TechCrunch’s Startup Battlefield is still one of the toughest filters in startup land. Thousands apply every year. Only 200 make it into the Startup Battlefield 200 cohort. Of those, 20 get on the main stage to battle for the Startup Battlefield Cup and a $100,000 check.
But the other 180 aren’t consolation prizes. Within that group, the enterprise category quietly showed where B2B software and AI are really heading next.
Here’s a rundown of all 32 enterprise-tech startups selected for the 2025 Battlefield 200, and why they earned a slot.
Why this list matters
Enterprise doesn’t usually get the splashy product videos or viral moments. Yet it’s where most of the AI, infrastructure and workflow innovation actually lands first.
This year’s enterprise Battlefield picks cluster around a few big themes:
- AI agents and copilots inside core business workflows
- Trust, safety and compliance for AI (hallucination, deepfakes, privacy)
- Tools that turn old, messy enterprise systems into usable data
- Platforms that make niche but critical operations (like research facilities or hardware logistics) work like SaaS
Now, the full list.
The 32 enterprise tech selectees
AI Seer — Builds multi-model AI systems to uncover “untruths” and authenticate information, including a real-time AI fact-checker and a next‑gen polygraph-style device for validating claims.
Atlantix Portal — A platform for aspiring founders, grounded in a searchable database of 6,000+ university research innovations, with examples of pitches, business plans and launch materials.
Billow AI — Financial-operations tooling that automates manual workflows using a mix of AI models beyond just LLMs, including voice technology.
Blok — Lets product teams run user testing with synthetic users — AI agents that represent key segments — to get faster insight than classic A/B tests or feedback surveys.
Breakout — An inbound sales development AI agent that sits on your website and turns static pages into interactive, personalized experiences that can answer questions and make recommendations.
Cashew Research — A next-gen market research platform that helps marketers design plans and survey proprietary customer panels of real humans, not synthetic AI data.
CODA — Builds AI avatars that translate spoken and written language into sign language, bringing advanced machine learning to accessibility for the deaf and hard-of-hearing community.
Collabwriting — A web-highlighting and annotation tool for saving content and notes across apps, with AI features like fact-checking and “knowledge triggers” that resurface relevant saved info when you need it.
Dextego — Creates AI coaching agents based on behavioral intelligence data to help employees improve skills in areas like leadership, sales, motivation and role‑playing.
Dobs AI — AI agents that sift through large volumes of unstructured enterprise documents to extract information and analytics, while keeping all data under the customer’s control and out of external model training.
Elloe AI — Tackles AI hallucinations by real-time fact-checking model outputs, acting as an independent AI auditor that doesn’t rely on the same LLMs it oversees.
Elroi — Handles user permissions and supplies user-consented datasets for AI training that comply with shifting global privacy regulations.
Etiq AI — A data-science AI copilot that taps into data sources to power AI code generation and agentic workflows, with a strong focus on data context to avoid hallucinations.
GRAVL — A storefront platform for core research facilities, effectively a “Shopify for science” that gives labs and facilities the web presence and back-office IT they need to license innovations.
Hypercubic — Captures institutional knowledge around aging mainframe applications and uses AI to help document and debug the legacy code that still runs a huge share of enterprise workloads.
JustAI — AI agents for marketing that promise end‑to‑end execution, from designing a marketing plan through to running campaigns and analyzing results.
KrosAI — Voice AI agents, including phone-number provisioning, focused on around 50 emerging-market countries and use cases like call centers, optimized for ultra‑low latency.
Libertify — Turns written documents like PowerPoints and PDFs into secure interactive AI videos, where the “document” can explain itself and answer questions without exposing the underlying data.
Maisa — A platform for running enterprise AI agents as auditable digital co‑workers, designed to handle complex processes and produce clear trails of what each agent did.
Mappa — A recruiting platform that layers in behavioral voice analysis, using models trained to correlate voice patterns with traits such as communication style, empathy and confidence.
mAy-I — Computer-vision tools for retailers and other foot‑traffic businesses that capture visitor data like gender, age group and in‑store journey — giving physical spaces analytics similar to e‑commerce.
Mendo — A training tool that helps employees learn how to use a company’s generative AI stack, share time‑saving tips and evolve best practices so no worker is left behind.
Nimblemind — A platform for healthcare organizations to structure, label and manage multimodal clinical data for AI, with automation, audit trails and APIs to keep things faster and safer.
Plurall AI — Builds proprietary multimodal deepfake-detection technology from the ground up, rather than bolting on a detector to an existing model.
PRVIEW — A workflow platform for PR teams that automates tracking of speaking and award programs, replacing the giant spreadsheet many comms teams use to chase opportunities.
Rayda — Helps IT teams equip and manage remote workers in more than 170 countries, from configuring and shipping devices to tracking hardware, offboarding and recycling.
Sponstar — A platform for marketers to turn any event or city into a treasure hunt, complete with quests, rewards and Pokémon Go–style experiences.
Unthread — A help desk built natively around Slack, watching conversations to capture issues early and track them without bolting Slack on as an afterthought.
Visualsyn — A platform for businesses to create, edit and publish immersive 3D and XR content to the web, including mobile capture powered by on‑device AI processing.
WeShop AI — An AI video agent that generates professional, shoot‑quality product photos from a prompt, aimed at e‑commerce players and influencers.
ZETIC.ai — A developer tool for deploying real‑time AI directly on users’ devices, letting AI app builders decouple growth from runaway cloud‑AI costs.
Zinnia — An AI-powered platform that gathers data and supports salespeople as an assistant, designed to boost human productivity rather than replace reps outright.
Looking ahead to Disrupt 2026
Startup Battlefield is a snapshot of where founders and VCs think the market is going. In 2025’s enterprise cohort, that means AI agents everywhere, but wrapped in compliance, explainability and hard-nosed business workflows.
TechCrunch Disrupt 2026 lands in San Francisco from October 13–15, with a 200‑plus‑session agenda and hundreds of exhibiting startups already in the pipeline. If this year’s enterprise picks are any guide, expect even more agent-first, AI-native tooling — and a lot more competition for that $100,000 cup.
