Most AI funding databases force every startup into a single category. A company building fraud detection for banks gets filed under “FinTech” and its AI capabilities vanish from infrastructure analysis. Bot Memo’s taxonomy system — 18 AI startup verticals, 104 cross-cutting tags, and 9 classification types — was built to solve that problem. It provides a structured set of ai industry categories purpose-built for tracking capital flows across the AI economy.
This article breaks down every component of the taxonomy, explains how it differs from legacy classification systems, and shows why multi-vertical attribution produces more accurate AI sector classification than the industry standard.
On this page
- What Are AI Startup Verticals and Why Do They Matter?
- Bot Memo's 18 AI Verticals: The Complete List
- 104 Tags Across 3 Categories: Infrastructure, AI Concepts, and Domain
- AI Native vs AI Augmented, AI Adjacent, AI Platforms: 4 Classification Types
- Why Multi-Vertical Attribution Makes the Numbers Add Up to More Than 100%
- How Bot Memo's Taxonomy Compares to Crunchbase, Dealroom, and PitchBook
- Frequently Asked Questions
What Are AI Startup Verticals and Why Do They Matter?
The industry vertical definition in an AI context is straightforward: a market segment defined by the primary industry a company serves. The vertical answers one question: who is the buyer?
“Health & Biotech” captures companies applying AI to drug discovery, diagnostics, and patient care. “Cybersecurity” captures those building AI-powered threat detection and network security.
Standard industry codes were not designed for this. The North American Industry Classification System (NAICS) assigns a single code per establishment based on primary activity.
A company like Xaira Therapeutics, which raised a $1 billion Series A in 2024 for AI-driven drug discovery, would land under a generic biotech NAICS code — erasing the AI dimension entirely. The Georgetown CSET research group documented this blind spot: NAICS assigns a single code per establishment based on primary activity, which cannot capture a company’s use of AI — a traditional pharmaceutical lab and one deploying AI for drug discovery receive the same code.
AI startup verticals exist because investors, founders, and analysts need to track capital flows within the AI economy, not just within healthcare or finance broadly. When Crunchbase reported that AI-related healthcare startups had pulled in $10.7 billion through late 2025 — already 24.4% higher than 2024 — that figure was only possible because their taxonomy separates AI healthcare from traditional healthcare. Bot Memo’s ai startup taxonomy does the same across all 18 verticals, with 13,532 deals tracked between 2023 and 2025.
Bot Memo’s 18 AI Verticals: The Complete List
Bot Memo’s taxonomy version 2.5.0 (released January 3, 2026) defines 18 ai industry categories. Each vertical represents a distinct buyer market where AI companies compete for revenue.
| # | Vertical | Definition | Example Deal |
|---|---|---|---|
| 1 | Cybersecurity | AI for threat detection, network security, identity management | Vega — $120M Series B for AI-powered enterprise threat detection (Feb 2026) |
| 2 | Developer Tools & AI Infrastructure | Tools for building, deploying, and monitoring AI systems | Cursor — AI-native code editor valued at $29.3B after raising $2.3B in 2025 |
| 3 | Education Technology | AI for adaptive learning, tutoring, institutional tools | AI tutoring platforms using reasoning models for personalized instruction |
| 4 | Energy & Sustainability | AI for renewable energy optimization, grid management, carbon tracking | AI-driven grid balancing for intermittent renewable sources |
| 5 | Enterprise Software | AI for internal business operations, workflow automation, ERP | AI copilots embedded in enterprise productivity suites |
| 6 | FinTech | AI for banking, payments, lending, credit scoring, fraud detection | AI-powered credit scoring replacing legacy FICO models |
| 7 | Food & AgriTech | AI for precision agriculture, food supply chain, crop monitoring | Computer vision for real-time crop disease detection |
| 8 | GovTech & Defense | AI for government operations, defense applications, public safety | Autonomous surveillance and document processing systems |
| 9 | HR Tech & People Analytics | AI for recruitment, workforce planning, employee analytics | Juicebox — $80M Series B for AI-native autonomous recruiting (March 2026) |
| 10 | Health & Biotech | AI for drug discovery, diagnostics, precision medicine, therapeutics | Xaira Therapeutics — $1B Series A for AI drug discovery |
| 11 | InsurTech | AI for underwriting, claims processing, risk assessment | Automated claims adjudication using computer vision |
| 12 | Legal Tech | AI for contract analysis, legal research, compliance automation | AI contract review reducing due diligence timelines from weeks to hours |
| 13 | Manufacturing & Industrials | AI for predictive maintenance, quality control, robotics | Digital twin platforms for factory optimization |
| 14 | Marketing & Sales Tech | AI for lead generation, content personalization, sales automation | WINN.AI — $18M Series A for real-time AI sales guidance |
| 15 | Media & Entertainment | AI for content generation, gaming, music, video production | Generative AI studios producing synthetic media at scale |
| 16 | Real Estate & Construction | AI for property valuation, construction management, spatial planning | AI-powered building information modeling (BIM) |
| 17 | Retail & E-Commerce | AI for demand forecasting, personalization, inventory optimization | AI recommendation engines driving conversion rate improvements |
| 18 | Transportation & Mobility | AI for autonomous vehicles, fleet management, logistics | Autonomous trucking companies targeting long-haul freight corridors |
These 18 ai startup verticals cover the full spectrum of the AI economy. The taxonomy uses a flat structure, meaning tags (covered below) are independent of verticals and can apply across any category.
104 Tags Across 3 Categories: Infrastructure, AI Concepts, and Domain
While verticals answer “who buys this?”, tags answer “what technology does this use?” and “what specific problem does this solve?” Bot Memo’s 104 ai startup tags are organized into three groups — forming a layered set of ai industry categories that cut across every vertical.
Infrastructure Tags (8)
Infrastructure tags identify the foundational technology layer a company operates on.
| Tag | Description |
|---|---|
| Blockchain | Distributed ledger technology for data integrity or tokenization |
| Cloud Computing | Cloud-native deployment and managed services |
| CI/CD | Continuous integration and deployment tooling for AI workflows |
| Edge AI | AI inference running on devices rather than centralized servers |
| GPU | Graphics processing unit infrastructure for training and inference |
| IoT | Internet of Things sensor networks generating AI training data |
| Quantum Computing | Quantum hardware or quantum-enhanced algorithms |
| Vector Database | Specialized databases for embedding storage and similarity search |
A company like Lambda, which raised $1.5 billion in its Series E (November 2025) for GPU cloud services, would carry the GPU and Cloud Computing ai infrastructure tags while sitting in the Developer Tools & AI Infrastructure vertical.
AI Concepts Tags (26)
These tags capture the AI methodology or paradigm a company uses. The full set of 26: AI Agents, AI Chips, AI Evaluation, AI Memory, AI Safety, Chip Making, Coding Agents, Continuous Learning, Generative AI, Human-in-the-Loop, Inference Optimization, Model Compression, Model Evaluation, Model Routing, Multimodal AI, Open Source, Privacy-Preserving AI, Reasoning Models, Robotics, Self-Improving, Small Language Models, Sovereign AI, Spatial Intelligence, Synthetic Data, Voice Agents, and World Models.
A vertical AI startup building autonomous sales agents would carry the AI Agents and Voice Agents concept tags, regardless of whether it sits in Marketing & Sales Tech or Enterprise Software. This flat structure means analysts can query “all companies using AI Agents” across every vertical — something hierarchical taxonomies cannot do.
Domain-Specific Tags (70)
Domain tags identify the business problem a company solves. Examples include: Drug Discovery, Fraud Detection, Autonomous Vehicles, Contract Analysis, Precision Medicine, Supply Chain Optimization, Threat Detection, Credit Scoring, Recruitment Automation, Predictive Maintenance, and 60 more.
WINN.AI, the Tel Aviv-based startup that raised $18M in Series A funding for AI-powered real-time sales coaching, carries the domain tags Sales Enablement and Sales Agents alongside the concept tag AI Agents. It sits in the Marketing & Sales Tech vertical.
Each layer adds analytical resolution without forcing a single-label constraint.
AI Native vs AI Augmented, AI Adjacent, AI Platforms: 4 Classification Types
Beyond verticals and tags, Bot Memo applies one of 9 ai company classification types to every deal. Understanding the distinction between ai native vs ai augmented is essential for accurate ai sector classification.
Four of these describe the company’s relationship to AI technology. The remaining five are filtering categories (Non-AI, Venture Firm, Acquisition, News Article, Insufficient Info) used to maintain data hygiene.
AI Native
AI is the product’s foundation. The company would not exist without artificial intelligence. Remove the AI and there is no product.
Example: Cursor is an AI-native code editor built by Anysphere. Its core product — predictive code completion, natural-language editing, and codebase-aware suggestions — requires large language models to function.
Without AI, Cursor is an empty text editor. The company reached a $29.3 billion valuation in November 2025 after closing a $2.3 billion Series D.
AI Augmented
An existing product category enhanced with AI features. The core product existed (or could exist) before AI was added.
Example: WINN.AI adds AI capabilities to established sales execution and enablement tools. CRM tools, call recording, and playbook management existed before AI.
WINN.AI augments them with real-time AI guidance during sales calls, automatically capturing key information and ensuring playbook adherence. The product is better with AI, but the category is not new.
AI Adjacent
The company enables AI development but does not run AI models itself. Infrastructure providers, data platforms, and tooling companies fall here.
Example: Lambda provides GPU cloud infrastructure for AI training and inference workloads. Lambda does not train or deploy its own models — it sells the compute that other companies use to do so. This makes it ai adjacent: essential to the AI ecosystem without being an AI product itself.
AI Platforms
Reserved for companies that train foundation models from scratch or design custom AI silicon. Bot Memo applies a strict 3-question litmus test:
- Does this company train its OWN foundation models from scratch?
- Does this company design custom silicon/chips specifically for AI workloads?
- Does this company operate GPU clusters sold specifically as AI training/inference compute?
A “yes” to question 1 or 2 qualifies the company as an AI Platform. OpenAI (GPT series), Anthropic (Claude), and Google DeepMind (Gemini) meet this threshold. Most companies that use foundation models via API do not — they are AI Native or AI Augmented.
This distinction matters for ai sector classification because AI Platforms represent a fundamentally different business model: capital expenditures measured in billions, research teams of hundreds, and competitive moats built on data and compute scale rather than domain expertise.
For a deeper breakdown of each classification with additional examples, see AI startup classifications explained.
Why Multi-Vertical Attribution Makes the Numbers Add Up to More Than 100%
A company building AI fraud detection for insurance companies touches two verticals: FinTech (fraud detection is a financial use case) and InsurTech (insurance is the buyer). Under single-label taxonomies, analysts must pick one. Under Bot Memo’s multi-vertical attribution system, the company’s full funding amount is attributed to both verticals.
This means vertical totals will sum to more than the dataset total — 30% more by design, not a data error.
Three Types of Multi-Attribution
Multi-Vertical: A company tagged with two or more verticals contributes its full funding amount to each. A $50M round for a company in both FinTech and InsurTech adds $50M to the FinTech total and $50M to the InsurTech total.
Multi-City: Companies headquartered in multiple cities (common for AI startups with distributed teams) contribute full funding to each city’s total. Bot Memo uses expand_multi_city_rows() rather than normalizing to a single metro, ensuring secondary hubs like Austin or Montreal are not undercounted.
Multi-Investor: Every investor in a syndicate receives full deal attribution. If five firms co-invest in a $100M round, each is credited with participating in $100M of deal activity. This is why investor totals always exceed the dataset total and why Bot Memo reports “participated in deals worth $X” rather than “invested $X.”
Multi-vertical attribution produces more accurate funding landscape analysis because it reflects how capital actually flows — across industry boundaries, not within neat single-label boxes. For the full methodology, see multi-attribution methodology.
How Bot Memo’s Taxonomy Compares to Crunchbase, Dealroom, and PitchBook
Each ai startup database uses different ai startup database categories and a different approach to ai startup taxonomy. Here is how the leading platforms compare across ai industry categories:
| Feature | Bot Memo | Crunchbase | Dealroom | PitchBook |
|---|---|---|---|---|
| AI-Specific Verticals | 18 verticals built for AI | 700+ industries (generic) | Innovation-focused tags | 50+ vertical codes (generic) |
| AI Classification Types | 4 types + litmus test | No AI-specific classification | AI/Deep Tech flag | No AI-specific classification |
| Tag System | 104 tags, 3 groups (flat) | Keyword-based categories | Innovation-based tagging | Sub-verticals |
| Multi-Vertical Attribution | Yes (full funding to each) | Single primary industry | Multiple tags allowed | Primary + secondary |
| Multi-Investor Attribution | Yes (full deal credit) | Per-investor tracking | Per-investor tracking | Lead vs. participant |
| Taxonomy Structure | Flat (tags cross verticals) | Hierarchical | Flat | Hierarchical |
| Update Frequency | Continuous (per pipeline run) | AI-validated daily | Weekly | Daily |
| AI Focus | AI-only dataset | All industries | All industries + innovation | All industries |
Crunchbase uses machine learning and AI to validate data across publications, but its 700+ industries are designed for the full economy, not specifically for ai startup verticals. Dealroom emphasizes startup-native taxonomy and innovation metrics, with particular strength in European ecosystems. PitchBook offers the deepest financial data but uses generic vertical codes inherited from traditional private equity classification.
Bot Memo’s advantage is specificity. By tracking only AI companies and applying a purpose-built ai startup taxonomy, every field — from classification type to domain tags — answers questions that AI investors actually ask: “How much capital went to AI-native cybersecurity startups using generative AI?” None of the general-purpose platforms can answer that query with a single filter.
Frequently Asked Questions
What are the main AI industry verticals?
Bot Memo tracks 18 ai startup verticals: Cybersecurity, Developer Tools & AI Infrastructure, Education Technology, Energy & Sustainability, Enterprise Software, FinTech, Food & AgriTech, GovTech & Defense, HR Tech & People Analytics, Health & Biotech, InsurTech, Legal Tech, Manufacturing & Industrials, Marketing & Sales Tech, Media & Entertainment, Real Estate & Construction, Retail & E-Commerce, and Transportation & Mobility. These verticals cover the full AI economy from AI startups by vertical with funding data.
How do you classify an AI startup?
Bot Memo uses a 9-type ai company classification system. The four core types — AI Native, AI Augmented, AI Adjacent, and AI Platforms — are defined by how central AI is to the product. AI Native means AI is the product. AI Augmented means AI enhances an existing product. AI Adjacent means the company enables AI without running models. AI Platforms means the company trains foundation models from scratch or designs AI chips. See AI startup classifications explained for the full framework.
How many types of AI startups are there?
Bot Memo’s ai startup taxonomy defines 9 classification types. Four describe a company’s relationship to AI technology: AI Native (AI is the product), AI Augmented (existing product enhanced with AI), AI Adjacent (enables AI without running models), and AI Platforms (trains foundation models or designs AI chips). The remaining five — Non-AI, Venture Firm, Acquisition, News Article, and Insufficient Info — are filtering categories used to maintain data hygiene and ensure only genuine AI companies appear in funding analysis.
What is an AI taxonomy?
An AI taxonomy is a structured classification system designed to categorize companies based on their relationship to artificial intelligence. Unlike generic industry codes (NAICS, SIC) that assign a single label per company, an ai startup taxonomy captures multiple dimensions: the industry vertical a company serves, the AI technologies it uses, and how central AI is to its product. Bot Memo’s taxonomy (version 2.5.0) combines 18 ai industry categories, 104 cross-cutting tags, and 9 classification types to provide that multi-dimensional view across 13,532 tracked deals.
What is the difference between horizontal AI and vertical AI?
Horizontal AI builds general-purpose tools that work across industries — think large language models or general image generators. Vertical AI builds specialized solutions for a single industry, such as AI for drug discovery in biotech or AI for threat detection in cybersecurity. Bessemer Venture Partners argues that vertical AI competes for labor budgets rather than IT budgets, enabling value capture far larger than generic horizontal platforms. Bot Memo’s 18 verticals track funding across vertical AI markets specifically.
What industries are most impacted by AI?
Based on Bot Memo’s dataset of 13,532 deals (2023-2025), the verticals attracting the most AI funding are Health & Biotech, Developer Tools & AI Infrastructure, Enterprise Software, FinTech, and Cybersecurity. AI cybersecurity startups raised record funding in 2025, with deal activity surging across the vertical. For current vertical breakdowns, see AI startups by vertical with funding data.
What does multi-vertical attribution mean in AI funding data?
Multi-vertical attribution means a company operating across multiple verticals has its full funding amount counted in each relevant vertical. A $50M round for a company in both FinTech and InsurTech adds $50M to both vertical totals. This causes vertical sums to exceed the dataset total by 30%, but it produces more accurate funding analysis than forcing every company into a single category. Bot Memo applies the same principle to multi-city and multi-investor data. See multi-attribution methodology for the complete explanation.
Bot Memo’s taxonomy (version 2.5.0) is updated continuously as new deals enter the pipeline. For definitions of funding types and stages or other technical terms, see the AI startup glossary.
About the author
Chintan Zalani
I'm the insight architect behind Bot Memo. I have spent the last decade building media assets on the internet. Bot Memo started as a simple project covering industry deep dives. Then I built a data pipeline for it. And now I love analyzing and covering all things AI startups and trends on top of our own data infrastructure.
