Startups > Blaize

Blaize

Blaize powers edge AI inference at 60% lower power than cloud solutions.

Series D $136M total Founded 2010 El Dorado Hills, California 300 employees

Blaize provides a full-stack edge AI platform combining purpose-built hardware (GSP-based processors delivering 16 TOPS inference at 7W power) with a no-code/low-code software suite (AI Studio) for rapid deployment. The platform enables real-time AI inference at the network edge with 60% greater efficiency and 10x lower latency than cloud-optimized alternatives. Serving automotive, surveillance, retail, and industrial automation verticals, Blaize eliminates the complexity of deploying AI on power- and thermally-constrained devices.

Problem solved

Edge AI deployment requires extreme power efficiency and low latency, but existing cloud-optimized AI solutions waste power and bandwidth when deployed on constrained devices.

Target customer

Automotive OEMs, surveillance infrastructure operators, industrial automation providers, and smart city integrators requiring low-latency, power-efficient AI at the edge.

Website LinkedIn Crunchbase Twitter / X

Founders

Dinakar Munagala

CEO & Co-Founder

22+ years leading graphics chip development teams; spent 12 years at Intel leading GPU microarchitecture and design.

Satyaki Koneru

CTO & Co-Founder

20+ years in system-on-chip design; held architect and design roles at Intel and NVIDIA.

Ke Yin

Co-Founder

12 years at Intel as architect and design engineer; led power reduction efforts on Intel Gen graphics achieving 30%+ power savings.

Funding history

Series A Unknown October 6, 2016 Led by Magna · Unknown

Series D Unknown July 26, 2021 Led by Franklin Templeton Investments · Unknown

Post-IPO Equity $106M April 29, 2024 Led by Unknown · Unknown

Post-IPO Private Placement (PIPE) $30M November 11, 2025 Led by Polar Asset Management Partners · Unknown

Total raised: $136M

Industries

Artificial Intelligence (AI) Computer Vision Information Technology Service Industry

Pricing

Custom pricing based on deployment requirements. Enterprise-level hardware and software platform with no publicly disclosed pricing; contact sales for quotes.

Notable customers

DENSO, Mercedes-Benz, Winmate Inc., Southeast Asia national-scale surveillance infrastructure operator (250,000+ endpoints), Thrive Logic (integration partner)

Integrations

Thrive Logic, Winmate Inc., NeoTensr

Tech stack

jQuery Migrate (JavaScript libraries) jQuery (JavaScript libraries) FancyBox (JavaScript libraries) Bootstrap (UI frameworks) Wistia (Video players) Popper RSS Open Graph HTTP/3 WordPress (Blogs) ShareThis (Widgets) HubSpot Analytics (Analytics) Matomo Analytics (Analytics) Linkedin Insight Tag (Analytics) Google Analytics (Analytics) Cloudflare Bot Management (Security) Twitter Emoji (Font scripts) PHP (Programming languages) cdnjs (CDN) Cloudflare (CDN) HubSpot (Marketing automation) MySQL (Databases) Google Tag Manager (Tag managers) Elementor (Page builders) Yoast SEO (SEO) WP Engine (PaaS) GoDaddy (Hosting)

Website

blaize.com/

Competitors

NVIDIA

NVIDIA dominates centralized data center AI; Blaize specializes in power-efficient edge inference with 60-70% lower power consumption.

Groq

Groq focuses on high-throughput inference acceleration; Blaize prioritizes extreme power efficiency and thermal constraints for edge deployment.

Hailo

Hailo offers edge AI accelerators; Blaize differentiates with a full-stack solution including proprietary hardware plus integrated no-code software platform.

Tenstorrent

Tenstorrent develops AI acceleration hardware; Blaize combines specialized silicon with an accessible software development suite for faster time-to-market.

SiMa.ai

SiMa.ai targets edge AI; Blaize delivers purpose-built GSP architecture designed specifically for inference with superior power efficiency.

Why this matters: Blaize represents a significant trend in AI infrastructure: specialization. As cloud giants dominate general-purpose AI, Blaize has captured the large and growing edge AI market by building hardware and software from the ground up for power-constrained, latency-critical inference. With $136M in funding, strategic automotive and infrastructure partnerships, and a $56M+ surveillance deployment contract, the company is demonstrating real market traction in a category increasingly critical to autonomous vehicles and smart infrastructure.

Best for: Automotive manufacturers, smart city infrastructure operators, and industrial companies deploying real-time AI at the edge where power consumption and latency are critical constraints.

Use cases

Autonomous Vehicle Perception

DENSO and Mercedes-Benz deploy Blaize's edge AI platform for real-time computer vision processing in vehicles. The low-power GSP architecture enables continuous environmental sensing and decision-making without draining vehicle power systems or creating thermal hotspots.

Smart City Surveillance at Scale

A Southeast Asian nation deploying 250,000+ intelligent surveillance endpoints uses Blaize's platform to process video streams locally for real-time threat detection and traffic optimization. Local processing eliminates bandwidth bottlenecks and latency issues inherent in cloud-based surveillance.

Rugged Edge Computing Integration

Winmate integrates Blaize's AI chips into rugged computing systems for field deployment in harsh industrial and outdoor environments. The platform's power efficiency and thermal optimization enable reliable AI operation in conditions where data center hardware fails.

Alternatives

NVIDIA Jetson General-purpose edge GPU platform; less specialized for power-constrained inference than Blaize's purpose-built GSP architecture.

Google Coral TPU Google's edge TPU for on-device ML; more focused on consumer devices and prototyping than enterprise automotive and infrastructure deployments.

Qualcomm Snapdragon Platforms Mobile-first AI acceleration; less optimized for stationary edge infrastructure and industrial automation use cases.

FAQ

What does Blaize do? +

Blaize provides a full-stack edge AI platform combining specialized hardware (GSP processors) and a no-code/low-code software suite (AI Studio) for deploying AI inference at the network edge with extreme power efficiency and low latency. The platform serves automotive, surveillance, industrial, and smart city applications where cloud processing creates unacceptable latency or power constraints.

How much does Blaize cost? +

Blaize uses custom enterprise pricing based on deployment scale and requirements. The company does not publicly disclose pricing; interested customers must contact sales for quotes tailored to their use case.

What are alternatives to Blaize? +

NVIDIA Jetson offers general-purpose edge GPU acceleration but with higher power consumption. Google Coral TPU targets consumer on-device ML. Groq, Hailo, and Tenstorrent provide competing edge AI accelerators, though Blaize differentiates with its integrated software platform and power efficiency focus.

Who uses Blaize? +

Target customers include automotive OEMs (DENSO, Mercedes-Benz), large-scale surveillance infrastructure operators, and industrial automation providers. Blaize serves organizations deploying AI on thousands to hundreds of thousands of edge devices where power efficiency and real-time performance are critical.

How does Blaize compare to NVIDIA? +

NVIDIA dominates centralized data center AI with powerful GPUs optimized for high throughput. Blaize specializes in distributed edge inference with purpose-built hardware delivering 60-70% lower power consumption and 10x lower latency, making it ideal for autonomous vehicles, surveillance, and industrial edge applications where NVIDIA solutions would be overbuilt.

Is Blaize publicly traded? +

Yes. Blaize went public on NASDAQ on January 14, 2025 via SPAC merger under the ticker symbol BZAI. The company has raised $136M total including $106M in a post-IPO equity round and $30M through a PIPE transaction in November 2025.