All funding news

AI funding news

100 recent AI rounds across our tracked sources.

P
🇸🇬PixVerseAI Video Generation

PixVerse builds AI video generation tools for creators and game developers to produce videos and interactive worlds from text prompts and images.

$439MSeries C
Investor undisclosed
A $439M Series C for an AI video gen tool signals the category has moved past hype into infrastructure—this isn't a bet on whether text-to-video works, it's capital flooding toward whoever owns the creator workflow. PixVerse is likely using this to build out API/platform layers and lock in game dev partnerships before the space consolidates. If you're building any creator tool (design, music, 3D), watch how they're positioning around game engines and interactive content—that's where the defensibility actually lives.
N
🇺🇸Nous ResearchAgent AI

Nous Research builds Hermes, an open-source AI agent platform with web search, coding, and vision capabilities that runs locally or in the cloud.

$75MSeries B
R
Robot Ventures
USV logo
A $75M Series B for an open-source agent platform signals that the market is betting hard on *local-first* AI inference as a defensible moat—not just API-dependent agents. If you're building in adjacent agent spaces (automation, workflow, specialized reasoning), this validates that investors see real differentiation in owning the model layer, not just the orchestration layer. Watch how they monetize: open-source + cloud hosting is the play, which means your TAM might be bigger than you think if you can own a piece of the inference stack.
P
🇳🇱PromptwatchAI Search Optimization

Promptwatch helps companies optimize their visibility in AI chatbots like ChatGPT and Claude through specialized content and search optimization.

$6.5MSeed
S
Seed + Speed Ventures
BA
A $6.5M seed for AI search optimization signals that founders are already treating ChatGPT/Claude discoverability as a real distribution channel worth paying for—not a nice-to-have. If you're building B2B SaaS or content products, this matters: the SEO playbook you know is fragmenting, and visibility in LLM outputs is becoming table stakes. Watch whether Promptwatch's customers are mostly migrating SEO budgets or spending net-new—that tells you if this is a category shift or just reallocation.
Helsing logo
🇩🇪HelsingAI Defense Technology

Helsing builds AI-powered defense technology for European militaries to enhance operational capabilities and decision-making.

$1.8B
G
Goldman Sachs
Lightspeed logo
A $1.8B round for a European defense AI startup signals that Western governments are moving past procurement inertia—they're actually writing checks for AI-native military tech, not just bolting AI onto legacy systems. If you're building B2G infrastructure software anywhere (supply chain, logistics, compliance), watch how Helsing structures its go-to-market with NATO allies; that playbook is about to get copied across every allied government's tech spending.
Helsing logo
🇩🇪HelsingAI Defense Technology

Helsing builds AI-powered defense technology for European militaries to enhance operational capabilities and decision-making.

$1.8BSeries B
Investor undisclosed
An $1.8B Series B for defense AI signals European governments are moving past procurement pilots into real deployment—this isn't theoretical anymore. If you're building B2G infrastructure software (logistics, supply chain, compliance), watch how Helsing structures its go-to-market with defense ministries; the sales cycles and contract mechanics will likely become the template for other critical infrastructure plays in Europe over the next 18 months.
A
🇨🇳AISHI TechnologyAI

AISHI Technology builds AI solutions for Chinese enterprises and developers.

$41.7MSeries C
Investor undisclosed
A $41.7M Series C for enterprise AI in China signals that the market has moved past chatbot novelty—this is about embedding AI into actual workflows at scale. At this stage and size, AISHI is likely burning cash on sales infrastructure and model fine-tuning for specific verticals (manufacturing, finance, logistics are the usual suspects). If you're building AI tooling for any regulated industry, watch how AISHI navigates compliance and data residency; Chinese enterprises have different requirements than Western ones, and their playbook could preview what you'll face in other geographies.
珞
🇨🇳

珞博智能

AI Toys

珞博智能builds AI-powered collectible toys for Gen Z consumers using interactive intelligence.

$14MPre-A
Investor undisclosed
A $14M pre-A for AI-powered collectibles signals China's consumer AI market is moving past chatbots into hardware/physical goods—this is where the real engagement loop lives. The ticket size suggests investors believe the unit economics work (likely high margins on digital-native toy mechanics), which matters if you're building any Gen Z consumer product with recurring engagement. If you're in gaming, social, or even education tech, watch how they monetize the AI layer—it's probably not just the toy, it's the companion app ecosystem.
幻
🇨🇳

幻码跃迁

Quantum Computing

幻码跃迁 builds quantum computing infrastructure that integrates quantum processors with AI applications.

UndisclosedSeed
Investor undisclosed
Z
🇨🇳ZujidynamicRobotics

Zujidynamic builds humanoid robots for industrial and commercial applications.

$200MPre-IPO
Investor undisclosed
A $200M pre-IPO round for a Chinese humanoid robot company signals that industrial robotics is moving past hype into deployment—someone's betting on near-term revenue, not just R&D. If you're building in adjacent hardware (vision systems, motion control, industrial software), this validates that customers are actually buying and integrating these systems now, which means your TAM just got real.
海
🇨🇳海艺AIAI

海艺AI builds AI-powered creative tools for digital content creators and designers.

$14MSeries B
Investor undisclosed
A $14M Series B for a Chinese AI creative tools startup signals that the market is still hungry for vertical AI applications—especially in design/content where creators have real willingness to pay. The round size suggests they're past product-market fit and moving into distribution/localization; expect them to build out team and expand use cases (video, 3D, etc.). If you're building in adjacent creator tools, watch how they're monetizing—Chinese SaaS often moves faster on freemium-to-paid conversion than Western equivalents, which could be a playbook.
V
🇨🇳Volant AerotechAerospace/eVTOL

Volant Aerotech builds commercial eVTOL aircraft for urban air mobility using compound wing technology.

$4.2BSeries C+
Investor undisclosed
A $420M Series C+ for a Chinese eVTOL startup signals that the category has moved past tech risk and into manufacturing/certification risk—investors are betting on execution at scale, not concept validation. This capital likely funds production ramp, regulatory pathway in China, and probably some international certification work. If you're building hardware in aerospace or adjacent regulated industries, watch how Volant navigates China's certification timeline; it'll be a template for how fast non-Western aerospace can move.
鹰
🇨🇳

鹰瞰智翼

RoboticsVerified

鹰瞰智翼 builds bionic flapping-wing flying robots for consumers using embodied AI and fluid simulation technology.

UndisclosedSeries A
元
元禾璞华
孚厚
Chinese VCs are still writing Series A checks in July 2026, which suggests the domestic tech market hasn't frozen—but the investor syndicate (元禾璞华 leading with two co-investors) hints at a smaller round than typical US Series A, likely $5-15M. If you're building B2B infrastructure or enterprise software in Asia, this signals that capital is flowing to founders who've already hit product-market fit with local customers, not just the hype cycle.
M
🇨🇳MiniMaxLLM

MiniMax builds consumer AI products and an open LLM platform for Chinese users.

UndisclosedStrategic
Investor undisclosed
Sovereign wealth funds + Chinese institutions moving into a strategic round signals patient capital is betting on this space's 5-10 year thesis—likely infrastructure or deep tech that needs geopolitical hedging. If you're building in adjacent verticals (energy, semiconductors, materials), watch whether this company's cap table becomes a template for how non-traditional LPs are now co-investing with traditional VCs on later-stage bets.
L
🇨🇳logicalqubitQuantum Computing

LogicalQubit provides cloud-based quantum computing access with physical and logical qubit capabilities for enterprises.

UndisclosedSeries A
G
Ge Weidong Hunton Investment
K
This is a Series A from relatively under-the-radar China-focused investors in mid-2026—not a signal of major category heat, more likely a regional play or niche vertical getting patient capital. The timing and investor profile suggest founders should watch if this company is solving something specific to Chinese market dynamics (regulatory, infrastructure, or consumer behavior) that might have parallels in other regulated markets. If you're building in fintech, logistics, or enterprise software for Asia, this round's quiet nature is actually the tell—big money is flowing to specific verticals without the usual hype cycle.
M
🇨🇳MiniMaxLLM

MiniMax builds consumer AI products and an open LLM platform for Chinese users.

Undisclosed
Investor undisclosed
L
🇨🇳Lingsi TechnologyAI

Lingsi Technology builds AI solutions for enterprise customers in China.

$70MSeries B
Investor undisclosed
A $70M Series B for enterprise AI in China signals that the market is past the chatbot phase—this is about vertical-specific automation with real ROI, not consumer toys. At this stage and size, Lingsi is likely building out sales/implementation teams and deepening their product moat in specific verticals (manufacturing, logistics, finance are common bets). If you're building B2B AI for regulated industries or complex workflows, watch how they navigate China's AI governance—their playbook on compliance and customer trust will matter more than their tech stack.
X
🇨🇳Xspark AIPhysical AIVerified

Xspark AI builds physical AI systems for autonomous robotics and industrial automation with trustworthy, scalable deployment.

$14MAngel
鼎
鼎晖VGC
初S深
+1
A $14M angel round for physical AI in China signals the market is past proof-of-concept—investors are betting on deployment at scale, not just research. At this stage and size, Xspark is likely burning cash on hardware iteration, real-world testing, and building the ops/safety infrastructure that actually matters for industrial customers. If you're building in robotics, autonomous systems, or even software that touches the physical world, watch how they solve the "trustworthy deployment" problem—that's the unsexy moat that wins these markets, not the model.
聆
🇨🇳

聆思科技

AI ChipVerified

聆思科技 builds edge AI inference chips for perception and cognitive models.

$70MSeries B
安
安徽省国资平台
合深天
+4
A $70M Series B for edge AI inference chips, led by regional Chinese state capital, signals the mainland is doubling down on domestic chip sovereignty for on-device AI—this isn't about competing globally yet, it's about reducing reliance on Nvidia for inference workloads. If you're building multimodal or vision models, watch whether 聆思 can actually ship volume; their funding mix (state + commercial VCs) suggests they have distribution channels but need to prove silicon-to-production velocity.
D
🇨🇳DeepCtrlsPhysical AI, Industrial AIVerified

DeepCtrls builds physics-AI hybrid models for autonomous control of electromechanical systems in industrial and commercial facilities.

UndisclosedSeries B
J
Jinko Energy
SCSequoia China logo
+3
A Chinese solar manufacturer raising Series B with state-backed capital (SDIC) + top-tier VCs signals the government is actively consolidating the supply chain—this isn't just market heat, it's industrial policy. If you're in energy hardware or climate tech, watch whether this company pivots downstream (storage, grid software) or stays pure manufacturing; that tells you where Beijing thinks the margin is moving.
M
🇨🇳Moqi IntelligenceAI

Moqi Intelligence builds AI-powered enterprise software for Chinese businesses to automate complex workflows and decision-making.

$140MAngel
Investor undisclosed
A $140M angel round for workflow automation in China signals that enterprise AI buyers there are willing to write massive checks early—likely because labor costs and process complexity make automation ROI obvious. Moqi's probably burning this on sales infrastructure and model fine-tuning for vertical use cases rather than R&D. If you're building B2B automation anywhere, watch whether they go horizontal (competing with UiPath) or stay vertical—that'll tell you if the market rewards generalists or specialists right now.
Ollama logo
OllamaLLM Infra

Ollama provides open-source infrastructure for running large language models locally on consumer hardware.

$65MSeries B
Theory Ventures logo
Theory Ventures
Local LLM inference is moving from hobbyist curiosity to infrastructure bet—a $65M Series B signals VCs believe there's real enterprise demand for on-device model serving, not just cloud API consumption. If you're building anything that needs low-latency inference, offline capability, or data residency guarantees, this validates that the tooling layer is maturing enough to be a real dependency rather than a workaround.
瑞
🇨🇳

瑞启深空

Commercial Space + Remote Sensing

瑞启深空 builds an AI-powered remote sensing satellite constellation for commercial Earth observation and data analytics.

$30.8MAngel
Investor undisclosed
A $30.8M angel round for a satellite constellation is unusually large and signals China's serious push to commercialize Earth observation—this isn't a tech demo, it's capital-intensive infrastructure play. They're likely burning this on satellite manufacturing, launch costs, and ground station ops, which means the market's moved past "can we build this" to "who captures the data monopoly." If you're building any location-intelligence or geospatial product, watch whether Chinese satellite data becomes cheaper/faster than Western alternatives—it'll reshape your data sourcing economics within 18 months.
灵
🇨🇳

灵睿智芯

AI

灵睿智芯 develops AI chips and semiconductor solutions for intelligent computing applications.

Undisclosed
Investor undisclosed
D
🇨🇳DeepCtrlsPhysical AI, Industrial AI

DeepCtrls builds physics-AI hybrid models for autonomous control of electromechanical systems in industrial and commercial facilities.

Undisclosed
J
Jinko Energy
S
P
ProceptionRobotics

Proception builds AI-powered robotic manipulation systems with evolutionary learning for precise, adaptive task execution.

$11MSeed
Investor undisclosed
An $11M seed for a robotics company signals investors are betting on embodied AI moving past simulation—they want real-world task learning, not just language models. If you're building in manufacturing automation, supply chain, or even autonomous systems, this validates that the bottleneck isn't capital anymore; it's solving the sim-to-real gap and proving repeatability at scale.
中
🇨🇳

中科新松

AI

中科新松develops AI-powered robotics and automation solutions for industrial manufacturing.

UndisclosedSeries A+
Investor undisclosed
Lovable logo
🇸🇪LovableDeveloper Tools

Lovable builds an AI coding assistant that lets developers build software through natural language and visual feedback.

$300M
Investor undisclosed
A $300M round for an AI coding assistant signals that the market is betting hard on LLMs actually shipping production code—not just drafting it. At this stage and size, Lovable is likely burning cash on compute, hiring senior engineers to improve code quality/safety, and building enterprise sales infrastructure. If you're building any developer tool, watch whether they can actually reduce churn; the graveyard of AI coding assistants is full of tools with great demos but users who bounce after the novelty wears off.
G
🇫🇷GradiumVoice AIVerified

Gradium builds voice AI tools and models for developers to create real-time speech applications with translation and edge optimization.

$30MSeed
Nvidia logo
Nvidia
FED
+4
Nvidia leading a $30M seed is a rare signal—they're not just investing, they're signaling that real-time voice AI at the edge is infrastructure-grade, not a feature. For founders in adjacent speech/audio spaces, this validates that the bottleneck isn't model quality anymore; it's latency and on-device inference, which means your moat is probably in optimization and deployment, not training data.
SambaNova Systems logo
🇺🇸SambaNova SystemsAI Chip

SambaNova builds AI inference chips and systems for enterprises running large language models on-premises or in the cloud.

$1BSeries F
G
General Atlantic
IST
+11
A $1B Series F for inference chips signals that enterprises are actually moving LLMs on-prem—not just talking about it—and that the market believes there's real margin in custom silicon vs. renting GPU capacity. SambaNova's getting this kind of capital because inference is becoming the recurring revenue play (models are trained once, inference runs forever), which means if you're building enterprise AI tooling, you should assume your customers will eventually want to run models locally for cost/latency/compliance reasons.
Z
🇨🇳Zhipu AILLM

Zhipu AI builds large language models and AI technology for enterprise and consumer applications in China.

$4BIPO
Investor undisclosed
A $4B IPO for a Chinese LLM player signals that Beijing's willing to let domestic AI companies scale past the US export controls—this is less about Zhipu's tech and more about geopolitical capital allocation. If you're building AI infrastructure or applications outside the US-China duopoly, watch whether this unlocks funding for non-US alternatives; it might mean VCs are hedging against US regulatory risk by backing regional players.
SambaNova logo
🇺🇸SambaNovaAI Chip Hardware

SambaNova builds specialized AI processors optimized for machine learning workloads.

$1B
Investor undisclosed
A $1B raise for an AI chip company in mid-2026 signals that specialized inference hardware is still seen as defensible against Nvidia's dominance—likely because customers want alternatives for specific workloads (language models, recommendation systems) where custom silicon can cut costs or latency. SambaNova probably uses this to scale manufacturing partnerships and land enterprise design wins before the window closes. If you're building AI infrastructure (vector DBs, model serving, optimization layers), watch whether SambaNova's customers actually adopt at scale—that tells you if the chip differentiation story holds or if Nvidia's software moat wins anyway.
B
🇨🇳BrainCoBrain-Computer Interface

BrainCo develops brain-computer interface technology that decodes neural signals for medical and commercial applications.

$2.8BPre-IPO
Investor undisclosed
A $2.8B pre-IPO round for a Chinese BCI company signals that neural decoding has moved from research curiosity to capital-scale commercialization—likely driven by regulatory clarity in China and real clinical/consumer traction. At this stage and size, they're probably scaling manufacturing, building out clinical evidence for specific indications (motor control, communication), and preparing for public markets. If you're in adjacent neurotech, assistive devices, or even consumer health hardware, watch their go-to-market playbook closely—this validates that the regulatory and reimbursement path exists, even if it's China-first.
S
🇨🇳Shengshi WeiseiRobotics

Shengshi Weisei builds embodied AI brain models for autonomous welding robots in industrial manufacturing.

UndisclosedSeries B
Investor undisclosed
H
🇺🇰HiveRobotics

Hive builds autonomous control software for industrial machines like diggers and warehouse equipment using proprietary AI models.

$15MPre-Series A
S
SuperSeed
VSN
A $15M pre-Series A for industrial autonomy signals that investors are past the 'can we do this?' phase and betting on execution at scale—this is real money for real problems, not research. Hive's likely burning this on: (1) expanding their AI model training pipeline with more machine types, (2) field deployment and customer support to prove repeatability beyond one use case, and (3) hiring ops/sales to move from pilot to production contracts. If you're building in adjacent robotics or industrial software, watch whether they can actually ship repeatably—that's the moat everyone claims but few prove.
量
🇨🇳

量感智能

AI

量感智能 builds AI solutions for enterprise applications in China.

UndisclosedAngel
Investor undisclosed
Norm logo
🇺🇸NormLegal AI

Norm builds an AI-native law firm for enterprises, using supervised AI agents to deliver legal services with outcome-based pricing instead of hourly billing.

$120MSeries C
Khosla Ventures logo
Khosla Ventures
BCCoatue logo
+6
A $120M Series C for legal AI signals that outcome-based pricing models are now fundable at scale—VCs are betting enterprises will pay for results, not hours. Norm's mix of institutional LPs (Vanguard, TIAA, New York Life) suggests this isn't hype; it's capital treating legal automation as infrastructure. If you're building any B2B service with opaque pricing or time-based billing, this round is a reminder that your unit economics might be the real blocker, not your product.
M
🇮🇳MowitoRobotics

Mowito enables industrial robot arms to learn manufacturing tasks from operators without coding using AI.

$3MPre-Seed
V
Version One Ventures
AUI
+4
A $3M pre-seed for no-code robot programming signals investors believe the bottleneck in manufacturing automation isn't hardware—it's the skill gap to deploy it. If you're building any kind of operator-facing interface (whether for drones, CNC, or warehouse systems), this validates that learning-from-demonstration beats traditional programming as a go-to-market wedge, especially in price-sensitive markets like India where labor costs make automation ROI-sensitive but engineering talent is scarce.
L
🇺🇰Luffy AIFinancial Services AI

Luffy AI builds AI-powered financial services tools for enterprises and financial institutions.

$10.3MSeries A
Investor undisclosed
A $10.3M Series A for enterprise fintech AI in mid-2026 suggests the market has moved past chatbot novelty—institutions are now willing to fund AI that touches actual financial workflows and compliance. If you're building B2B AI in regulated spaces, this validates that investors see defensibility in domain-specific models over generic LLMs, and that enterprise sales cycles are finally closing at scale.
N
🇨🇳NICEBBConsumer AI

NICEBB builds an AI-powered recording device for parents to track and understand their children's development and behavior.

UndisclosedAngel
Investor undisclosed
Norm AI logo
Norm AILegal AI

Norm AI builds AI solutions for legal professionals to automate document review, research, and contract analysis.

$120MSeries C
Khosla Ventures logo
Khosla Ventures
A $120M Series C for legal AI signals that enterprise software buyers are finally willing to pay for AI that reduces billable hours—the unit economics work when you're displacing expensive knowledge work. If you're building AI for any professional services (accounting, consulting, architecture), this validates that the buyer will fund adoption if you can credibly show ROI on labor cost, not just efficiency theater.
Norm AI logo
Norm AILegal AI

Norm AI builds AI solutions for legal professionals to automate document review, research, and contract analysis.

$120M
Investor undisclosed
A $120M raise for legal AI in mid-2026 signals that document-heavy verticals are finally getting serious capital—the ROI math on automation is working. Norm's likely burning this on: (1) expanding their model's domain specificity beyond contracts into litigation/compliance, (2) enterprise sales/implementation, and (3) staying ahead of OpenAI's own legal product roadmap. If you're building in accounting, insurance claims, or healthcare records—spaces with similar document density and liability concerns—watch how Norm structures their go-to-market; their playbook for selling to risk-averse professionals will matter more than their tech.
E
🇨🇳Even RealitiesAI

Even Realities builds AI-powered solutions for Chinese enterprises.

UndisclosedPre-B
Investor undisclosed
W
🇨🇳Wujie ZhihangAutonomous Systems

Wujie Zhihang develops autonomous systems software for industrial and logistics applications in China.

$706KAngel
Investor undisclosed
A $700K angel for autonomous logistics in China signals the market is still in founder-validation mode—not yet at the Series A scale where you'd see $5M+ checks. If you're building autonomous systems anywhere, watch whether Wujie's next round comes from tier-1 Chinese VCs (Baidu's presence here matters) or stays angel-only; that'll tell you if the unit economics actually work or if it's still a technical proof-of-concept problem.
S
🇺🇸SyntiantAI Software

Syntiant builds AI and machine learning software solutions backed by Intel.

UndisclosedIPO
Investor undisclosed
E
🇨🇳Even RealitiesAI

Even Realities builds AI-powered solutions for Chinese enterprises.

UndisclosedPre-Series B
Investor undisclosed
华
🇨🇳

华源智因

AI Drug Discovery

华源智因 uses AI to model human cells and predict drug efficacy for pharmaceutical development.

UndisclosedSeed
Investor undisclosed
A
🇨🇳AI导出鸭Developer Tools

AI导出鸭 solves text encoding issues when exporting AI-generated content across formats.

UndisclosedAngel
Investor undisclosed
智平方 logo
🇨🇳

智平方

AI

智平方 builds AI-powered logistics optimization software for supply chain enterprises in China.

$700M
Investor undisclosed
A $700M round for logistics optimization in China signals that enterprise AI is moving past pilots into infrastructure—this isn't venture-scale anymore, it's growth equity or strategic capital betting on supply chain digitization as table stakes. If you're building B2B SaaS in regulated verticals (fintech, manufacturing, energy), watch how 智平方 navigates compliance and integration complexity; that playbook matters more than the dollar amount.
忆
🇭🇰

忆生科技

Robotics

忆生科技 builds memory systems for robots using AI, enabling machines to learn and retain experiences.

UndisclosedAngel
Investor undisclosed
C
🇨🇳Chang Guang SatelliteSatellite Imaging

Chang Guang Satellite operates the Jilin-1 constellation to provide high-resolution Earth observation and geospatial intelligence data.

$700M
长
长发集团
陆诚东
+2
A $700M round for a satellite constellation operator signals China's doubling down on sovereign geospatial intelligence—this isn't venture capital, it's state-backed infrastructure play. If you're building location/mapping products or defense-adjacent geospatial tools, watch whether Jilin-1 data becomes a commodity input (cheaper, more accessible) or stays locked behind Chinese state channels; that determines if you need your own satellite layer or can build on top of existing constellations.
Etched logo
🇺🇸EtchedChip Design

Etched designs custom AI chips manufactured by TSMC to compete with Nvidia's offerings.

$120MSeries A
Investor undisclosed
A $120M Series A for custom AI chips signals that the market believes there's real margin to capture by going around Nvidia's software lock-in—but only if you can actually ship and prove performance parity. Etched is likely burning this on tape-outs, TSMC relationships, and hiring deep chip talent; the bet is that custom silicon for specific workloads (inference, training, etc.) can undercut Nvidia's pricing enough to matter. If you're building AI infrastructure or ops tooling, watch whether Etched's chips actually get adopted at scale—if they do, your entire go-to-market changes because suddenly your customers have a real alternative to the Nvidia moat.
光
🇨🇳

光象科技

Robotics

光象科技 builds embodied AI systems for robotics applications using physical intelligence.

UndisclosedAngel
Investor undisclosed
G
🇨🇳Gestalt TechnologyAI

Gestalt Technology builds AI solutions for enterprise customers in China.

$58.8MAngel+
Investor undisclosed
A $58.8M angel-plus round for an enterprise AI company in China signals that late-stage angel syndicates are now comfortable writing checks at Series A scale—likely because China's AI infrastructure play is maturing past hype. Gestalt is probably burning this on sales/implementation teams and model fine-tuning for specific verticals (manufacturing, logistics, finance), not R&D. If you're building B2B AI for regulated industries anywhere, watch how they navigate compliance and customer concentration; China's playbook often previews what Western enterprises will demand in 18 months.
Kling AI logo
🇨🇳Kling AIVideo Generation

Kling AI generates high-quality videos from text and images using advanced AI models.

$3B
Investor undisclosed
A $3B round for a video generation model signals that Chinese AI labs are now competing directly with Western incumbents on scale and capital—this isn't a Series A, it's a war chest. Kling is likely burning this on compute infrastructure and dataset acquisition to close the quality gap with OpenAI's Sora and Runway, which means the video generation space just got a lot more expensive to enter. If you're building in multimodal AI or content creation tools, watch whether Kling's output quality actually justifies the spend—if it does, your moat just got thinner.
Unitree logo
🇨🇳UnitreeRobotics

Unitree builds humanoid robots powered by physical AI for industrial and commercial applications.

$619MIPO
Investor undisclosed
A Chinese robotics company hitting $619M in IPO valuation signals that physical AI—not just LLMs—is now fundable at scale, especially outside the US. Unitree likely uses this capital to scale manufacturing, build out the sales/support infrastructure for enterprise deployments, and fund the next generation of actuators/sensors. If you're building in robotics, autonomous systems, or even enterprise software, watch how they allocate this: the bottleneck isn't usually the AI model anymore, it's getting hardware into production and keeping customers happy.
V
Venice AIAI

Venice AI builds an open-source AI platform for developers to build and deploy applications.

$65MSeries A
Investor undisclosed
A $65M Series A for an open-source AI platform signals that the market is still hungry for developer infrastructure plays, even as the space gets crowded—investors are betting on open-source as a defensible moat against closed API lock-in. If you're building AI tooling, this validates that developers will pay for deployment/ops simplicity over raw model access, so your GTM should emphasize workflow integration rather than model novelty.
K
🇨🇳Kunlun ChipAI Chip Design

Kunlun Chip designs and manufactures AI training and inference processors for large-scale deployments across China.

$50BIPO
Investor undisclosed
A $50B IPO for a Chinese AI chip maker signals that domestic semiconductor alternatives to Nvidia are now production-ready at scale—this isn't speculative anymore. If you're building inference or training infrastructure anywhere, the cost structure just shifted: Chinese cloud providers will have a real cost advantage within 18 months, which means your pricing model needs to account for that competitive pressure. Watch whether their margins hold or compress; that tells you if the real bottleneck is fab capacity or design.
脸
🇨🇳

脸谱心智

World Models

脸谱心智 builds world models technology for AI systems to understand and simulate physical environments.

Undisclosed
Investor undisclosed
Kling AI logo
🇨🇳Kling AIVideo Generation

Kling AI generates high-quality videos from text and images using advanced AI models.

$30B
Investor undisclosed
A $30B valuation for a text-to-video model signals that video generation is now treated as infrastructure, not a feature—investors are betting on winner-take-most dynamics similar to LLMs. The capital scale suggests Kling is being positioned to compete directly with OpenAI's Sora and Runway, which means the real moat isn't the model itself but distribution and compute access. If you're building in adjacent generative media (3D, audio, interactive content), watch whether Kling's funding unlocks cheaper inference costs—that's when the category gets real.
Together AI logo
🇺🇸Together AILLM Infra

Together AI builds open-source LLM infrastructure and APIs for developers to run and fine-tune large language models.

$800MSeries C
Investor undisclosed
An $800M Series C for LLM infrastructure in mid-2026 signals that open-source model serving has moved from nice-to-have to essential—likely driven by enterprises wanting inference cost parity with closed APIs without vendor lock-in. Together is probably using this to scale compute capacity, build out managed services, and compete directly with Anthropic/OpenAI on latency and pricing. If you're building any AI product that needs custom models or cost-sensitive inference, this validates that the infrastructure layer is consolidating fast—you'll want to lock in pricing/partnerships now before the winners emerge.
B
🇨🇳Bose QuantumQuantum Computing

Bose Quantum develops quantum computing technology and solutions for enterprise applications.

UndisclosedPre-IPO
Investor undisclosed
K
🇨🇳KlingVideo Generation

Kling generates videos from text and images using AI, enabling creators and businesses to produce content at scale.

$30B
Investor undisclosed
A $30B valuation for a text-to-video model signals the category has moved past hype into infrastructure pricing—this is OpenAI/Anthropic territory, not startup. If you're building in generative media (3D, audio, code), watch whether Kling's scale actually converts to moat or if it becomes a commodity layer; the real margin will be in vertical applications, not the model itself.
VAST logo
🇨🇳VASTWorld Models

VAST builds world models that enable AI systems to understand and simulate complex environments.

$140MSeries A3
Investor undisclosed
A $140M Series A3 for world models signals that simulation-as-infrastructure is moving from research toy to production dependency—someone's betting this becomes the backbone for robotics, autonomous systems, or synthetic data pipelines. At this stage and size, VAST is likely burning cash on compute for training massive environment simulators and building out enterprise go-to-market. If you're building anything that needs to predict or control physical/digital systems (autonomous vehicles, game engines, industrial automation), watch whether world models become a commodity layer or stay vertically integrated—that determines your build-vs-buy calculus in 18 months.
Kling AI logo
🇨🇳Kling AIVideo GenerationVerified

Kling AI generates high-quality videos from text and images using advanced AI models.

$3BStrategic
Alibaba logo
Alibaba
Tencent logoCG
+10
A $3B strategic round from Alibaba + Tencent signals China is treating video generation as infrastructure, not a feature—expect aggressive deployment into their ecosystems (e-commerce, short-form content, ads). If you're building in adjacent generative media (3D, audio, interactive), watch how Kling gets distribution; Chinese incumbents rarely fund competitors without integration plans, so the real moat here is access to their platforms, not the model itself.
DeepSeek logo
🇨🇳DeepSeekLLM

DeepSeek builds large language models for developers and enterprises using advanced AI research.

$7B
Investor undisclosed
A $7B raise for a Chinese LLM player signals serious capital is flowing to non-US AI labs—likely betting on cost-efficient model training and enterprise adoption in Asia. At this stage and size, DeepSeek is probably building inference infrastructure, expanding model families, and securing compute capacity to compete on speed and price rather than pure capability. If you're building AI tooling or applications, watch whether they're winning enterprise deals in China/SEA—that's the real tell for whether Western LLM dominance is actually eroding.
P
🇨🇳Perception Era TechnologyRobotics

Perception Era Technology builds electronic skin and tactile sensing hardware for robots to perceive and manipulate objects through touch.

UndisclosedAngel
S
Songhe Capital
A single angel check from Songhe Capital in mid-2026 doesn't tell you much—this is pre-seed noise, not a market signal. Without knowing the company, stage, or amount, the only real takeaway is that Songhe is still writing small checks, which matters only if you're in their geographic/sector focus (likely China-adjacent or Asia-focused tech). Skip this unless you're specifically tracking that fund's thesis.
MobAI logo
🇨🇳MobAIInteractive Narrative AIVerified

MobAI builds AI-powered interactive storytelling experiences that let users create and explore dynamic narratives.

UndisclosedAngel
赤
赤子城科技
A Chinese investor leading an angel round in July 2026 suggests continued appetite for early-stage bets despite macro headwinds—worth watching if you're in Asia-Pacific. The company is likely burning this on product-market fit validation and initial go-to-market, so if you're building B2B SaaS or a consumer app in a similar vertical, this signals your TAM probably has enough dry powder to support multiple funded competitors.
Amity Robotics logo
🇹🇭Amity RoboticsRobotics

Amity Robotics builds AI-powered robotic kiosks and mobile robots for customer-facing automation in hospitality, retail, and real estate.

$7MSeed
E
East Ventures
5
A $7M seed for physical robots in Southeast Asia signals investors are betting on labor arbitrage—where robot capex beats local wage inflation in hospitality/retail. If you're building software for these verticals, this matters: Amity's success validates that customers will actually deploy unfamiliar hardware, which means your SaaS layer (scheduling, analytics, customer data) just got a clearer path to adoption if you can integrate with their fleet.
DexForce logo
🇨🇳DexForceRobotics

DexForce builds humanoid robots and embodied AI software for industrial and commercial automation.

$14MStrategic
Investor undisclosed
China's state-backed funds leading a $14M strategic round into humanoid robotics signals Beijing is moving past R&D into commercialization—this isn't venture money, it's industrial policy capital. If you're building automation software or hardware that needs to integrate with physical systems, watch whether DexForce's stack becomes a de facto standard in Chinese manufacturing; if it does, you're either building on top of it or competing against subsidized incumbents.
TwelveLabs logo
TwelveLabsVideo AI

TwelveLabs builds AI models that understand and search video content at scale for developers and enterprises.

$100MSeries B
Investor undisclosed
A $100M Series B for video understanding signals that multimodal AI infrastructure is moving past hype into actual enterprise deployment—companies are willing to bet serious capital on the plumbing layer. If you're building anything that needs to index, search, or extract meaning from video (compliance, content moderation, autonomous systems), this validates that the underlying models are mature enough to be a dependency, not a research project.
EquiLibre Technologies logo
🇨🇿EquiLibre TechnologiesReinforcement Learning / Quantitative Trading

EquiLibre Technologies builds reinforcement learning algorithms for quantitative trading, partnering with hedge funds to trade billions daily across stocks and crypto.

UndisclosedSeries A
Creandum logo
Creandum
A Series A from Creandum in mid-2026 signals they're still writing checks in a selective market—Creandum typically backs European founders with strong unit economics, so this company likely proved retention or revenue traction before raising. If you're building B2B SaaS in Europe, this is a reminder that the bar for Series A hasn't dropped; investors want to see the business work before they fund growth.
Kimi logo
🇨🇳KimiLLM

Kimi is an AI assistant by Moonshot AI offering LLM-powered API services and developer tools for enterprises.

Undisclosed
Investor undisclosed
轨
🇨🇳

轨道辰光

Space Computing Infrastructure

轨道辰光 builds space computing infrastructure for satellite and orbital applications using AI-optimized systems.

$67K
Investor undisclosed
A $100K check into Chinese space-computing infrastructure suggests the category is still pre-product or pre-traction—this is exploration capital, not validation. If you're building orbital software or ground stations, watch whether Chinese VCs start writing bigger checks here; right now it signals interest but not yet conviction in the commercial unit economics.
K
🇨🇳Kuawei IntelligenceEmbodied AI

Kuawei Intelligence builds embodied AI systems for robotics and autonomous applications.

$14M
Investor undisclosed
A $14M Series A for embodied AI in China signals the category is moving past research into deployment—this isn't pure robotics, it's the software layer that makes robots actually useful. If you're building in autonomous systems, perception, or sim-to-real tooling, watch whether Chinese embodied AI companies start outpacing Western ones on cost and iteration speed; that's the real competitive pressure.
A
🇸🇬Akro AIDocument Intelligence

Akro AI extracts data from unstructured documents for regulated industries using privacy-preserving AI models.

$700KPre-Seed
A
Amigos Venture Capital
A $700k pre-seed for document AI in regulated industries signals investors are betting on privacy-first extraction as table stakes—not a differentiator. If you're building compliance tooling or workflow automation that touches PDFs/contracts, this validates that on-device or federated inference is becoming the expected baseline, not a premium feature.
SpinQ logo
🇨🇳SpinQQuantum Computing

SpinQ builds quantum computing hardware and software for enterprises and researchers.

$14MSeries D
Investor undisclosed
China's quantum hardware play just raised $14M at Series D—a sign that enterprise quantum applications (likely optimization, simulation) are moving past pure R&D into pilot deployments. If you're building infrastructure for AI or scientific computing, watch whether SpinQ's software stack becomes a distribution moat; right now, quantum's real bottleneck is usable abstractions, not just raw qubit count.
Z
🇨🇳ZhipingfangEmbodied AI

Zhipingfang builds neuromorphic AI architectures for robotics inspired by the human nervous system.

$700M
Investor undisclosed
A $700M Series round for neuromorphic robotics in China signals serious state backing for embodied AI as infrastructure, not just research—the investor mix (sovereign funds, pharma, liquor conglomerates) suggests this is being positioned as foundational tech for manufacturing and logistics. If you're building robotics or embodied AI anywhere, watch whether neuromorphic approaches actually reduce latency/power vs. transformer-based alternatives in production; that's the bet being made here.
X Square Robot logo
🇨🇳X Square RobotRobotics

X Square Robot builds embodied AI systems combining foundation models, custom hardware, and data pipelines for real-world robotic applications.

UndisclosedSeries C
Investor undisclosed
自
🇨🇳

自变量

AI

自变量 builds AI solutions for Chinese enterprises.

UndisclosedSeries C
Investor undisclosed
State-backed AI fund + Sequoia China co-leading a Series C in mid-2026 signals Beijing is actively deploying capital into domestic AI infrastructure plays—likely a foundation model, chip design, or inference layer company. If you're building AI applications in China or competing against Chinese teams, watch whether this company's tech becomes a de facto standard; state-backed rounds often come with distribution advantages that reshape competitive moats.
X Square Robot logo
🇨🇳X Square RobotRobotics

X Square Robot builds embodied AI systems combining foundation models, custom hardware, and data pipelines for real-world robotic applications.

Undisclosed
Investor undisclosed
Mykare logo
🇮🇳MykareAI Healthcare Operations

Mykare builds an AI operating system for hospitals to automate patient scheduling, acquisition, and feedback collection via voice agents.

$3.2MSeed
Investor undisclosed
A $3.2M seed for hospital ops automation signals that AI voice agents are finally moving past chatbots into revenue-touching workflows—this is about replacing FTEs, not just deflecting support tickets. If you're building in adjacent healthcare operations (billing, discharge, referral workflows), the fact that Mykare landed Andrew Parker + a Middle Eastern family office suggests there's real appetite for India-first healthtech that solves acute labor shortages, not just US-market problems.
P
Philo AIAI

Philo AI builds AI solutions for enterprise customers.

$10MSeries A
Investor undisclosed
A $10M Series A for an enterprise AI solutions company in mid-2026 suggests the market is still hungry for vertical-specific AI plays, but the bar for generalist 'AI solutions' is getting higher—this likely means Philo has a defensible wedge (specific use case, customer type, or data moat) rather than just wrapping an LLM. At this stage and size, expect them to hire sales/CS heavily and lock in 5-10 marquee customers; if you're building adjacent enterprise AI, watch whether they're winning on implementation speed or on outcomes—that tells you if the bottleneck is still 'getting AI to work' or 'getting customers to adopt it.'
Sand.ai logo
🇨🇳Sand.aiVideo AI / World Models

Sand.ai builds video understanding AI and world models for developers and enterprises using deep video analysis.

$100M
Investor undisclosed
A $100M round for video understanding in China signals serious enterprise appetite for multimodal AI beyond text—likely driven by manufacturing, logistics, and surveillance use cases where video data is abundant but underutilized. At this stage and size, Sand.ai is probably building toward a platform play: APIs for developers + verticalized solutions for large enterprises, which means they're betting video understanding becomes as commoditized as image recognition. If you're building in robotics, autonomous systems, or any space that needs real-time scene understanding, watch how they price and distribute—that'll tell you whether video AI stays a custom engineering problem or becomes a plug-and-play layer.
Anthropic logo
🇺🇸AnthropicLLM Infra

Anthropic builds large language models and AI systems for developers and enterprises.

UndisclosedSeries H
Investor undisclosed
Series H in mid-2026 with undisclosed investors is a red flag—either the round is still being quietly assembled or the company is avoiding public signal. If this is a real H-round, the company likely needs $50M+ for geographic expansion or to defend against a well-funded competitor; founders in adjacent spaces should watch whether this company starts aggressive hiring or M&A, which signals the category is consolidating faster than expected.
智平方 logo
🇨🇳

智平方

AI

智平方 builds AI-powered logistics optimization software for supply chain enterprises in China.

$700M
Investor undisclosed
A $700M round for logistics optimization in China signals that supply chain software is now a venture-scale category there—likely driven by e-commerce density and manufacturing complexity making traditional routing obsolete. At this check size, they're probably building out enterprise sales, regional infrastructure, and possibly expanding into adjacent verticals (warehousing, last-mile). If you're building B2B SaaS for operations in Asia, this shows that deep domain software (not just dashboards) can command massive capital when it solves a unit economics problem at scale.
DeepSeek logo
🇨🇳DeepSeekLLM

DeepSeek builds large language models for developers and enterprises using advanced AI research.

$700MStrategic
Investor undisclosed
A $700M strategic round for a Chinese LLM player signals that inference efficiency and cost-per-token are now table stakes—DeepSeek's likely using this to scale production and compete on price against OpenAI/Claude in enterprise deals. If you're building any AI application layer (agents, RAG, fine-tuning platforms), this matters because your unit economics just got tighter; you can't assume customers will pay for premium models when a well-funded alternative runs 10x cheaper.
Rocket logo
🇮🇳RocketAI-Powered Development Platform

Rocket builds an AI-powered platform for creating apps, websites, and dashboards from natural language instructions.

UndisclosedSeries A
S
Susquehanna International Group
SAccel logoS
DeepSeek logo
🇨🇳DeepSeekLLM

DeepSeek builds large language models for developers and enterprises using advanced AI research.

$7B
Investor undisclosed
N
🇨🇳Nianxiang TechnologyNeural Interface / Brain-Computer InterfaceVerified

Nianxiang builds a non-invasive wristband that captures neural signals for gesture control and hand tracking in AI applications.

$14KAngel
Y
Yongjun Xingmang
PY
A Chinese angel round for neural interface hardware in mid-2026 suggests the gesture-control-as-input layer is still pre-product; if this were proven, you'd see Series A by now. The real signal: wristband-based neural capture is attracting early capital in China specifically, which means either the regulatory path is clearer there or the cost structure works at scale in that market—worth watching if you're building any input layer for embodied AI.
DeepSeek logo
🇨🇳DeepSeekLLM

DeepSeek builds large language models for developers and enterprises using advanced AI research.

$714MSeries A
Investor undisclosed
A $714M Series A for a Chinese LLM company signals that frontier model development is no longer a winner-take-all game—there's real capital flowing to non-US players with differentiated research. The round size and investor mix (Tencent, JD, NetEase) suggests DeepSeek is being positioned as infrastructure for China's AI stack, not just a research lab. If you're building AI applications or tools, this matters because it means your moat can't rely on exclusive access to one model provider anymore.
群
🇨🇳

群青智能

Robotics

群青智能 builds embodied AI robots for manufacturing and industrial automation.

$14M
Investor undisclosed
A $14M Series A for Chinese embodied robotics in mid-2026 signals that hardware-software integration for factory floors is finally moving past proof-of-concept—investors are betting on the unit economics working at scale. If you're building any kind of perception or control layer (vision, motion planning, fleet coordination), watch how 群青智能 prices their robots and what margin they're targeting; that'll tell you whether the real money is in the hardware or the software stack sitting on top.
Peec logo
🇩🇪PeecAI Search

Peec builds AI-powered search technology for discovering and accessing information efficiently.

Undisclosed
Investor undisclosed
协
🇨🇳

协创数据

AI InfrastructureVerified

协创数据 builds intelligent computing infrastructure and data storage solutions for enterprises using AI-optimized systems.

$1.1BStrategic
Investor undisclosed
A $1.1B strategic round for Chinese AI infrastructure in mid-2026 signals that compute capacity—not just models—is still the bottleneck enterprises are willing to pay for, especially outside the US. The scale suggests this is likely going toward expanding GPU clusters and proprietary storage systems to serve domestic enterprises locked out of Western chips. If you're building data-heavy applications (analytics, search, recommendation), watch whether 协创数据's infrastructure becomes the default stack for Chinese enterprises—it could reshape where you need to optimize for latency and cost.
Patronus AI logo
🇺🇸Patronus AIAI Evaluation & TestingVerified

Patronus AI builds simulation environments to stress-test and evaluate AI agents in complex real-world scenarios.

$50MSeries B
G
Greenfield Partners
NLightspeed logoD
+1
A $50M Series B for AI evaluation signals that enterprises are moving past model benchmarks—they need to stress-test agents before deployment, which means real production risk is now top-of-mind. If you're building any AI product that touches workflows or decisions, this validates that your customer's procurement process will include a "does this break in edge cases" phase, so bake testability into your product early rather than bolting it on later.
脸
🇨🇳

脸谱心智

World Models

脸谱心智 builds world models technology for AI systems to understand and simulate physical environments.

Undisclosed
Investor undisclosed
深
🇨🇳

深度机智

AI

深度机智 builds AI solutions for enterprise customers in China.

Undisclosed
Investor undisclosed
念
🇨🇳

念象科技

Human-Computer Interaction

念象科技 builds embodied data entry platforms for next-generation human-computer interaction using AI.

$1.4MAngel
Investor undisclosed
A $1.4M angel for embodied data entry in China signals someone believes gesture/spatial UI is ready to move past R&D—likely betting on AR glasses adoption or industrial automation where hands-free input matters. If you're building any form of input layer (voice, gesture, eye-tracking), watch whether this team ships a real product or stays in the research phase; that'll tell you if the hardware ecosystem is actually there yet.
S
Scaled CognitionAI

Scaled Cognition builds AI systems that learn and improve from human feedback at scale.

$100MSeries A
Investor undisclosed
A $100M Series A for a feedback-loop AI company signals that post-training optimization (RLHF-adjacent work) is now a standalone business, not just an internal capability—investors believe there's defensible margin in the infrastructure layer between raw models and production systems. If you're building anything that needs to adapt to user behavior or domain-specific performance, this validates that outsourcing the feedback-collection and model-tuning piece is becoming a real option rather than DIY.
Cosm logo
🇺🇸CosmImmersive Technology

Cosm builds immersive entertainment venues using advanced projection and spatial technology for mass audiences.

$100MStrategic
S
Sony Pictures
Sony's $100M strategic bet signals that immersive venues are moving from novelty to real estate play—they're not funding tech, they're funding locations and content pipelines. If you're building spatial computing, AR, or event tech, this matters because Sony just signaled they'll fund the *venue layer* if the unit economics work, which means your software/hardware needs to integrate into their deployment model, not compete with it.