Quantinuum builds quantum computing hardware and software for enterprise applications.
UndisclosedIPO
Investor undisclosed
All funding news
AI — funding news
100 recent AI rounds across our tracked sources.
N
Nuoxin Jia develops neural interfaces and biosensors for brain-computer communication applications.
Undisclosed
Investor undisclosed
Tencent backing a Shanghai-based play in May 2026 signals continued appetite for China-native tech, though the lack of disclosed amount and sparse investor list suggests either a smaller check or a strategic/corporate round rather than institutional validation. If you're building in gaming, social, or fintech in Asia, watch whether this company's go-to-market strategy relies on Tencent's distribution—that's the real tell of what Tencent thinks will work in the next 18 months.
Anthropic builds safe, interpretable large language models for developers and enterprises.
UndisclosedSeries H
Anthropic builds safe, interpretable large language models for developers and enterprises.
$9.1B
Investor undisclosed
A $9.1B round for an LLM company in mid-2026 signals that frontier model development is still consolidating around a handful of well-capitalized players—this isn't seed-stage category validation, it's late-stage arms race funding. Anthropic's likely burning this on compute (training runs, inference infrastructure) and hiring top researchers to stay competitive with OpenAI/others on capability and safety claims. If you're building AI applications or tooling, watch whether this capital translates to cheaper/faster API access or if it just funds a capability gap that makes it harder for smaller models to compete.
M
Moxin AI designs efficient AI processors like SparsePrime® for cost-effective AI computing.
$140MSeries C
Investor undisclosed
A $140M Series C for a Chinese AI chip startup signals that efficiency-focused inference hardware is still attracting serious capital—likely because the margin between commodity GPUs and custom silicon keeps widening as models scale. Moxin's probably using this to scale manufacturing and land design wins with hyperscalers who are tired of paying Nvidia's tax on every inference call. If you're building any inference-heavy product (search, recommendations, agents), watch whether these alternative chips actually ship at scale—it could reshape your unit economics in 18-24 months.
Groq builds specialized AI inference chips and cloud services for developers and enterprises to run large language models faster and cheaper.
$650M
Investor undisclosed
A $650M raise for inference chips in mid-2026 signals the market has moved past 'will LLMs be useful' to 'who owns the margin on running them.' Groq's bet is that specialized hardware beats general-purpose GPUs on cost/latency—if this round closes at a high valuation, it means enterprises are actually willing to switch inference providers, which changes the unit economics for anyone building LLM applications. If you're building on top of LLMs, watch whether Groq's inference speed actually moves the needle on your product differentiation or if it stays a cost-optimization play for large-scale deployments.
歌
🇨🇳
歌迷圈子
Entertainment歌迷圈子 builds a concert and live performance platform for fans to discover, engage, and connect around events.
$700KSeed
Investor undisclosed
A $700k seed for a concert discovery platform in China signals investors still see room to consolidate fragmented live event discovery—but the modest check size suggests they're betting on niche fan engagement rather than competing head-to-head with Douyin/WeChat's existing event ecosystems. If you're building community tools around any passion vertical (gaming, sports, music), watch how they monetize beyond ticketing; the real margin is likely in fan-to-fan commerce or creator tools, not just discovery.
Triomics builds AI software for oncology practices to automate clinical trial matching, appointment prep, and chart summarization.
$22MSeries B
A $22M Series B for oncology workflow automation signals that VCs are betting on AI-as-glue for fragmented clinical systems—not just diagnosis. Triomics is likely using this to expand from trial matching into the full patient journey (EHR integration, insurance pre-auth, follow-up scheduling). If you're building in adjacent healthcare verticals (rheumatology, cardiology), watch whether they're hiring for horizontal platform plays or staying vertical-focused; that tells you if the TAM is really about oncology or about replacing the broken appointment/documentation layer everywhere.
F
FunloomAI is an AI content co-creation platform that handles repetitive tasks so creators can focus on creative work.
UndisclosedPre-A
Investor undisclosed
DeepSeek builds large language models for developers and enterprises using advanced AI research.
UndisclosedSeries A
State-backed Chinese capital moving into Series A suggests Beijing is actively de-risking semiconductor or advanced manufacturing plays—this isn't opportunistic, it's strategic. The 国家大基金 (National IC Fund) typically writes checks when they want to build domestic supply chains, so if this company is in chips, materials, or tooling, expect regulatory tailwinds but also eventual pressure to localize. If you're building infrastructure in adjacent sectors (EDA, packaging, materials science), watch whether follow-on rounds stay domestic or if they need foreign capital—that'll tell you how much geopolitical friction you're actually facing.
墨
🇨🇳
墨芯
AI Chip/Hardware墨芯 builds cost-effective AI chips and hardware to reduce computational expenses for AI workloads.
$140MSeries C
Investor undisclosed
A $140M Series C for a Chinese AI chip startup signals that inference cost reduction is still table stakes—not a solved problem despite all the GPU competition. If you're building AI applications, this matters: the economics of your product depend heavily on whether inference costs keep dropping or plateau, and this round suggests the market still believes there's 2-3x efficiency gains left to capture. Watch whether they're targeting specific workloads (vision, language, etc.) or going broad—that'll tell you if specialized chips are winning or if general-purpose efficiency is the real moat.
Moxin builds sparse computing acceleration cards that optimize AI workload performance and efficiency.
$140M
Investor undisclosed
A $140M raise for a Chinese sparse computing accelerator in May 2026 signals that inference optimization—not just training—is now a capital-intensive category. Moxin likely uses this to scale manufacturing and build out software/compiler support, which is the real moat in accelerators. If you're building any AI infrastructure that touches model serving or edge deployment, watch whether sparse compute becomes table-stakes for your cost model.
M
Moffett AI builds sparse computing chips for AI inference acceleration using dual-sparsity algorithms.
$14MSeries C
A $14M Series C for a sparse inference chip startup signals China's doubling down on edge AI efficiency—this isn't about raw compute, it's about inference cost at scale. Moffett's dual-sparsity angle (model + hardware level) suggests they're solving a real bottleneck: running LLMs cheaply on-device. If you're building inference infrastructure or edge AI products, watch whether sparse chips actually move the needle on latency vs. just power—that's the bet here.
Moxin builds sparse computing acceleration cards that optimize AI workload performance and efficiency.
$140MSeries C
Investor undisclosed
A $140M Series C for a Chinese AI accelerator in mid-2026 signals that sparse compute—not just raw throughput—is now table stakes for inference economics. This money likely funds manufacturing scale and go-to-market in US/EU despite geopolitical friction, which means the efficiency gains must be real enough to justify the regulatory headwind. If you're building any inference-heavy product (search, recommendations, agents), watch whether sparse cards actually move your unit economics; if Moxin's gaining traction, it's a signal that customers care more about cost-per-inference than peak performance.
Z
Zhongke Diwuji builds universal physical AI models for industrial robotics using few-shot learning.
UndisclosedSeries A
China's doubling down on semiconductor supply chain localization—this round's investor mix (state-backed funds + industry players) signals the government is actively de-risking chip manufacturing dependencies. If you're building infra or tooling for chip design/fab, expect tailwinds in China but also watch for export controls tightening around your tech.
新
🇨🇳
新智具身
Robotics新智具身 develops tactile sensing technology that gives robots human-like touch perception.
$140KAngel
Investor undisclosed
A $100K angel for tactile sensing in robotics is a signal that China's hardware AI stack is filling in the gaps—touch perception is genuinely hard and underinvested compared to vision/language. If you're building robot software or manipulation tasks, this matters because tactile feedback is the bottleneck between lab demos and real-world dexterity; watch whether this team's approach (likely MEMS sensors or learned representations) becomes table stakes for your use case.
C
China National Aviation operates a major state-owned airline providing domestic and international air transportation services.
$2.8BStrategic
Investor undisclosed
A $2.8B strategic round into a state-owned Chinese airline in mid-2026 signals Beijing is doubling down on aviation infrastructure as a geopolitical/economic lever—likely fleet modernization and route expansion rather than operational innovation. If you're building logistics, supply chain, or B2B travel software, watch whether this capital flows toward digital infrastructure or stays siloed in hardware; that tells you if Chinese carriers will actually adopt third-party platforms or build walled gardens.
凌
🇨🇳
凌云智矿
AI4Earth凌云智矿 uses AI and earth observation to provide mining intelligence and site monitoring solutions.
$10MPre-A
Investor undisclosed
A $10M pre-A for mining intelligence signals that earth observation + AI is moving from climate/ESG theater into operational ROI—mining companies care about ore grade prediction and equipment downtime more than carbon scores. If you're building in supply chain visibility, logistics optimization, or industrial asset monitoring, watch how they monetize: per-site subscription, usage-based pricing on imagery, or embedded into mining software. The China play matters too—domestic mining ops have less regulatory friction to adopt new tools than Western counterparts, making it a faster validation market.
Stilta builds AI-powered patent analysis software for IP teams to automate research and discovery of owned patents.
$10.5MSeed
A10M seed for patent automation signals that IP teams are finally willing to pay for AI tooling to replace manual grunt work—this is less about market timing and more about the category maturing past skepticism. Stilta's likely burning this on product depth (handling complex patent taxonomies), sales to in-house counsel, and probably some enterprise integrations. If you're building any workflow automation for knowledge workers (contracts, compliance, research), watch how Stilta prices and sells: legal teams are notoriously sticky once you own their process, so their GTM playbook matters more than the product itself.
Lucis builds AI-powered preventative health insights for individuals using predictive analytics.
$20MSeries A
General Catalyst leading a $20M Series A for predictive health signals suggests the bar for clinical-grade AI validation is rising—this isn't a consumer wellness play, it's infrastructure for risk stratification. If you're building in adjacent health data spaces (genomics, wearables, claims), watch whether Lucis's model actually reduces downstream costs or just predicts them; that determines if payers fund the category or let it stay consumer-only.
L
Lightwheel AI generates training data and simulation environments for embodied AI systems and industrial robotics applications.
UndisclosedPre-A+
Q
Qiaoitian Intelligence builds AI solutions for enterprise customers in China.
$14MSeries A+
Investor undisclosed
A $14M Series A+ for a China-focused enterprise AI play signals that the market is still funding applied AI for specific verticals rather than broad foundation models—investors are betting on execution in regulated markets where relationships matter more than raw compute. If you're building B2B AI tools, watch whether Qiaoitian's go-to-market relies on system integrators or direct sales; that'll tell you if enterprise AI adoption in Asia is moving faster than the West.
OpenRouter provides a unified API gateway for developers to access multiple LLMs from different providers through a single integration.
$113MSeries B
A $113M Series B for an LLM router signals that the multi-model abstraction layer is now table stakes—not a nice-to-have. CapitalG's bet suggests the real value isn't in picking winners between Claude/GPT/Llama, but in letting developers switch seamlessly as pricing and performance shift. If you're building any AI product that touches multiple model providers, this validates that your customers will demand portability; if you're building *for* developers, watch how OpenRouter uses this capital to lock in distribution before the model providers build this in-house.
B
Blue Dot Touch builds six-axis force sensors for collaborative robots, surgical robots, and precision assembly systems.
UndisclosedSeries C++
A Series C++ in May 2026 with SAIC Jinkong + a consortium including Sequoia China signals robotics/automation is still pulling institutional capital in Asia, though the round structure (multiple co-leads) suggests either a competitive process or risk-sharing on a higher valuation. At this stage, money goes to manufacturing scale, field deployment, and likely regulatory/safety validation—not product-market fit. If you're building adjacent hardware or automation software, watch whether this company's go-to-market is direct enterprise or through system integrators; that choice will constrain your own distribution options.
A $21M round to a Chinese enterprise AI company in mid-2026 suggests the market is still hungry for vertical AI solutions, but the investor (米奥会展, primarily an exhibition/events company) is an odd signal—this looks more like a strategic/corporate investor than VC conviction. If you're building B2B AI tools, watch whether Chinese enterprises are consolidating around fewer vendors or still fragmenting across point solutions.
桥
🇨🇳
桥田智能
AI桥田智能builds AI solutions for enterprise automation and intelligent systems.
$14M
Investor undisclosed
A $14M round for a Chinese enterprise AI automation play in mid-2026 suggests the market is still hungry for vertical-specific automation tools, but the lack of named investors and sparse details make it hard to read whether this is genuine momentum or a smaller regional round. If you're building workflow automation or RPA-adjacent products, watch whether Chinese teams start competing on cost/speed in your verticals—they often do once they've proven the model domestically.
B
Blue Dot Touch builds six-axis force sensors for collaborative robots, surgical robots, and precision assembly systems.
UndisclosedSeries C
阶
🇨🇳
阶跃星辰
AI阶跃星辰builds large language models and AI infrastructure for enterprise applications in China.
$21MStrategic
Investor undisclosed
A $21M strategic round for a Chinese LLM/infra play in mid-2026 signals that enterprise AI adoption in China is moving past pilot phase—strategic investors (likely from established tech/enterprise groups) are betting on infrastructure stickiness rather than model differentiation. If you're building B2B AI tools, this matters: it suggests the real margin is in vertical-specific deployment and compliance layers for regulated industries, not in competing on base model performance.
A $14M round for an enterprise AI shop in China in mid-2026 suggests the market is still hungry for vertical-specific AI applications rather than horizontal platforms—but the undisclosed investors and sparse details make it hard to read whether this is genuine traction or capital recycling. If you're building B2B AI tools, watch whether Chinese enterprise AI companies start shipping faster than their US counterparts; if they do, it's a sign that implementation speed (not model quality) is becoming the real moat.
T
Tianji Intelligence builds intelligent robotic systems for industrial and commercial applications.
$140M
Investor undisclosed
A $140M round for Chinese enterprise AI software signals that the market is willing to fund AI tooling at scale in Asia, even as Western investors get pickier about unit economics. At this stage and size, Tianji is likely building out sales infrastructure and expanding their product suite across verticals—not burning cash on R&D. If you're building B2B AI in any region, watch whether they can actually retain customers at enterprise pricing; that's the real test of whether this category has legs.
Peec AI helps brands optimize their visibility in AI search results through generative engine optimization dashboards.
$21MSeries A
Investor undisclosed
A $21M Series A for an AI search optimization tool signals that brands are treating generative search as a distinct, measurable channel—not just SEO 2.0. This is real budget reallocation happening now, which means if you're building anything that touches brand visibility or search, you need a generative engine strategy, not just a Google one.
Anthropic builds safe, interpretable large language models for developers and enterprises.
Undisclosed
Investor undisclosed
Anthropic builds safe, interpretable large language models for developers and enterprises.
Undisclosed
Sequoia + Dragoneer + Altimeter moving together signals serious conviction in a specific thesis—this isn't a follow-on, it's a new category bet with tier-1 dry powder. Without the company name, the real tell is the investor *combination*: Dragoneer typically leads growth rounds in proven models, Altimeter hunts for scale plays, and Greenoaks backs infrastructure. If you're building in fintech, enterprise software, or infrastructure, watch what this round actually funds—the investor mix hints at either a breakout growth moment or a new playbook being validated.
Hark builds next-generation GPU infrastructure and AI hardware for enterprises scaling large language models.
$700MSeries A
A $700M Series A with chip makers as co-investors signals the GPU shortage narrative has shifted from scarcity to *specialization*—enterprises need custom inference hardware, not just more H100s. If you're building any LLM application layer, this validates that your unit economics depend on hardware partnerships you probably can't negotiate alone yet, so either plan for that dependency or start thinking about how to abstract it away.
Hark builds next-generation GPU infrastructure and AI hardware for enterprises scaling large language models.
$700M
Investor undisclosed
A $700M raise for GPU infrastructure in May 2026 signals the market still believes there's room for non-NVIDIA players in enterprise LLM deployment—but the co-investment from NVIDIA, AMD, and Intel suggests this is less about disruption and more about carving out a managed-services layer on top of commodity chips. If you're building anything that touches model serving, inference optimization, or enterprise GPU orchestration, this validates that customers will pay for abstraction layers that hide hardware complexity, even if they're not buying the silicon itself.
H
HiDream AI builds a 200B+ parameter multimodal language model for processing text, images, and other data types.
Undisclosed
Investor undisclosed
中
🇨🇳
中科沌序
Collective AGI, Autonomous SystemsVerified中科沌序 builds collective AGI systems for autonomous society applications and low-altitude safety using multi-agent coordination.
UndisclosedAngel
Chinese early-stage VCs are still actively deploying into angel rounds in 2026, suggesting confidence in founder-led deal flow despite macro headwinds. If you're building B2B infrastructure or SaaS in Asia, this signals that local capital is moving faster than Western equivalents at the seed stage—worth considering for your fundraising geography.
MatriQ develops logical qubits and commercial quantum computing solutions for enterprise applications.
Undisclosed
Investor undisclosed
F
Funlinc develops and expands entertainment IP across the full industry chain for global audiences.
UndisclosedStrategic
Investor undisclosed
Secai Marche connects fresh food suppliers to HORECA buyers across Southeast Asia with integrated supply chain financing and payments.
Undisclosed
Acelen builds sustainable aviation fuel from biomass in Brazil to decarbonize commercial aviation.
$1.5BStrategic
A $1.5B strategic check from a sovereign wealth fund signals SAF is moving from venture bet to infrastructure play—the capital requirements are too high for traditional VC, and offtake agreements with airlines are now bankable enough for patient capital. If you're in climate tech, this shows the path to scale isn't venture rounds; it's securing strategic LPs + long-term contracts early, then raising massive growth capital once unit economics are proven.
Modal builds serverless infrastructure for running AI models and LLM applications at scale.
$355MSeries C
A $355M Series C for LLM infra in mid-2026 signals the market has moved past 'will enterprises adopt AI?' to 'how do we actually run this stuff cheaply and reliably?'—Modal's bet is that serverless beats traditional GPU clouds for variable workloads. If you're building any tool that needs to call models on-demand (agents, RAG systems, batch processing), watch how Modal prices compute vs. alternatives; that unit economics floor will constrain what's viable in your category.
SHAREBOT provides on-demand robot rental and scheduling for industrial, commercial, and logistics operations.
UndisclosedSeries A+
Chinese mega-cap family offices and conglomerates are now leading Series A+ rounds—a shift from pure VC that signals confidence in late-stage China tech despite regulatory headwinds. If you're building infrastructure or B2B SaaS in Asia, this round type means patient capital is available, but expect strategic stakeholders who want operational integration, not just returns.
E
Energy Bridge Technology builds prefabricated superconducting power modules for AI data centers, reducing energy loss and cooling complexity.
Undisclosed
Gumloop enables non-technical employees to build and deploy AI agents for automating complex workflows without coding.
$50MSeries B
A $50M Series B for no-code AI agents signals that workflow automation is moving past the hype phase—investors are betting on adoption velocity among non-technical users, not just technical feasibility. Gumloop's mix of tier-1 VCs (Benchmark, First Round) plus strategic players (Shopify) suggests the real money is in enterprise workflows, not consumer toys. If you're building in adjacent automation spaces (RPA, business process software, internal tools), this validates that buyers will pay for speed-to-deployment over customization—and that the bottleneck is now distribution and trust, not technology.
AutoAgents.ai deploys specialized AI agents to automate complex, multi-step knowledge work for enterprises.
UndisclosedPre-A
Gimlet Labs builds AI inference orchestration software that runs workloads across diverse hardware types.
$80MSeries A
An $80M Series A for inference orchestration signals that the market is past the 'which GPU to buy' phase and into 'how do we run workloads across whatever hardware we already own'—a pragmatic shift as GPU scarcity eases and cost optimization becomes the real lever. Gimlet's likely burning this on sales/go-to-market and R&D to support more silicon types (TPUs, custom accelerators, CPUs), which means they're betting enterprises will standardize on orchestration layers rather than lock into single vendors. If you're building any layer of the inference stack (quantization, routing, caching), watch whether Gimlet becomes the de facto control plane—it changes what you optimize for.
Runway builds AI video generation tools for creators using general world models.
$10M
Investor undisclosed
Runway's $10M round with Nvidia as a backer signals that video generation is moving from research toy to creator infrastructure—Nvidia wouldn't touch this if inference costs weren't approaching viability. The check size + stage suggests they're burning cash on compute to improve quality/speed rather than hiring, so watch their model release cadence over the next 6 months as the real signal. If you're building any content tool (design, music, animation), this validates that end-users will adopt AI generation if latency drops below ~30 seconds—worth testing that assumption in your own product.
Kompas VC invests in industrial tech and manufacturing startups solving supply chain, infrastructure, and sustainability challenges across Europe, US, and China.
$174.4MFund Raise
Investor undisclosed
A $174M industrial tech fund closing in 2026 signals that deep-tech manufacturing is finally getting institutional dry powder—LPs are betting on supply chain reshoring and decarbonization as structural, not cyclical, trends. If you're building in adjacent infrastructure or climate tech, this means there's now real capital velocity behind the boring stuff; the bar is less about hype and more about unit economics and regulatory tailwinds.
Eclipse is a venture capital firm investing in physical AI startups across transportation, energy, infrastructure, and defense.
$1.3B
Investor undisclosed
A $1.3B fund closing for physical AI in 2026 signals that the infrastructure/defense/energy bet has moved from hype to institutional capital allocation—this isn't a thesis fund, it's a deployment vehicle. If you're building in robotics, autonomous systems, or hardware-software stacks, this means your TAM just got validated by serious LPs, but also that you'll face well-funded competitors within 18 months.
Nyne builds an intelligence layer that helps AI agents understand human behavior from digital footprints to enable autonomous purchasing and scheduling.
$5.3MSeed
A $5.3M seed for behavioral inference in agent workflows signals investors believe autonomous purchasing/scheduling is the next bottleneck after basic agent reasoning—this is pre-hype money, not FOMO. Nyne's likely burning this on training data pipelines and inference infrastructure to make agents actually *useful* at intent prediction rather than just task execution. If you're building agent tooling or workflows, watch whether behavioral data becomes a defensible moat or a commodity—Nyne's traction will tell you which.
N
Neysa Networks provides GPU-powered cloud infrastructure for enterprises and startups to train, customize, and deploy AI models at scale.
$600MStrategic
A $600M strategic round from Blackstone signals that GPU infrastructure is graduating from startup-land to institutional infrastructure—this isn't venture capital, it's real estate money treating compute like data centers. If you're building any AI application layer (agents, RAG, fine-tuning platforms), watch whether Neysa's pricing or availability becomes a constraint; right now, the bottleneck is shifting from 'can we get GPUs?' to 'can we afford them at scale,' and that's where your unit economics break or hold.
ComfyUI provides a node-based workflow platform for creators to fine-tune generative AI outputs across images, video, and audio.
$30MSeries B
A $30M Series B for a node-based AI workflow tool signals that the market is moving past "one-click generation" toward creator-grade control—investors are betting on the prosumer/professional segment, not consumer. This money likely funds sales/GTM expansion and deeper integrations with existing creative software (Adobe, DaVinci, etc.). If you're building any AI tool for creators, watch whether ComfyUI's node-based UX becomes table stakes or stays niche—it'll tell you whether your audience wants simplicity or power.
Cerebras builds AI-optimized semiconductor processors for training and inference workloads.
UndisclosedIPO
Investor undisclosed
Aera Technology builds AI-powered business automation software for enterprise operations teams to optimize decision-making and workflows.
UndisclosedStrategic
Sycamore builds an agentic orchestration platform for enterprises to deploy and manage AI agents across infrastructure and data systems.
$65MSeed
A $65M seed for agent orchestration signals that enterprises are past the chatbot phase—they're actually trying to deploy agents into production systems, and the plumbing to manage that is now fundable. If you're building any kind of multi-step automation or workflow tool, this validates that buyers will pay for the control layer that sits between agents and their existing infrastructure, not just the agents themselves.
Sequen builds real-time personalization infrastructure for consumer companies using large event models, enabling TikTok-style ranking without massive datasets.
$16MSeries A
A $16M Series A for personalization infrastructure signals that the market is finally ready to pay for ranking-as-a-service—the gap between TikTok's proprietary advantage and what most companies can actually build is still massive. If you're building consumer products, this matters because it means real-time personalization is moving from "nice-to-have" to table stakes, and the infrastructure to do it cheaply is now fundable; if you're in adjacent ranking/discovery spaces (search, recommendations, content), watch whether Sequen's "large event models" actually reduce the data moat or just shift it.
Cognichip uses AI to automate chip design, helping engineers reduce development costs by 75% and timelines by 50%.
$60MSeries B
A $60M Series B for chip design automation signals that the bottleneck isn't AI capability anymore—it's applying it to capital-intensive, long-cycle industries where even 50% timeline cuts unlock real value. If you're building in hardware, manufacturing, or any domain with expensive iteration loops, this validates that customers will pay for tools that compress development cycles, not just cut costs.
Positron designs AI chips optimized for machine learning inference and training workloads.
$230MSeries B
A $230M Series B for an inference chip startup signals that the market believes custom silicon for ML workloads is moving past the hype phase—this isn't about replacing NVIDIA, it's about the wedge of domain-specific efficiency. If you're building any AI infrastructure layer (software, tooling, or adjacent hardware), this validates that customers are willing to fund the full stack, not just rent compute from cloud providers.
S
Status AI lets users roleplay as any character across universes in a gamified social platform with user-generated worlds.
$17MSeed and Series A
Investor undisclosed
A $17M seed-to-Series A for character roleplay social is a strong signal that consumer AI agents are moving past chatbots into persistent, social experiences—investors are betting the moat is community + world-building, not just the model. If you're building in creator tools, gaming, or social, this validates that users will spend time in AI-generated spaces if there's social proof and progression mechanics baked in. Watch how they monetize: if it's cosmetics/battle pass rather than API access, that's your playbook for consumer AI retention.
Synthetic builds AI-powered autonomous bookkeeping for software startups, automatically generating accrual-based financial statements.
$10MSeed
A $10M seed for back-office automation signals that VCs are betting on AI agents handling non-core ops at startups—this is less about bookkeeping being sexy and more about the unit economics of replacing junior accountants. If you're building any founder-facing SaaS, watch how Synthetic prices and what % of customers actually adopt: that adoption curve will tell you whether founders will tolerate AI handling their financial records, which matters for any compliance-adjacent product.
Anthropic builds safe, interpretable large language models for developers and enterprises.
$50B
Investor undisclosed
A $50B round for Anthropic in 2026 signals that frontier LLM capability is still the primary battleground—not applications or verticalization. This scale of capital suggests investors believe the moat is in model weights and inference efficiency, not distribution. If you're building AI tooling or vertical SaaS, watch whether Anthropic's next moves are toward cheaper inference or proprietary datasets; that'll tell you if the cost of your LLM dependency is about to compress or stay sticky.
A
AI Health Fund is a venture firm investing in early-stage healthcare AI startups, offering capital and founder support through its Treehub residency program.
$10MFund Formation
Investor undisclosed
A $10M healthcare AI fund closing in 2026 signals that LPs still see runway in the space despite the hype cycle cooling—but the real tell is the residency model, which suggests founders need more than capital (likely regulatory/clinical pathway help). If you're building in adjacent regulated verticals (biotech, medtech, fintech compliance), watch how Treehub structures founder support; that playbook probably works for your space too.
a16z crypto is a venture capital firm investing in cryptocurrency and blockchain startups.
$2.2BFund Raise
Investor undisclosed
a16z closing a $2.2B crypto fund in May 2026 signals institutional conviction that the regulatory fog has cleared enough for mega-scale bets—this isn't a defensive fund, it's offensive capital. If you're building in adjacent infrastructure (AI agents, on-chain data, custody), expect a16z to be aggressively sourcing deals in your space as portfolio hedges, which means higher valuations but also more pressure to have a crypto angle ready to pitch.
ElevenLabs builds human-quality AI voices for enterprises to automate customer interactions across languages.
$500MSeries D
Investor undisclosed
A $500M Series D for voice AI signals that enterprise automation via voice has moved past proof-of-concept—the investor mix (asset managers + strategic corporates like Salesforce, Santander, KPN) suggests real deployment at scale, not hype. If you're building any customer-facing automation (chat, support, IVR), you should assume voice quality is now table-stakes and that ElevenLabs' pricing/latency will become the benchmark you're measured against.
Anthropic builds safe, interpretable large language models for developers and enterprises.
$50B
Investor undisclosed
A $50B round for Anthropic in mid-2026 signals that frontier LLM capability is still the primary battleground—not applications or verticalization. This scale of capital suggests investors believe the moat is in model weights and inference efficiency, not distribution. If you're building in AI tooling or enterprise AI, this is a reminder that your defensibility likely depends on either owning a unique data flywheel or embedding deeply enough that switching costs matter more than raw model performance.
SiFive designs customizable RISC-V chip architectures for enterprises building CPUs and embedded systems.
$400M
A $400M round with Nvidia as co-investor signals serious momentum behind RISC-V as a real alternative to Arm for custom silicon—this isn't hype, it's Nvidia hedging and validating the architecture. SiFive is likely using this to scale design services and tooling for enterprises building their own chips, which means the bottleneck shifting from ISA choice to actually executing custom silicon designs. If you're building any infrastructure around chip design, manufacturing, or verification, watch whether RISC-V adoption accelerates enough to justify parallel tooling investments.
Scope builds AI-powered inspection workflows for industrial operations to automate visual quality control and asset monitoring.
$20MSeries A
Index backing a $20M Series A for industrial visual AI signals that the category has moved past POC hell—customers are now willing to fund scaled deployments, not just pilots. Scope is likely using this to build out vertical-specific models (automotive, pharma, etc.) and hire sales to land enterprise contracts. If you're building any kind of inspection or monitoring tool, watch how they position around ROI metrics; industrial buyers care about cost-per-defect-caught, not just accuracy numbers.
M
MemoraX AI builds AI-powered memory and knowledge management tools for knowledge workers.
UndisclosedSeed+
Investor undisclosed
DEEP Robotics builds high-performance quadruped robots for industrial inspection, research, and emergency rescue.
$350.4MIPO
Investor undisclosed
A Chinese quadruped robotics company hitting IPO at $350M+ signals that hardware-as-a-service for industrial inspection has moved from R&D theater to actual revenue—likely recurring contracts with utilities, factories, and emergency services. If you're building software for physical infrastructure (autonomous monitoring, predictive maintenance, site analytics), this validates that customers will pay for robots to do the legwork, which means your software layer just became more defensible and valuable.
Scope builds AI-powered inspection workflows for industrial operations to automate visual quality control and asset monitoring.
$20M
Investor undisclosed
A $20M Series A for industrial visual inspection suggests the market has moved past pilot purgatory—customers are now willing to fund multi-year deployments if defect detection ROI is clear. If you're building any kind of automated monitoring (supply chain, infrastructure, compliance), this validates that enterprises will pay for AI that replaces expensive manual audits, not just optimizes existing workflows.
Lexroom builds AI-powered legal services for law firms and enterprises to automate document analysis and legal workflows.
$50MSeries B
A $50M Series B for legal doc automation signals that enterprise buyers are finally willing to pay for AI that reduces billable hours—the unit economics work when you're displacing expensive labor. If you're building workflow automation in any regulated vertical (finance, healthcare, compliance), watch how Lexroom structures their go-to-market: they're likely selling to risk-averse procurement teams, which means your sales cycle and proof-of-concept requirements just got a template.
Commure builds AI solutions for healthcare providers to improve clinical workflows and patient outcomes.
$70M
Investor undisclosed
A $70M raise for healthcare AI workflow tooling signals that providers are finally moving past pilots—they're committing real capital to reduce clinician burden, not just chase efficiency gains on paper. If you're building in adjacent verticals (insurance ops, pharma supply chain, medical devices), this validates that healthcare buyers will fund solutions that touch their core operational pain, but you'll need to prove ROI within 12-18 months, not 3 years.
Dust builds an enterprise platform for teams to create and deploy AI agents that collaborate across business functions.
$40MSeries B
A $40M Series B for an agent orchestration platform signals that enterprises are past the chatbot phase and actually deploying multi-agent workflows—this is real workflow automation money, not AI experimentation money. Dust is likely using this to build out compliance/security features, expand sales motion into Fortune 500, and probably acquire or integrate with data warehouse/observability tools (note Datadog and Snowflake in the round). If you're building anything that touches cross-functional processes (finance ops, customer success automation, supply chain), watch how Dust positions agent governance—that's the moat that'll matter in 18 months.
Nectar Social builds an AI agent OS for marketers to automate social media management, moderation, and commerce across platforms.
$30MSeries A
A $30M Series A for social media automation signals that marketing AI is moving past chatbots into workflow replacement—investors are betting on agents that actually execute, not just suggest. Nectar's likely burning this on hiring (especially ML/platform engineers) and expanding platform coverage, since social APIs are a constant moving target. If you're building any kind of multi-platform orchestration tool, watch how they handle the moderation + commerce angle—that's where the real margin lives, not just posting.
G
AI-powered earbuds with built-in camera that enable users to query information about their surroundings through voice, with on-device processing and privacy-first design.
Undisclosed
Investor undisclosed
An embodied AI startup developing robotics and physical AI systems.
$140KAngel
Investor undisclosed
Y
AI data company with revenue exceeding 100M RMB
UndisclosedPre-A++
Investor undisclosed
L
An AI research lab founded by former Alibaba Tongyi Qwen head Lin Junyang, focused on foundational large language models.
Undisclosed
Investor undisclosed
Video-generation AI model spun off from Kuaishou for creating AI-generated videos.
$2BStrategic
Investor undisclosed
China's doubling down on video gen as a defensible moat—$2B strategic from Tencent signals they're treating this like infrastructure, not a feature. If you're building in adjacent generative media (3D, audio, interactive), watch how Kling monetizes; the playbook will likely flow through Tencent's ecosystem first, then get copied globally within 18 months.
Chinese embodied AI startup that develops AI technology and manufactures robot hardware, including robot dogs and full-size humanoid robots.
$700KPre-A
A $700K pre-A for embodied AI in China signals the category is still in hardware-heavy, capital-intensive early days—this isn't a software-first play. The investor mix (traditional auto capital via SAIC, family offices) suggests robotics funding is flowing through established industrial players rather than pure-play VCs, which matters if you're building in adjacent hardware spaces and wondering where patient capital actually sits.
Moonshot AI develops Kimi Operator, an AI assistant platform.
$2B
Investor undisclosed
A $2B raise for a Chinese LLM inference/agent platform signals that post-training compute and agentic workflows are still attracting massive capital despite market saturation—likely because inference margins remain defensible if you can own the operator layer. Moonshot is probably using this to scale inference infrastructure and build distribution moats around autonomous task execution rather than just chat. If you're building in automation or workflow tooling, watch whether they're bundling Kimi Operator into enterprise products; that's the playbook that could compress your TAM.
Chinese AI startup developing AI models and audio synthesis technology (StepAudio TTS).
$2.5BStrategic
Investor undisclosed
A $2.5B strategic round from hardware/telecom players (ZTE, OmniVision, Longcheer) signals Chinese AI infrastructure is consolidating around vertically-integrated stacks—these investors need embedded audio/voice models for devices, not just API access. If you're building voice or multimodal products, watch whether StepFun's TTS becomes the default in Chinese consumer hardware; that's a distribution moat Western startups don't have.
DeepSeek builds large language models for developers and enterprises using advanced AI research.
$700MSeries A
A $700M Series A for a Chinese LLM company signals that frontier model development is now a capital-intensive, geographically distributed race—this isn't a US-only game anymore. The consortium backing (state funds + Alibaba/Tencent) suggests China is treating LLM infrastructure as strategic, which means if you're building AI applications, expect faster iteration cycles and lower inference costs from multiple competing model providers. If you're in adjacent AI infrastructure (evals, fine-tuning, deployment), watch how DeepSeek's model performance forces repricing of the entire stack.
Moonshot AI develops Kimi Operator, an AI assistant platform.
$2B
A $2B round for an AI assistant platform signals China's willingness to deploy massive capital into LLM inference/deployment layers—this isn't about model weights anymore, it's about operational AI. The Meituan + China Mobile backing suggests the money flows toward companies that can integrate AI into existing super-app ecosystems rather than standalone models. If you're building agent infrastructure or workflow automation, watch whether Western investors start matching this capital intensity on similar plays, or if the gap widens.
Chinese AI startup that develops the Kimi conversational AI model with advanced coding capabilities and multi-agent collaboration features.
$2B
A $2B round for a Chinese LLM in mid-2026 signals that coding-capable models are now table stakes for funding, not differentiators—the real bet here is Moonshot's multi-agent architecture and whether it can compete with OpenAI/Claude in enterprise workflows. The Meituan + state-backed investor combo suggests China's AI strategy is consolidating around fewer, better-capitalized players rather than spreading bets. If you're building agent orchestration or enterprise AI tooling, watch whether Moonshot's multi-agent approach becomes the expected UX baseline—it'll force you to either lean into that or find a different moat.
W
Chinese startup developing embodied AI and humanoid robots with unified full-body large models and General Action Expert (GAE) technology.
UndisclosedPre-A+
SHAREBOT provides on-demand robot rental and scheduling for industrial, commercial, and logistics operations.
UndisclosedPre-A
Ultra-portable e-paper secondary display devices that attach magnetically to smartphones, featuring AI-enabled capabilities.
$140K
Investor undisclosed
AI infrastructure company building custom chips and infrastructure to enable them to communicate effectively, focusing on full-stack solutions and open standards for scalable AI infrastructure.
$180MSeries B
Investor undisclosed
A $180M Series B for AI infrastructure in April 2026 signals that custom silicon for AI is moving past the hype phase—investors are betting on full-stack plays that own the chip-to-software layer, not just one. If you're building AI applications or services, this matters because the cost and latency floor you're competing against just got lower; if you're in adjacent infrastructure (networking, memory, orchestration), Upscale's focus on open standards is a tell that interop, not lock-in, is becoming table stakes.
Sycamore builds an agentic orchestration platform for enterprises to deploy and manage AI agents across infrastructure and data systems.
$65MSeed
Gimlet Labs builds AI inference orchestration software that runs workloads across diverse hardware types.
$80MSeries A
