CS

Cerebras Systems

CBRS

Hardware

AI chip company building wafer-scale processors for fast AI training and inference.

Supplied by: TSMC

Supplies OpenRouter

Total funding—

Market cap$44.4B

Employees708

Founded2015

HeadquartersSunnyvale, USA

StatusPublic

Products

Selected offerings — the AI products recorded here, not the vendor's full catalogue.

• Cerebras Inference Model API Sold on its own

  Hosted inference service that serves open-weight models on Cerebras wafer-scale hardware, billed per token.

• Cerebras Training Cloud Platform Sold on its own

  Managed training/fine-tuning cloud for models from 1B to tens-of-trillions of parameters without manual sharding or model-parallelism configuration, billed pay-per-hour or pay-per-model (Cerebras-run custom training).

Inside Cerebras CS-4

• Cerebras WSE-3 Turbo (WSE-3T) Hardware Through partners

  Wafer-scale processor — now WSE-3 Turbo (WSE-3T), four trillion transistors and 900,000 AI cores — sold inside Cerebras' rack-scale systems: three WSE-3T processors power the CS-4, the successor to the WSE-3-based CS-3.

Closest alternatives 3

NNVIDIAPublic · $5.1TCerebras' own blog claims its WSE3 chip outperforms Nvidia's Blackwell generation by close to 5x on the same inference workload, a direct compute-purchase claim.3 of 3 capabilitiesShared product typeLarger scaleSourced competitorBuy data-centre AI computeCompareSSambaNovaPrivate · $11BSacra names Cerebras as a specialized-silicon rival to SambaNova in AI accelerator hardware. SambaNova's RDU chip and Cerebras's WSE chip both carry buy-ai-compute.2 of 3 capabilitiesShared product typeSmaller scaleSourced competitorBuy data-centre AI computeCompareGGroqPrivate · $3.5BContrary Research's Cerebras company report, Competition section -- names Groq as Cerebras's direct rival in inference. Reciprocally corroborated by Reuters (republished on investing.com, 2025-12-24): "Groq's primary rival in the approach is Cerebras Systems."2 of 3 capabilitiesShared product typeSmaller scaleSourced competitorBuild on a hosted model APIServe a model in productionCompare

Similar products 71

How this list is ranked ↗

IIntelPublic · $463.7BCerebras WSE-3 Turbo (WSE-3T) and Intel Gaudi AI Accelerator do comparable work on accelerator silicon and model training; both also serve buyers who need to buy data-centre AI compute; larger scale (public).3 of 3 capabilitiesShared product typeLarger scaleCompareAAMDPublic · $785.1BCerebras WSE-3 Turbo (WSE-3T) and AMD Instinct do comparable work on accelerator silicon and model training; both also serve buyers who need to buy data-centre AI compute; larger scale (public).3 of 3 capabilitiesShared product typeLarger scaleCompareAAAxelera AIPrivate · $450M raisedCerebras WSE-3 Turbo (WSE-3T) and Axelera Metis AIPU do comparable work on accelerator silicon; both also serve buyers who need to buy data-centre AI compute; similar scale (private).2 of 3 capabilitiesShared product typeCompareQQualcommPublic · $174.9BCerebras WSE-3 Turbo (WSE-3T) and Qualcomm Cloud AI 100 Ultra do comparable work on accelerator silicon; both also serve buyers who need to buy data-centre AI compute; larger scale (public).2 of 3 capabilitiesShared product typeLarger scaleCompareHHuaweiPrivateCerebras WSE-3 Turbo (WSE-3T) and Atlas do comparable work on accelerator silicon and model training; both also serve buyers who need to buy data-centre AI compute; larger scale (private).3 of 3 capabilitiesShared product typeCompareDTDell TechnologiesPublic · $360.7BCerebras WSE-3 Turbo (WSE-3T) and Dell Pro Max AI Desktop do comparable work on model training; both also serve buyers who need to buy data-centre AI compute; larger scale (public).2 of 3 capabilitiesShared product typeLarger scaleCompareGGraphcorePrivateCerebras WSE-3 Turbo (WSE-3T) and IPU (Intelligence Processing Unit) do comparable work on accelerator silicon and model training; both also serve buyers who need to buy data-centre AI compute; Graphcore's scale not recorded.3 of 3 capabilitiesShared product typeCompareTTenstorrentPrivate · $2.6BCerebras WSE-3 Turbo (WSE-3T) and Blackhole do comparable work on accelerator silicon; both also serve buyers who need to buy data-centre AI compute; smaller scale (private).2 of 3 capabilitiesShared product typeSmaller scaleCompareBBasetenPrivate · $13BCerebras Training Cloud and Baseten Training do comparable work on model training; both also serve buyers who need to train or fine-tune a model; smaller scale (private).2 of 3 capabilitiesShared product typeSmaller scaleCompareTATogether AIPrivate · $8.3BCerebras Inference and Together Inference both serve buyers who need to build on a hosted model API and serve a model in production; smaller scale (private).2 of 3 capabilitiesShared product typeSmaller scaleCompareBBroadcomPublic · $1.7TCerebras WSE-3 Turbo (WSE-3T) and Broadcom custom silicon do comparable work on accelerator silicon; both also serve buyers who need to buy data-centre AI compute; larger scale (public).1 of 3 capabilitiesShared product typeLarger scaleCompareFAFireworks AIPrivate · $17.5BCerebras Inference and Fireworks Nexus both serve buyers who need to build on a hosted model API and serve a model in production; smaller scale (private).2 of 3 capabilitiesShared product typeSmaller scaleCompareHRHorizon RoboticsPublic · $8.8BCerebras WSE-3 Turbo (WSE-3T) and Journey 6 Series do comparable work on accelerator silicon; both also serve buyers who need to buy data-centre AI compute; smaller scale (public).2 of 3 capabilitiesShared product typeSmaller scaleCompareCCiscoPublic · $433.3BCerebras WSE-3 Turbo (WSE-3T) and Cisco AI PODs do comparable work on model training; both also serve buyers who need to buy data-centre AI compute; larger scale (public).3 of 3 capabilitiesShared product typeLarger scaleCompareAAmazonPublic · $2.8TCerebras WSE-3 Turbo (WSE-3T) and AWS Trainium do comparable work on accelerator silicon and model training; both also serve buyers who need to buy data-centre AI compute; larger scale (public).3 of 3 capabilitiesShared product typeLarger scaleCompareAMApplied MaterialsPublic · $380.7BCerebras Training Cloud and SmartFactory AI IQworks do comparable work on model training; both also serve buyers who need to train or fine-tune a model; larger scale (public).1 of 3 capabilitiesShared product typeLarger scaleCompareCCCactus ComputePrivateCerebras Training Cloud and Cactus Platform do comparable work on model training; both also serve buyers who need to train or fine-tune a model; Cactus Compute's scale not recorded.2 of 3 capabilitiesShared product typeCompareSHSK HynixPublic · $874.5BCerebras WSE-3 Turbo (WSE-3T) and HBM3E do comparable work on model training; both also serve buyers who need to buy data-centre AI compute; larger scale (public).2 of 3 capabilitiesShared product typeLarger scaleCompareALAI21 LabsPrivate · $1.4BCerebras Training Cloud and AI21 Post-Training do comparable work on model training; both also serve buyers who need to train or fine-tune a model; smaller scale (private).2 of 3 capabilitiesShared product typeSmaller scaleCompareHTHailo TechnologiesAcquired ($340M deal)Cerebras WSE-3 Turbo (WSE-3T) and Hailo AI accelerators do comparable work on accelerator silicon; Hailo Technologies is acquired, with no independent scale.2 of 3 capabilitiesShared product typeCompareMTMarvell TechnologyPublic · $214.9BCerebras WSE-3 Turbo (WSE-3T) and Marvell custom AI silicon do comparable work on accelerator silicon; both also serve buyers who need to buy data-centre AI compute; larger scale (public).1 of 3 capabilitiesShared product typeLarger scaleCompareDDatabricksPrivate · $190BCerebras Training Cloud and Databricks Model Training do comparable work on model training; both also serve buyers who need to train or fine-tune a model; larger scale (private).2 of 3 capabilitiesShared product typeLarger scaleCompareCCrusoePrivate · $10BCerebras Training Cloud and Serverless Fine-Tuning do comparable work on model training; both also serve buyers who need to train or fine-tune a model; smaller scale (private).2 of 3 capabilitiesShared product typeSmaller scaleCompareAAlphabet (Google)Public · $4.2TCerebras WSE-3 Turbo (WSE-3T) and Google Cloud TPU do comparable work on accelerator silicon and model training; both also serve buyers who need to buy data-centre AI compute; larger scale (public).3 of 3 capabilitiesShared product typeLarger scaleCompareNGNebius GroupPublic · $61.6BCerebras Inference and Nebius Token Factory both serve buyers who need to build on a hosted model API and serve a model in production; same layer and the same scale band (mid-cap).2 of 3 capabilitiesShared product typeSame scale bandCompareLALiquid AIPrivateCerebras Training Cloud and LEAP do comparable work on model training; both also serve buyers who need to train or fine-tune a model; Liquid AI's scale not recorded.2 of 3 capabilitiesShared product typeCompareKEKakao EnterprisePrivateCerebras Training Cloud and Kubeflow do comparable work on model training; both also serve buyers who need to train or fine-tune a model; Kakao Enterprise's scale not recorded.2 of 3 capabilitiesShared product typeCompareTTolokaPrivateCerebras Training Cloud and Toloka Train do comparable work on model training; both also serve buyers who need to train or fine-tune a model; Toloka's scale not recorded.1 of 3 capabilitiesShared product typeCompareLLambdaPrivate · $1.5B raisedCerebras Inference and Lambda Inference both serve buyers who need to build on a hosted model API and serve a model in production; similar scale (private).2 of 3 capabilitiesShared product typeCompare44ParadigmPublic · $2BCerebras Training Cloud and 4Paradigm Sage HyperCycle do comparable work on model training; both also serve buyers who need to train or fine-tune a model; smaller scale (public).2 of 3 capabilitiesShared product typeSmaller scaleCompareMAMistral AIPrivate · $13.8BCerebras Training Cloud and Mistral Forge do comparable work on model training; both also serve buyers who need to train or fine-tune a model; smaller scale (private).1 of 3 capabilitiesShared product typeSmaller scaleCompareAAlibabaPublic · $273.4BCerebras Inference and Alibaba Cloud Model Studio both serve buyers who need to build on a hosted model API and serve a model in production; larger scale (public).2 of 3 capabilitiesShared product typeLarger scaleCompareHCHuawei CloudPrivateCerebras Inference and MaaS both serve buyers who need to build on a hosted model API and serve a model in production; Huawei Cloud's scale not recorded.2 of 3 capabilitiesShared product typeCompareMModalPrivate · $466M raisedCerebras Training Cloud and Modal Notebooks do comparable work on model training; both also serve buyers who need to train or fine-tune a model; similar scale (private); ships developer tool and infrastructure service rather than the same layer.2 of 3 capabilitiesDifferent layerCompareAHArm HoldingsPublic · $282.8BCerebras WSE-3 Turbo (WSE-3T) and Arm CSS for Mobile 2 do comparable work on accelerator silicon; larger scale (public).2 of 3 capabilitiesShared product typeLarger scaleCompareW&Weights & BiasesAcquiredCerebras Training Cloud and W&B Models do comparable work on model training; both also serve buyers who need to train or fine-tune a model; Weights & Biases is acquired, with no independent scale.2 of 3 capabilitiesShared product typeCompareCMComet MLPrivate · $70M raisedCerebras Training Cloud and Comet Experiment Tracking do comparable work on model training; both also serve buyers who need to train or fine-tune a model; similar scale (private).1 of 3 capabilitiesShared product typeCompareHFHugging FacePrivate · $4.5BCerebras Inference and Inference Endpoints both serve buyers who need to serve a model in production; smaller scale (private).2 of 3 capabilitiesShared product typeSmaller scaleCompareRRunPodPrivateCerebras Inference and Public Endpoints both serve buyers who need to build on a hosted model API and serve a model in production; RunPod's scale not recorded.2 of 3 capabilitiesShared product typeCompareEEleutherAINonprofitCerebras Training Cloud and GPT-NeoX do comparable work on model training; both also serve buyers who need to train or fine-tune a model; EleutherAI is a nonprofit, with no independent scale; ships developer tool rather than the same layer.1 of 3 capabilitiesDifferent layerCompareOOraclePublic · $454.4BCerebras Inference and Generative AI Capabilities both serve buyers who need to build on a hosted model API and serve a model in production; larger scale (public).2 of 3 capabilitiesShared product typeLarger scaleCompareDDecartPrivate · $4BCerebras Inference and Cogito both serve buyers who need to build on a hosted model API and serve a model in production; smaller scale (private).2 of 3 capabilitiesShared product typeSmaller scaleCompareTMThinking Machines LabPrivate · $12BCerebras Training Cloud and Tinker do comparable work on model training; both also serve buyers who need to train or fine-tune a model; smaller scale (private); ships API service rather than the same layer.2 of 3 capabilitiesDifferent layerSmaller scaleCompareNNeo4jPrivate · $2BCerebras Training Cloud and Neo4j Graph Data Science do comparable work on model training; both also serve buyers who need to train or fine-tune a model; smaller scale (private).1 of 3 capabilitiesShared product typeSmaller scaleCompareLLabelboxPrivate · $189M raisedCerebras Training Cloud and Horizon do comparable work on model training; both also serve buyers who need to train or fine-tune a model; similar scale (private).2 of 3 capabilitiesShared product typeCompareBBaiduPublic · $31.6BCerebras Inference and Baidu Qianfan both serve buyers who need to build on a hosted model API and serve a model in production; same layer and the same scale band (mid-cap).2 of 3 capabilitiesShared product typeSame scale bandCompareTTuringPrivate · $2.2BCerebras Training Cloud and RL environments do comparable work on model training; both also serve buyers who need to train or fine-tune a model; smaller scale (private); ships agent platform and data service rather than the same layer.1 of 3 capabilitiesDifferent layerSmaller scaleCompareUUnslothPrivateCerebras Training Cloud and Unsloth do comparable work on model training; both also serve buyers who need to train or fine-tune a model; Unsloth's scale not recorded; ships developer tool rather than the same layer.2 of 3 capabilitiesDifferent layerCompareAAppenPublicCerebras Training Cloud and AI Data Platform do comparable work on model training; both also serve buyers who need to train or fine-tune a model; Appen's scale not recorded.1 of 3 capabilitiesShared product typeCompareSSiemensPublic · $244.7BCerebras Training Cloud and Rapidminer AI Studio do comparable work on model training; both also serve buyers who need to train or fine-tune a model; larger scale (public).2 of 3 capabilitiesShared product typeLarger scaleCompareSASnorkel AIPrivate · $1.3BCerebras Training Cloud and Snorkel Flow do comparable work on model training; both also serve buyers who need to train or fine-tune a model; smaller scale (private).1 of 3 capabilitiesShared product typeSmaller scaleCompareMMicrosoftPublic · $3.7TCerebras Training Cloud and Azure Machine Learning do comparable work on model training; both also serve buyers who need to train or fine-tune a model; larger scale (public).2 of 3 capabilitiesShared product typeLarger scaleCompareKKuaishouPublic · $18.6BCerebras Training Cloud and Vanchin do comparable work on model training; both also serve buyers who need to train or fine-tune a model; same layer and the same scale band (mid-cap).2 of 3 capabilitiesShared product typeSame scale bandCompareNCNAVER CloudPrivateCerebras Training Cloud and NCLUE do comparable work on model training; both also serve buyers who need to train or fine-tune a model; NAVER Cloud's scale not recorded.2 of 3 capabilitiesShared product typeCompareBFBlack Forest LabsPrivate · $3.3BCerebras Inference and FLUX API both serve buyers who need to build on a hosted model API; smaller scale (private).1 of 3 capabilitiesShared product typeSmaller scaleCompareTTencentPublic · $510.1BCerebras Training Cloud and Tencent Cloud TI Platform do comparable work on model training; both also serve buyers who need to train or fine-tune a model; larger scale (public).2 of 3 capabilitiesShared product typeLarger scaleCompareIMInsilico MedicinePublic · $4.3BCerebras Training Cloud and Biology42: Generative Biologics do comparable work on model training; both also serve buyers who need to train or fine-tune a model; smaller scale (public).1 of 3 capabilitiesShared product typeSmaller scaleCompareDDataikuPrivate · $3.7BCerebras Training Cloud and Dataiku do comparable work on model training; both also serve buyers who need to train or fine-tune a model; smaller scale (private).2 of 3 capabilitiesShared product typeSmaller scaleCompareIIBMPublic · $216.6BCerebras Training Cloud and IBM watsonx.ai do comparable work on model training; both also serve buyers who need to train or fine-tune a model; larger scale (public).2 of 3 capabilitiesShared product typeLarger scaleCompareAASMLPublic · $670.1BCerebras WSE-3 Turbo (WSE-3T) and ASML DUV lithography systems both serve buyers who need to buy data-centre AI compute; larger scale (public).Shared product typeLarger scaleCompareKKioxiaPublic · $177.9BCerebras WSE-3 Turbo (WSE-3T) and KIOXIA CM9-V Series both serve buyers who need to buy data-centre AI compute; larger scale (public).Shared product typeLarger scaleCompareAAdobePublic · $115.9BCerebras Training Cloud and Adobe Firefly Custom Models do comparable work on model training; both also serve buyers who need to train or fine-tune a model; larger scale (public).2 of 3 capabilitiesShared product typeLarger scaleCompareYYandexPublicCerebras Training Cloud and Yandex DataSphere do comparable work on model training; both also serve buyers who need to train or fine-tune a model; Yandex's scale not recorded.2 of 3 capabilitiesShared product typeCompareSESamsung ElectronicsPublic · $1.2TCerebras WSE-3 Turbo (WSE-3T) and Samsung Datacenter SSD both serve buyers who need to buy data-centre AI compute; larger scale (public).Shared product typeLarger scaleCompareSSenseTimePublic · $7.7BCerebras WSE-3 Turbo (WSE-3T) and SenseCore do comparable work on model training; both also serve buyers who need to buy data-centre AI compute; smaller scale (public).2 of 3 capabilitiesShared product typeSmaller scaleCompareSSpace42Public · $2.6BCerebras Training Cloud and GIQ do comparable work on model training; smaller scale (public).1 of 3 capabilitiesShared product typeSmaller scaleCompareOOllamaPrivate · $88M raisedCerebras Inference and Ollama Cloud both serve buyers who need to build on a hosted model API and serve a model in production; similar scale (private).1 of 3 capabilitiesShared product typeCompareHHumainPrivateCerebras WSE-3 Turbo (WSE-3T) and HUMAIN COMPUTE do comparable work on model training; both also serve buyers who need to buy data-centre AI compute; Humain's scale not recorded; ships API service, agent platform and 1 more rather than the same layer.2 of 3 capabilitiesDifferent layerCompareANArista NetworksPublic · $246.4BCerebras WSE-3 Turbo (WSE-3T) and 7060X5 Series both serve buyers who need to buy data-centre AI compute; larger scale (public).Shared product typeLarger scaleCompareNRNous ResearchPrivate · $70M raisedCerebras Inference and Nous Portal both serve buyers who need to build on a hosted model API; similar scale (private); ships AI agent, API service and 1 more rather than the same layer.1 of 3 capabilitiesDifferent layerCompareRRunwayPrivate · $5.3BCerebras Inference and Runway API both serve buyers who need to build on a hosted model API; smaller scale (private).1 of 3 capabilitiesShared product typeSmaller scaleCompare

Closest alternatives are pairs a vendor page itself compares. Similar products are matched from what each product does; nothing in that group is asserted as a rivalry.

Sources

Total funding

Checked — none published

Aggregators disagree widely on Cerebras' pre-IPO cumulative total ($1.9B to $3.4B across sources), and neither its own Series G (Sep 2025) nor Series H (Feb 2026) press release states a cumulative figure. Left null rather than guessed. Cerebras Systems went public on Nasdaq as CBRS on 2026-05-14, an IPO that raised $5.55B.

Market cap

Third-party database · 27 Aug 2026

Market cap $44.35B as of 2026-08-27, quoted natively in USD — no fx conversion. Replaces the stale pre-IPO $8.1B Series G valuation from September 2025, which predates the 2026-05-14 IPO entirely.

Employees

News reporting · Jun 2026

Forbes company profile for Cerebras Systems, "Employees" stated as of June 2026. Read 2026-08-27.

Founded

Wikipedia · 27 Aug 2026

Wikipedia infobox "founded" field, read 2026-08-27.

Public listing

Third-party database · 27 Aug 2026

Cerebras Systems IPO'd on Nasdaq as CBRS on 2026-05-14. Unlike the three HKEX listings, this is a straightforward USD-quoted Nasdaq listing with no currency conversion needed.

Open Cerebras Systems in the directory

Something wrong here? Send a correction — quote this company id: cerebras-systems.