |
| Weekly active users | WAU on consumer assistants. ChatGPT reportedly above 300M, Gemini above 350M and Claude above 30M as of 2025. The headline traction metric for the consumer tier. |
| Token volume | Daily or monthly tokens processed across API and first-party products. The cleanest read on real underlying activity at model labs. |
| Revenue per token | Realised revenue divided by tokens served. Falling sharply as model prices compress; useful for tracking pricing erosion at the API tier. |
| Compute spend ratio | Inference and training compute as a share of revenue. Anthropic and OpenAI have run this above 100% during heavy training years; trends down as inference economics improve. |
Main AI software players globally
The most active AI software companies and category leaders globally.
| Company | HQ | Overview |
|---|---|---|
OpenAI openai.com |  San Francisco | Maker of ChatGPT and the GPT model family. Reported $11B+ revenue run-rate by late 2025; raised $6.6B in October 2024 at a $157B post-money valuation, with a further round reported at $300B+ in 2025. Microsoft is the principal cloud partner and investor. |
Anthropic anthropic.com | San Francisco | Maker of the Claude model family. Closed Series E at $61.5B in late 2024; reported to be raising at $170B+ in 2025. Amazon committed $8B in cumulative investment by late 2024; Google is a meaningful additional investor. |
Google DeepMind deepmind.google |  London | Alphabet's combined AI research unit, formed by merging Google Brain and DeepMind in 2023. Ships Gemini across consumer (Gemini app), enterprise (Gemini Enterprise), developer (Gemini API) and on-device (Gemini Nano in Android and Pixel) surfaces. London HQ with major US research presence. |
Meta AI ai.meta.com | Menlo Park | Meta's in-house AI organisation; ships Llama open-weight models and the Meta AI assistant across Facebook, Instagram, WhatsApp and Ray-Ban Meta glasses. Acquired a 49% stake in Scale AI at a $29B valuation in mid-2025 and brought Alexandr Wang in to lead a new superintelligence unit. |
xAI x.ai | Palo Alto | Elon Musk's AI lab; ships the Grok model family and Grok consumer app integrated with X. Raised $6B Series B in 2024 at a $50B valuation, with further rounds reported at $75B+ in 2025. Operates the Colossus 100k-H100 supercluster in Memphis. |
Mistral AI mistral.ai |  Paris | European frontier model lab; ships Mistral, Mixtral, Codestral and Pixtral open-weight models alongside Le Chat. Raised $640M Series B in mid-2024 at a $6B valuation; principal European challenger to US labs with Microsoft, Nvidia, General Catalyst and Andreessen Horowitz on the cap table. |
Perplexity perplexity.ai | San Francisco | Answer-engine combining web search with frontier models. Raised at $9B valuation in late 2024 and again at $14B in early 2025; reported ARR above $50M by mid-2025. Backed by IVP, NEA, Bezos Expeditions and Nvidia. |
Glean glean.com | Palo Alto | Enterprise AI search and assistant across SaaS apps and internal documents. Raised $260M Series F in mid-2025 at $7.25B valuation; reported ARR above $100M. Backed by Kleiner Perkins, Sequoia, Lightspeed and Coatue. |
Harvey harvey.ai | San Francisco | Vertical AI for elite law firms; built on OpenAI models with custom legal training. Raised at $5B in mid-2025 with reported ARR above $75M. Deployed at Allen & Overy, PwC, A&O Shearman and majority of the AmLaw 100. |
Cursor (Anysphere) cursor.com | San Francisco | AI-native code editor; reportedly hit $500M ARR in 2025 with a $9.5B valuation round. Backed by Thrive Capital, A16z, Benchmark and Index Ventures; widely adopted by both startup and enterprise engineering teams. |
Raising your next round?
Fractional CFO services covering model, narrative and data room - end-to-end fundraise support.
Key AI software market trends
The shift from chat to agents, vertical AI displacing horizontal SaaS and open-weight models closing the gap are reshaping AI software right now.
Frontier model economics under pressure
API pricing for equivalent capability has fallen roughly 10x since GPT-4 launched. OpenAI, Anthropic, Google and Mistral are competing on price at the commodity tier while investing into agentic and reasoning capability at the frontier. Margins at the API layer are compressing even as absolute revenue grows.
Shift from chat to agents
Anthropic Claude Computer Use (October 2024), OpenAI Operator (January 2025), Google Project Mariner and Microsoft Copilot Actions have reset competitive dynamics from chat completion to multi-step autonomous tasks. Sierra, Decagon, Cognition Devin, 11x.ai and Adept (acquired by Amazon) are pricing closer to the cost of replaced labour than to per-seat SaaS.
Vertical AI displacing horizontal SaaS
Harvey (legal), Abridge (clinical documentation), Sierra (customer support), Cresta (sales coaching), EvenUp (personal injury) and Hebbia (financial research) are landing seven-figure ACVs by replacing workflow software plus the analyst hours that sit on top of it. Incumbents (Salesforce Agentforce, ServiceNow Now Assist) bundling aggressively in response.
Capital flight to model labs
OpenAI, Anthropic, xAI and Mistral raised at $300B, $170B, $75B+ and $6B respectively across 2024-25; combined fundraising at the four leading labs is comfortably above $50B. Round sizes have outgrown venture and now include sovereign capital (PIF, Mubadala, MGX and Qatar Investment Authority), hyperscaler equity, structured debt against compute commitments, and revenue-share deals with cloud platforms.
Inference cost decoupling from training
Distillation, quantisation, speculative decoding and inference-optimised silicon (Groq, Cerebras, SambaNova and Nvidia Blackwell) have made high-quality inference dramatically cheaper. Application companies with disciplined caching and routing are now reaching 70-80% gross margins despite token-priced upstream models.
Open-weight models closing the gap
Meta Llama, Mistral, DeepSeek and Qwen (Alibaba) are within months of frontier capability at zero licensing cost. DeepSeek-V3 and R1 in early 2025 reset perceptions of cost-efficient training. Enterprises deploying open-weight models on dedicated infrastructure for data residency and predictable cost are growing fast.
Similar verticals to AI software
Explore niches like AI infrastructure & semiconductors and AI consulting.
Explore other sectors
We know tech inside & out.
We live and breath tech - true understanding of how startups operate is fundamental at what we do.
Recent insights across AI software and beyond
Talk to us
Schedule a call to get a health check on your business and see how we could help.
Fractional CFO
- Fractional CFO for Software
- Fractional CFO for AI & ML
- Fractional CFO for Fintech
- Fractional CFO for Consumer internet
- Fractional CFO for Digital media
- Fractional CFO for E-commerce & marketplaces
- Fractional CFO for Consumer products
- Fractional CFO for Mobility
- Fractional CFO for Digital health
- Fractional CFO for Industrial technology
- Fractional CFO for Digital infrastructure
- Fractional CFO for IT services
Stages
Countries
- UK Fractional CFO
- Ireland Fractional CFO
- France Fractional CFO
- Germany Fractional CFO
- Spain Fractional CFO
- Portugal Fractional CFO
- Italy Fractional CFO
- Netherlands Fractional CFO
- Belgium Fractional CFO
- Switzerland Fractional CFO
- Austria Fractional CFO
- Denmark Fractional CFO
- Sweden Fractional CFO
- Norway Fractional CFO
- Finland Fractional CFO
- Poland Fractional CFO
- Estonia Fractional CFO
- US Fractional CFO
- Canada Fractional CFO
- Mexico Fractional CFO
- Brazil Fractional CFO
- UAE Fractional CFO
- Australia Fractional CFO
Cities
- London Fractional CFO
- Dublin Fractional CFO
- Paris Fractional CFO
- Berlin Fractional CFO
- Madrid Fractional CFO
- Lisbon Fractional CFO
- Milan Fractional CFO
- Amsterdam Fractional CFO
- Brussels Fractional CFO
- Zurich Fractional CFO
- Vienna Fractional CFO
- Copenhagen Fractional CFO
- Stockholm Fractional CFO
- Oslo Fractional CFO
- Helsinki Fractional CFO
- Warsaw Fractional CFO
- Tallinn Fractional CFO
- New York Fractional CFO
- Toronto Fractional CFO
- Mexico City Fractional CFO
- São Paulo Fractional CFO
- Dubai Fractional CFO
- Sydney Fractional CFO