Cerebras Takes Aim at Nvidia With New AI System Built for Faster Chatbots
Synopsis
The Nvidia rival has launched a new AI server system built around three dinner-plate-sized chips, targeting the inference workloads that power responses from chatbots such as Anthropic’s Claude.
Cerebras CS-4 is the latest server from the company, which allows returning results generated by AI models, and thus directly competing with Nvidia in the GPU-dominated space.
The system is equipped with 3x WSE-3 Turbo chips on pluggable modules, based on Cerebras Nexus server architecture. The present system was manufactured using TSMC’s 5-nanometer technology with 50 fewer chips than its predecessor, according to the company.
The company claims that the CS-4 offers up to 30x faster inference than GPUs, although the figure quoted was based on internal comparisons and has not been officially confirmed.
Cerebras expects to produce the next-generation chip in 2027, while Feldman stated that the company will be able to provide 600 Megawatts of computing power by the end of the year.
AI infrastructure spending keeps rising
The latest Worldwide AI Spending Forecast from Gartner shows that the worldwide market for AI is set to grow to $2.59 trillion in 2026, an increase of 47 percent from 2025. More than 45 percent of that spending is projected to be allocated to AI infrastructure, including semiconductors for networking and processing, AI-optimized systems, and related platforms.
In the meantime, the market currently remains dominated by US-sources, including Nvidia, and other semiconductor firms. According to Reuters, chipmakers’ investments in AI infrastructure are expected to exceed $740 billion this year.
Cerebras CS-4 is set to bolster the company’s financials, coming on the heals of a strong second quarter, according to the company’s latest earnings report on August 12.
Revenue came in at $180.1 million compared with the $103.3 million compared to the same period last year and an adjusted loss of $6.9 million compared with a loss of $7.4 million versus $450.5 million for the quarter. Its core business, excluding certain items, brought in nearly $210 million in GAAP net loss.
At the same time, the company’s cloud business generated $126 million in revenue, a jump of 281 percent year-over-year.
Cerebras announced a higher revenue guidance for full year 2026, which is set to come in at the top end of the range of $880 million to $890 million. The company also reported signing a multi-year and multi-billion dollar contract with OpenAI for 750 Megawatts of processing power.
The recent launch of CS-4 is set to provide additional momentum behind the Australian-based and US-based company’s efforts to target AI service providers who are looking to either develop or scale their offerings as demand for computing power continues to surge, with companies investing in servers, networking, and data centers to meet the requirements.
Source: Reuters
Pooja Malik is a business journalist with over six years of experience covering startups, entrepreneurship, and emerging trends. She has previously worked with leading media platforms such as YourStory Media and BW BusinessWorld, where she reported on business, policy, and market developments. Currently, she serves as Editor at The Inspirepreneur Magazine, where she writes and edits stories across business, lifestyle, and travel, with a focus on clarity, accuracy, and reader relevance.
You Might Also Like
Strengthening Investor Confidence Through Accountability
Alibaba’s latest AI chip: Can it outperform the competition?
