What are the most valued startups in the AI infrastructure market?
Download our beautiful pitch about the AI infrastructure market
In our AI infrastructure market deck, you will find everything you need to understand the market
The AI infrastructure market now includes public semiconductor companies worth more than $100 billion, fast-growing GPU clouds, AI data platforms, networking specialists, and new computing architectures.
This ranking tracks 100 companies across the AI infrastructure ecosystem and orders them using the latest reported or estimated valuation available in our dataset.
We update this list every month to reflect new funding rounds, public-market movements, acquisitions, and other valuation signals.
And if you want to better understand this new industry, you can download our pitch covering the AI infrastructure market.
A quick summary table
| Metric | Value |
|---|---|
| Most valuable AI infrastructure startup | Cambricon, $103.1B |
| Second most valuable AI infrastructure startup | Nebius, $52.7B |
| Median AI infrastructure valuation | Approximately $850M |
| Share of AI infrastructure valuation captured by the top 10 | Approximately 69.0% |
| Top AI infrastructure valuation vs. median | Approximately 121x |
| Median AI infrastructure valuation-to-capital-raised ratio | Approximately 4.6x |
| AI infrastructure startups valued at $1B+ | 46 |
This market map, featured in our AI infrastructure market deck, highlights top companies and startups in the AI infrastructure market
Top startups in the AI infrastructure market ranked by valuation
Here is an updated table that ranks the top startups in the AI infrastructure market based on their latest reported or estimated valuations.
If you want more details about their fundraising activity, you can check our list of the startups who have raised the most funding in the AI infrastructure market.
| # | Startup Name | What They Do | Current Valuation ($) | Valuation Confidence Level | Valuation Type | Evidence Status | Total Funding ($) | Funding Confidence Level |
|---|---|---|---|---|---|---|---|---|
| 1 | Cambricon | Cloud and edge AI processors | $103.1B | Full Confidence | Public Market Cap | Observed | $101M | Not provided |
| 2 | Nebius | Full-stack AI cloud infrastructure | $52.7B | Full Confidence | Public Market Cap | Observed | ~$3.5B+ | Not provided |
| 3 | CoreWeave | GPU cloud for AI workloads | $44.6B | Full Confidence | Public Market Cap | Observed | ~$12.5B | Not provided |
| 4 | Cerebras Systems | Wafer-scale AI compute systems | $41.0B | Full Confidence | Public Market Cap | Observed | $2.92B | Not provided |
| 5 | MetaX Integrated Circuits | General-purpose GPU computing chips | $39.7B | Full Confidence | Public Market Cap | Observed | $304M | Not provided |
| 6 | Moore Threads | Full-function GPU computing platform | $36.8B | Full Confidence | Public Market Cap | Observed | $1.245B | Not provided |
| 7 | Crusoe | Builds large-scale AI data centers | $24.0B–$30.0B | Partial Confidence | Active Raise Valuation | Estimated | ~$2.8B | Not provided |
| 8 | VAST Data | AI-native data infrastructure platform | $24.0B–$30.0B | Strong Confidence | Announced Private Round Valuation | Observed | $881M | Not provided |
| 9 | Etched | Transformer-specific AI inference chips | $21.0B | Full Confidence | Announced Private Round Valuation | Observed | ~$1.8B | Not provided |
| 10 | Fluidstack | Builds infrastructure for AI labs | $17.0B–$19.0B | Partial Confidence | Active Raise Valuation | Estimated | ~$871M | Not provided |
| 11 | Baseten | AI model inference infrastructure | $13.0B–$18.0B | Strong Confidence | Announced Private Round Valuation | Observed | $2.1B | Not provided |
| 12 | Nscale | Vertically integrated AI infrastructure cloud | $14.6B | Full Confidence | Announced Private Round Valuation | Observed | ~$8.5B | Not provided |
| 13 | Biren Technology | Data-center AI GPU accelerators | $13.8B | Full Confidence | Public Market Cap | Observed | $1.91B | Not provided |
| 14 | Lambda | GPU cloud and AI factories | $12.0B–$15.0B | Partial Confidence | Active Raise Valuation | Estimated | ~$5.6B+ | Not provided |
| 15 | Iluvatar CoreX | Chinese general-purpose GPU accelerators | $12.8B | Full Confidence | Public Market Cap | Observed | $334M | Not provided |
| 16 | SambaNova Systems | Enterprise AI chips and systems | $11.0B | Full Confidence | Announced Private Round Valuation | Observed | ~$2.48B | Not provided |
| 17 | Enflame Technology | Cloud AI training inference chips | $8.8B–$9.2B | Strong Confidence | IPO or Listing Range Valuation | Observed | $746M | Not provided |
| 18 | Tenstorrent | AI processors and RISC-V IP | $5.4B–$5.9B | Strong Confidence | Proxy-Based Estimate | Estimated | $1.84B | Not provided |
| 19 | Vultr | Independent global AI cloud platform | $4.0B–$5.5B | Partial Confidence | Comparables-Based Estimate | Estimated | ~$333M | Not provided |
| 20 | Modal | Serverless cloud for AI workloads | $4.7B | Full Confidence | Announced Private Round Valuation | Observed | $465M | Not provided |
| 21 | Lightmatter | Photonic AI data-center interconnects | $4.3B–$4.9B | Strong Confidence | Proxy-Based Estimate | Estimated | $822M | Not provided |
| 22 | Unconventional AI | Energy-efficient next-generation AI computers | $4.5B | Full Confidence | Announced Private Round Valuation | Observed | $475M | Not provided |
| 23 | Positron | Energy-efficient AI inference accelerators | $3.5B–$5.0B | Partial Confidence | Active Raise Valuation | Implied | $306M | Not provided |
| 24 | Lightelligence | Photonic computing and optical interconnects | $4.1B | Full Confidence | Public Market Cap | Observed | $247M | Not provided |
| 25 | Ayar Labs | Optical chip-to-chip AI interconnects | $3.8B | Full Confidence | Announced Private Round Valuation | Observed | $965M | Not provided |
| 26 | MatX | Accelerators for large AI models | $3.0B–$4.2B | Partial Confidence | Implied Valuation from Raise | Implied | $605M | Not provided |
| 27 | Celestial AI | Photonic fabric for AI clusters | $3.5B | Full Confidence | Acquisition Value | Observed | $594M | Not provided |
| 28 | Groq | High-speed AI inference cloud | $3.5B | Full Confidence | Announced Private Round Valuation | Observed | $4.26B | Not provided |
| 29 | FuriosaAI | Data-center AI inference accelerators | $2.2B–$3.0B | Partial Confidence | Active Raise Valuation | Estimated | ~$737M | Not provided |
| 30 | Rebellions | AI inference chips and systems | $2.3B | Full Confidence | Announced Private Round Valuation | Observed | $850M | Not provided |
| 31 | Axelera AI | Power-efficient edge AI processors | $1.8B–$2.8B | Partial Confidence | Implied Valuation from Raise | Estimated | $481M | Not provided |
| 32 | DEEPX | Edge AI inference processors | $2.1B–$2.3B | Strong Confidence | Announced Private Round Valuation | Observed | $128M | Not provided |
| 33 | Tachyum | Universal processors for AI datacenters | $1.5B–$2.8B | Partial Confidence | Implied Valuation from Raise | Implied | ~$291M | Not provided |
| 34 | d-Matrix | Data-center generative AI inference chips | $2.0B | Full Confidence | Announced Private Round Valuation | Observed | $450M | Not provided |
| 35 | MinIO | AI-focused object storage software | $1.4B–$2.4B | Partial Confidence | Revenue or ARR Multiple Estimate | Estimated | $126M | Not provided |
| 36 | Voltage Park | Large-scale GPU cloud infrastructure | $1.5B–$2.0B | Low Confidence | Acquisition Value | Estimated | ~$900M | Not provided |
| 37 | Anyscale | Distributed AI computing platform | $1.6B–$1.7B | Strong Confidence | Acquisition Value | Observed | $260M | Not provided |
| 38 | TensorWave | AMD-powered AI compute cloud | $1.6B | Full Confidence | Announced Private Round Valuation | Observed | ~$498M | Not provided |
| 39 | SiMa.ai | Physical and edge AI platforms | $1.5B–$1.7B | Strong Confidence | Announced Private Round Valuation | Implied | $732M | Not provided |
| 40 | Moffett AI | Sparse-computing AI accelerator chips | $1.0B–$1.8B | Partial Confidence | Implied Valuation from Raise | Implied | $173M | Not provided |
| 41 | Recogni (now Tensordyne) | AI inference accelerator systems | $1.2B–$1.5B | Strong Confidence | Announced Private Round Valuation | Observed | $259M | Not provided |
| 42 | WEKA | High-performance AI data platform | $1.0B–$1.6B | Partial Confidence | Revenue or ARR Multiple Estimate | Estimated | $415M | Not provided |
| 43 | DustPhotonics | Silicon-photonics optical connectivity chips | $1.3B | Full Confidence | Acquisition Value | Observed | $96M | Not provided |
| 44 | Qumulo | Enterprise AI data storage platform | $900M–$1.3B | Partial Confidence | Revenue or ARR Multiple Estimate | Estimated | $346M | Not provided |
| 45 | RunPod | Developer-focused GPU AI cloud | $1.0B | Full Confidence | Announced Private Round Valuation | Observed | ~$120M | Not provided |
| 46 | Hammerspace | AI data orchestration platform | $800M–$1.2B | Partial Confidence | Revenue or ARR Multiple Estimate | Estimated | $157M | Not provided |
| 47 | Enfabrica | AI cluster networking silicon | $900M–$1.0B | Strong Confidence | Acquisition Value | Estimated | $260M | Not provided |
| 48 | Kneron | Edge AI chips and software | $800M–$1.1B | Partial Confidence | Comparables-Based Estimate | Estimated | ~$236M | Not provided |
| 49 | Axiado | AI data-center security processors | $800M–$1.1B | Partial Confidence | Implied Valuation from Raise | Implied | $220M | Not provided |
| 50 | DataCrunch / Verda | European sovereign AI cloud | $700M–$1.0B | Partial Confidence | Revenue or ARR Multiple Estimate | Estimated | ~$219M | Not provided |
| 51 | Ori | GPU cloud and AI infrastructure | $700M–$1.0B | Partial Confidence | Acquisition Value | Estimated | $145M | Not provided |
| 52 | Extropic | Thermodynamic processors for probabilistic AI | $600M–$1.0B | Low Confidence | Proxy-Based Estimate | Estimated | $14M | Not provided |
| 53 | AttoTude | Dielectric AI interconnect technology | $700M–$850M | Strong Confidence | Announced Private Round Valuation | Implied | $143M | Not provided |
| 54 | WhiteFiber | GPU cloud and data centers | $657M | Full Confidence | Public Market Cap | Observed | $183M | Not provided |
| 55 | Mesh Optical Technologies | AI data-center optical transceivers | $580M–$650M | Partial Confidence | Acquisition Value | Implied | $50M | Not provided |
| 56 | Avicena | MicroLED optical chip interconnects | $450M–$650M | Partial Confidence | Implied Valuation from Raise | Implied | $120M | Not provided |
| 57 | EnCharge AI | Analog in-memory AI accelerators | $430M–$600M | Partial Confidence | Proxy-Based Estimate | Estimated | ~$163M | Not provided |
| 58 | Lepton AI | GPU cloud and inference platform | $300M–$700M | Partial Confidence | Acquisition Value | Estimated | $11M | Not provided |
| 59 | Cornami | FHE and parallel computing accelerators | $450M–$550M | Partial Confidence | Implied Valuation from Raise | Implied | ~$151M+ | Not provided |
| 60 | NexGen Cloud | European GPU cloud infrastructure | $400M–$550M | Partial Confidence | Comparables-Based Estimate | Estimated | ~$93M | Not provided |
| 61 | GMI Cloud | Enterprise GPU cloud infrastructure | $350M–$500M | Partial Confidence | Proxy-Based Estimate | Estimated | ~$93M | Not provided |
| 62 | Graphcore | Intelligence processors for AI workloads | $350M–$500M | Low Confidence | Acquisition Value | Estimated | $682M | Not provided |
| 63 | Ethernovia | Ethernet chips for physical AI | $390M–$430M | Strong Confidence | Announced Private Round Valuation | Observed | $156M | Not provided |
| 64 | EdgeQ | Programmable 5G and AI chips | $300M–$450M | Partial Confidence | Comparables-Based Estimate | Estimated | $126M | Not provided |
| 65 | SF Compute | Marketplace for GPU compute | $300M–$400M | Strong Confidence | Announced Private Round Valuation | Estimated | $55M | Not provided |
| 66 | Lightbits Labs | Software-defined NVMe cloud storage | $200M–$400M | Partial Confidence | Comparables-Based Estimate | Estimated | $102M | Not provided |
| 67 | Kinara | Edge AI neural processing units | $284M | Full Confidence | Acquisition Value | Observed | ~$54M | Not provided |
| 68 | Quadric | Programmable edge AI processor IP | $250M–$300M | Strong Confidence | Implied Valuation from Raise | Implied | ~$90M | Not provided |
| 69 | Alluxio | Accelerates AI data access | $200M–$300M | Partial Confidence | Comparables-Based Estimate | Estimated | $74M | Not provided |
| 70 | Esperanto Technologies | RISC-V AI accelerator processors | $180M–$320M | Low Confidence | Proxy-Based Estimate | Estimated | ~$124M | Not provided |
| 71 | Prophesee | Event-based neuromorphic vision sensors | $180M–$280M | Partial Confidence | Implied Valuation from Raise | Implied | $153M | Not provided |
| 72 | Hailo | Edge AI acceleration processors | $150M–$300M | Low Confidence | Proxy-Based Estimate | Estimated | $451M | Not provided |
| 73 | Xscape Photonics | Photonics for AI data fabrics | $200M–$240M | Strong Confidence | Announced Private Round Valuation | Implied | $94M | Not provided |
| 74 | Parasail | Distributed AI inference cloud | $160M–$270M | Partial Confidence | Implied Valuation from Raise | Implied | ~$42M | Not provided |
| 75 | BrainChip | Neuromorphic edge AI processors | $210M | Full Confidence | Public Market Cap | Observed | $0.7M | Not provided |
| 76 | Salience Labs | Photonic switches for AI clusters | $160M–$230M | Partial Confidence | Implied Valuation from Raise | Implied | $42M | Not provided |
| 77 | MemryX | Low-power edge AI accelerators | $150M–$220M | Partial Confidence | Comparables-Based Estimate | Estimated | $51M | Not provided |
| 78 | Luminous Computing | Photonics-based AI supercomputing hardware | $120M–$230M | Low Confidence | Proxy-Based Estimate | Estimated | ~$126M | Not provided |
| 79 | NeuReality | AI inference networking and infrastructure | $120M–$210M | Low Confidence | Implied Valuation from Raise | Estimated | ~$70M | Not provided |
| 80 | Mythic | Analog AI inference processors | $130M–$180M | Partial Confidence | Announced Private Round Valuation | Implied | $297M | Not provided |
| 81 | Vast.ai | Marketplace for distributed GPU compute | $100M–$180M | Low Confidence | Revenue or ARR Multiple Estimate | Estimated | ~$4M | Not provided |
| 82 | io.net | Decentralized GPU compute network | $100M–$180M | Partial Confidence | Proxy-Based Estimate | Estimated | ~$40M | Not provided |
| 83 | Rain Neuromorphics (Rain AI) | Neuromorphic AI accelerator chips | $100M–$180M | Low Confidence | Proxy-Based Estimate | Estimated | ~$146M | Not provided |
| 84 | Flex Logix | Embedded FPGA and AI IP | $100M–$180M | Low Confidence | Acquisition Value | Estimated | $84M | Not provided |
| 85 | Tigris Data | Distributed object storage for AI | $120M–$160M | Strong Confidence | Implied Valuation from Raise | Implied | $32M | Not provided |
| 86 | Oriole Networks | Photonic networking for AI systems | $100M–$140M | Partial Confidence | Comparables-Based Estimate | Estimated | $35M | Not provided |
| 87 | Hyperbolic | Open-access distributed GPU cloud | $90M–$140M | Partial Confidence | Comparables-Based Estimate | Estimated | $20M | Not provided |
| 88 | Ambient Scientific | Ultra-low-power edge AI processors | $80M–$140M | Partial Confidence | Active Raise Valuation | Implied | $51M | Not provided |
| 89 | Volumez | Cloud data infrastructure orchestration | $80M–$120M | Low Confidence | Comparables-Based Estimate | Estimated | $53M | Not provided |
| 90 | Ephos | Glass-based photonic computing chips | $75M–$120M | Partial Confidence | Comparables-Based Estimate | Estimated | $18M | Not provided |
| 91 | Lucidean | Coherent data-center optical interconnects | $75M–$110M | Partial Confidence | Implied Valuation from Raise | Estimated | $18M | Not provided |
| 92 | Hyperlume | MicroLED optical chip interconnects | $92M | Full Confidence | Acquisition Value | Observed | $25M | Not provided |
| 93 | Pliops | Data-processing storage accelerators | $65M–$75M | Strong Confidence | Acquisition Value | Estimated | $210M | Not provided |
| 94 | Blaize | Edge AI compute processors | $69M | Full Confidence | Public Market Cap | Observed | ~$246M | Not provided |
| 95 | Cortical Labs | Biological computers using human neurons | $45M–$80M | Partial Confidence | Comparables-Based Estimate | Estimated | $11M | Not provided |
| 96 | Untether AI | Energy-efficient AI inference chips | $20M–$60M | Low Confidence | Proxy-Based Estimate | Estimated | ~$168M | Not provided |
| 97 | Solidus Ai Tech (ACN) | Tokenized GPU and AI cloud | $25M–$38M | Low Confidence | Proxy-Based Estimate | Estimated | $9M | Not provided |
| 98 | Gyrfalcon Technology | Low-power edge AI accelerators | $15M–$35M | Low Confidence | Proxy-Based Estimate | Estimated | $1.5M | Not provided |
| 99 | NovuMind | AI inference processor intellectual property | $15M–$30M | Low Confidence | Proxy-Based Estimate | Estimated | $16M | Not provided |
| 100 | Genesis Cloud | European GPU cloud services | $10M–$30M | Low Confidence | Acquisition Value | Estimated | ~$4M | Not provided |
This chart, included in our AI infrastructure market deck, compares the 2026 size of the AI infrastructure market with other markets of similar size
Key valuation trends in the AI infrastructure market
Insights
- The AI infrastructure market is highly concentrated. The top 10 companies represent about $410.9 billion of value at valuation midpoints, roughly 69% of the approximately $595.6 billion represented by all 100 companies.
- Chinese AI chip companies have reached enormous public valuations relative to disclosed funding. Cambricon, MetaX, Moore Threads, Biren, Iluvatar CoreX, and Lightelligence represent about $210 billion of value on roughly $4.1 billion of disclosed historical funding.
- AI infrastructure valuations vary dramatically even at similar funding levels. DEEPX is valued around $2.1 billion to $2.3 billion after raising $128 million, while Rebellions reached a similar valuation after raising $850 million.
- Optical connectivity has become a meaningful AI infrastructure category of its own. Lightmatter, Ayar Labs, Celestial AI, Lightelligence, DustPhotonics, Enfabrica, Xscape Photonics, and Hyperlume together represent more than $18 billion of value.
- Strategic buyers are paying substantial prices for technologies that remove AI-cluster bottlenecks. Celestial AI, DustPhotonics, Enfabrica, and Hyperlume generated roughly $5.8 billion of combined acquisition value.
- AI infrastructure business models have very different capital requirements. Nscale is valued at $14.6 billion after roughly $8.5 billion of funding, while Modal reached $4.7 billion after raising only $465 million.
- Alternative AI processor architectures are now a major investment category. Cerebras, Etched, and SambaNova alone represent roughly $73 billion of combined valuation, separate from conventional GPU infrastructure providers.
- Developer-focused AI infrastructure can reach large valuations without hyperscale financing. RunPod reached a $1 billion valuation after roughly $120 million of funding, showing the potential economics of software-led GPU platforms.
This chart, included in our AI infrastructure market deck, shows why CoreWeave is winning in AI infrastructure
A few words about our methodology
As you can see, we built a database that ranks startups in the AI infrastructure market based on their current valuation.
Estimating valuations in AI infrastructure is not always straightforward. The market includes public semiconductor companies, private GPU clouds, data infrastructure platforms, networking specialists, and early-stage hardware companies. Many companies do not publicly disclose their valuation.
To build this ranking, we applied a structured valuation methodology and cross-checked information across multiple reliable sources.
Whenever possible, we relied on direct disclosures. These include announced valuations from completed funding rounds, public filings for listed companies, or official acquisition prices.
When an AI infrastructure company is publicly listed, we use its current market capitalization as the reference valuation.
If a company was acquired and no independent valuation can reasonably be estimated today, we use the acquisition price as the main reference point.
When an AI infrastructure startup recently raised capital but the valuation was not disclosed, we estimate the implied valuation using typical dilution levels for that stage of fundraising.
In some cases, we also estimate valuations using operating metrics such as revenue, ARR, cloud capacity, customer traction, or commercial deployment, combined with valuation multiples from comparable AI infrastructure companies.
When direct financial data is not available, we may rely on carefully selected comparable companies and other signals such as hiring growth, investor quality, product traction, infrastructure deployment, or technical adoption.
Hardware and cloud infrastructure companies can require very different amounts of capital. We therefore do not treat cumulative funding as a direct proxy for valuation.
All estimates follow a strict evidence hierarchy. Recent funding rounds with announced valuations carry the most weight, followed by public-market data, acquisition values, strong operating metrics, and comparable company analysis.
We also carefully evaluate the age of every data point. Recent information carries more weight, while older data is treated cautiously and adjusted conservatively when necessary.
Whenever information is uncertain or incomplete, we clearly distinguish between confirmed facts and reasonable inferences.
Because valuation data is not always fully public, each company in the AI infrastructure ranking is assigned a valuation confidence level based on the reliability, recency, and consistency of the available evidence.
Full confidence means the valuation is supported by strong and recent evidence. Strong confidence means the estimate is well supported but includes minor inference. Partial confidence means the estimate relies more heavily on indirect signals. Low confidence means available information is limited or inconsistent.
When confidence is lower, we take a more conservative approach by widening the valuation range. This helps reflect the uncertainty and increases the probability that the true valuation falls within the estimated range.
The supplied source tables did not include a separate funding confidence level. For that reason, we show funding confidence as "Not provided" rather than creating a label that is not supported by the underlying dataset.
This reflects how we conduct all our research, including the work behind our report covering the AI infrastructure market.
In a world where LLMs hallucinate and unreliable information is everywhere, our goal is simple: provide data you can trust.
If you want the full detail on a specific valuation estimate, feel free to contact us and we will gladly explain.
Finally, know that we update the dataset once per month, so come back here if you need fresh information.
This chart, included in our AI infrastructure market deck, shows annual funding in AI infrastructure startups
Related blog posts
- Is TensorWave really worth $1.55B?
- AI Infrastructure: what are the top startups now?
- Which startups have raised the most funding in AI infrastructure?
