Guozhen AIGlobal AI field notes and model intelligence

Realtime AI News

Edge AI chip startup Acrab raises $130M Series B as GΞLIX 1 heads to production

Acrab, a Singapore-based AI compute infrastructure company, has closed a $130 million Series B with Vertex Growth among the participants, bringing total funding past $480 million for production capacity, ecosystem and next-generation platform development. The startup, founded in 2024, has moved its first edge AI chip GΞLIX 1 into customer onboarding and scaled production preparation alongside the Agent Box personal AI center.

Published

Acrab, a Singapore-based AI compute infrastructure company, has completed a $130 million Series B round with Vertex Growth among the investors, according to DealStreetAsia as reported by Chinese tech media QbitAI. The round pushes Acrab's cumulative funding past $480 million.

The new capital will go toward expanding production capacity, building out the technology ecosystem, and accelerating development of the next-generation AI compute platform.

Founded in 2024, Acrab unveiled its first-generation edge AI chip, GΞLIX 1, and the Agent Box personal AI center less than a month ago; the first-generation products have already entered customer onboarding and are moving toward scaled production — an unusually fast pace for a chip startup.

According to the company, the GΞLIX chip was taped out in November 2025, passed 12 key validations within six hours of power-on, and booted Linux and Android within 24 hours; the chip and Agent Box were formally launched in July.

GΞLIX 1 is built around a claimed 700 TOPS of peak compute, with unified memory, CPU-NPU co-design, prefill processing and agent orchestration in a single platform. Acrab's own tests report a prefill speed of 1,416 tokens per second running a Gemma 26B A4B model with a 40K-token context cache, which the company says is seven times faster than leading edge AI compute platforms — figures that still need third-party verification.

Agent Box is positioned as a Personal AI Center: a desktop device that runs hundred-billion-parameter models locally with long-term memory, multimodal interaction, tool calling and agent orchestration, while doubling as a reference system for AI PCs, AI NAS, automotive, industrial equipment and robots.

Acrab's thesis is that the cloud compute market is already crowded, and that as models compress and agents multiply, high-frequency, privacy-sensitive, low-latency tasks will shift to the edge — where devices will need a new compute foundation rather than just a faster NPU. CEO Ken Phua, a former co-CEO of Arm China, brings server-class chip and systems software experience to the bet.

The real test ahead is which mass-production devices GΞLIX 1 lands in, when the first customers appear, and whether shipments become sustained — the difference between technical validation and commercial validation.

Why it matters

With over $480M in cumulative funding and its first chip entering production preparation, Acrab is betting that agentic AI moves compute to the edge — a thesis that will be tested by real device adoption.

AcrabAI ChipFunding
Back to realtime news

Nearby Updates

All

08/13, 20:04

Grok 4.6 arrives, topping GPT-5.6 Sol and Fable 5 Max on agent benchmarks at $2 per million tokens

SpaceXAI has released Grok 4.6, which official benchmarks show beating GPT-5.6 Sol and Fable 5 Max on GDPVal-AA v2 and other agentic tests while priced at just $2 per million input tokens and $6 per million output tokens. The model is live in Grok Build, Cursor, Grok Bot and the API — with OpenRouter, Vercel and Cloudflare also carrying it — and is tuned for long-horizon agent tasks.

08/13, 19:53

Embodied-data startup SCALEFORCE closes two funding rounds in 40 days to build physical-AI infrastructure

Embodied-intelligence data infrastructure startup Yuanpoint Technology (SCALEFORCE) has completed a new funding round worth tens of millions of yuan, with investors including Hengxu Capital, Kailian Capital and a top domestic embodied-AI industry player — its second round within 40 days. The proceeds will fund the MatrixOS physical-AI operating system, a large-scale data production network, and team expansion.

08/13, 19:29

Claude clears all Hadamard matrices below order 2000, striking another open problem off the math list

Chinese tech outlet QbitAI reports that Anthropic's Claude has cleared all Hadamard matrices below order 2000, removing a long-standing open problem in combinatorics from the to-solve list. The report credits the mathematician's approach as much as the model itself, a fresh sign that large models can contribute to serious mathematical research.

08/13, 21:52

Report: SpaceX Races Toward 10GW as AI Compute War Enters Musk's “Comfort Zone”

Odaily reports that SpaceX is racing toward a 10-gigawatt compute target, pushing the AI compute war into what it calls Musk's comfort zone. The report argues that power and infrastructure, not chips alone, are becoming the decisive battleground in the AI arms race.