CohereCohere builds enterprise large language models and AI agents focused on regulated industries and sovereign deployments.
AppleBuilds the Vision Pro spatial computer and visionOS — Apple's bet that head-mounted displays become the next personal computing platform. The M5 Vision Pro starts at $3,499 with 2x on-device AI inference performance.
LegoraLegora builds a collaborative AI platform that automates research, drafting, and review work for lawyers.
AnthropicCreates Claude — the frontier LLM family that powers Claude Code, the terminal-based coding agent that became the breakout developer tool of 2025.
Liquid AILiquid AI builds general-purpose foundation models based on liquid neural networks, an alternative to transformer architectures.
SierraSierra builds conversational AI agents for enterprise customer-experience workflows, replacing tier-1 support teams.
Cognition AICreator of Devin, the first autonomous AI software engineer that can plan, write, test, and deploy complete software projects from natural-language specifications.
Mistral AIMistral AI develops open-weight foundation models and conversational AI applications for enterprise reasoning, code generation, and sovereign AI deployment across multiple languages.
HarkHark builds multimodal AI models and AI-native hardware devices designed to be a universal interface between humans and machines.
PerplexityPerplexity is an AI answer engine that delivers cited, conversational answers and is expanding into an agentic browser, Comet.
HarveyHarvey builds domain-specific AI agents for law firms and corporate legal and professional-services teams.
MetaReleases the open-weight Llama model family including Code Llama, which enables on-premise and self-hosted AI coding tools for enterprises with data-residency requirements.
OpenAIBuilds frontier large language models including GPT and Codex that power the majority of AI coding tools via API, plus first-party products like ChatGPT and Codex CLI.
World LabsWorld Labs builds large world models that generate and reason about 3D spatial environments.
Google DeepMindGoogle DeepMind is an AI research lab, now a Google subsidiary, that builds Gemini multimodal models, the AlphaFold protein structure system, open Gemma models and SIMA 2 agents for developers, enterprises and researchers.
WriterWriter is a full-stack enterprise generative-AI platform built on its own family of Palmyra models.
Reflection AIReflection AI is a frontier lab founded by ex-DeepMind researchers building open large language models and autonomous coding agents.
Moonshot AIMoonshot AI is a Beijing foundation-model lab behind the Kimi chatbot and the open-weight Kimi K2 model family.
NablaNabla builds an ambient AI assistant that listens to clinical encounters and generates medical notes for clinicians.
RekaReka builds custom multimodal AI models and platforms for enterprises.
Thinking Machines LabThinking Machines Lab is an AI research startup founded by former OpenAI CTO Mira Murati building frontier models and customization tools.
MiniMaxMiniMax is a Shanghai-based foundation-model lab behind the ABAB language models and the Hailuo video and audio generators.
GensparkGenspark builds a general-purpose 'Super Agent' that turns natural-language prompts into completed multi-step tasks.
Zhipu AIZhipu AI is a Tsinghua-incubated Chinese large-model developer behind the GLM family and the Z.ai platform.
Sakana AISakana AI is a Tokyo-based research lab building nature-inspired, evolutionary AI foundation models.
ManusManus is a Singapore-headquartered developer of a general-purpose autonomous AI agent that executes multi-step tasks such as web browsing, code writing, and data analysis from natural-language instructions. It is offered as a subscription service.
DeepSeekDeepSeek builds open-weight AI models and reasoning systems for code generation and general language tasks, distributed globally without proprietary restrictions.
xAIxAI is Elon Musk's frontier AI lab, building the Grok family of large language models and the Colossus supercomputer.
DecartDecart builds real-time, interactive world models and the inference stack that runs them.
AMI LabsAMI Labs is Yann LeCun's world-model lab building JEPA-based systems that learn from real-world sensor data rather than language alone.
GleanGlean is an enterprise AI platform that combines connected work search with AI assistants and agents across a company's apps and data.
H CompanyH Company is a Paris-based AI lab building agentic foundation models and autonomous agents for enterprise and consumer automation.
StepFunStepFun is a Shanghai multimodal foundation-model lab founded by ex-Microsoft executive Jiang Daxin, behind the Step model family.
01.AI01.AI is Kai-Fu Lee's Beijing large-model startup behind the open Yi model family, now focused on commercial AI solutions.
AI21 LabsAI21 Labs is an Israeli foundation-model company building enterprise LLMs and AI systems, including the Jamba model family.
Baichuan IntelligenceBaichuan Intelligence is a Beijing large-model startup founded by ex-Sogou CEO Wang Xiaochuan, building the open Baichuan model series.
ByteDance SeedByteDance Seed is the AI research unit of ByteDance that develops Seed foundation models for language, image, and video generation. Its models power consumer and developer products for chat, creative, and marketing use.
HebbiaHebbia builds AI agents for knowledge work, used by finance and legal teams to analyze large document sets.
MagicMagic builds frontier code-generation models with ultra-long context windows aimed at automating software engineering.
MicrosoftMicrosoft ships Copilot and its own MAI models, makes VS Code and owns GitHub, runs one of the largest security businesses, and is OpenAI's largest partner.
NdeaNdea is an AI lab pursuing AGI through deep learning-guided program synthesis.
QwenQwen is a China-based AI lab, founded in 2023 in Hangzhou, that builds open-weight large language and image models, plus an open-source coding agent, for developers and enterprises.
Safe SuperintelligenceSafe Superintelligence (SSI) is a research lab founded by Ilya Sutskever building safe superintelligence as its sole product.
Hugging FaceOperates AI model hub and inference API hosting 500,000+ models including Flux, Stable Diffusion, and others.
InstinctInstinct builds a personal AI agent that people text or call to book travel, order groceries, buy tickets and make phone calls, using its own phone number and computer to finish each task.
SarvamSarvam is a Bengaluru AI lab that trains foundation models for Indian languages and sells voice agents, document AI and speech, translation and model APIs to businesses and government.
Tencent HunyuanTencent Hunyuan is Tencent's foundation-model team in Shenzhen, building the Hy family of open-weight language models and Hunyuan 3D and video generation models for Tencent products and developers.Venture capital deployed · 2026 YTD
Sector market cap
Largest raise · trailing 12mo
Catalysts ahead · next 12mo
Companies tracked
Nothing on the calendar yet.
Nothing on the calendar yet.
Mega-rounds ($1B+) dominate 2025–2026, with 11 rounds above $1B versus only two in 2023–2024. Median round size jumped to $300M+. Seed rounds exploded upward: AMI Labs $1B seed, Thinking Machines $2B seed. Smaller rounds ($50–200M) remain present but marginal to sector momentum.
Late-stage capital overwhelms the sector. Series D+ rounds and growth/late-stage rounds account for ~65% of total capital deployed in past 18 months. Early-stage ($100M Series A/B) still funded, but mega-rounds at Series C–E and growth stages capture 80%+ of headline activity. Seed tier bifurcated into tiny and stratospheric.
Nvidia, Andreessen Horowitz, and Sequoia Capital lead the most rounds (6–8 each across 2024–2026). New entrants Greenoaks, Mitsubishi UFJ Financial, and Autodesk emerged as round leads. Established LPs Accel and Khosla remain active. xAI, Safe Superintelligence, and Cohere show multiple consecutive rounds from same or overlapping backer clusters.
The AI agent security pause is a feature, not a bug—and investors should treat it as a market signal.
The AI safety reckoning is colliding with the open-weights movement—and investors are caught in the crossfire.
AI agents are winning in implementation, not just intelligence.
DeepSeek’s rise exposes a widening rift between sovereign AI models and the infrastructure they depend on.
As of 2026-08-01
