$60B AI chip darling Cerebras almost died early on, burning $8M a month

TL;DR

Buying for a business?Offer from Amazon

Get business pricing on tech for your team

• Business-only prices and quantity discounts
• Tax-exempt purchasing
• Multiple users, one account, clear invoices

Create a free Business account →Free to register

As an affiliate, we earn on qualifying purchases.Hide

Cerebras Systems almost went bankrupt in 2019 after spending $8 million monthly trying to solve a technical challenge in creating massive AI chips. Today, it is a publicly traded company valued at $60 billion, highlighting its recovery and success.

Cerebras Systems, now a publicly traded company valued at approximately $60 billion, nearly collapsed in 2019 after burning nearly $200 million trying to develop a revolutionary AI chip. The company’s founder, Andrew Feldman, confirmed that during that period, they spent about $8 million per month on the project, which faced insurmountable technical challenges at the time.

Founded in 2015, Cerebras aimed to create a single, massive chip by turning an entire silicon wafer into one processor, a feat no other company had achieved. The project encountered severe packaging and cooling problems, as the chips were 58 times larger than standard microprocessors and required innovative solutions to manage heat and data transfer. By July 2019, after numerous failures and extensive trial and error, the team finally succeeded in manufacturing and operating the first functional prototype.

At that time, the company was in dire financial straits, with CEO Andrew Feldman revealing that they had exhausted nearly $200 million in the process. Despite the setbacks, the breakthrough allowed Cerebras to move forward, eventually establishing key partnerships, including with OpenAI, which loaned the company $1 billion secured by warrants. Today, Cerebras is a major player in AI hardware, serving clients like OpenAI and AWS, and achieving a market valuation of about $60 billion following its IPO.

Why It Matters

This development underscores the resilience and innovation within the semiconductor industry, especially in the AI hardware sector. Cerebras’ near failure and subsequent success illustrate the high risks and potential rewards of pushing technological boundaries. The company’s story also highlights how overcoming technical challenges can lead to market dominance and significant financial gains, impacting AI research and deployment at a global scale.

Background

In 2019, Cerebras was attempting to solve the unprecedented challenge of packaging and cooling a giant wafer-scale chip designed for AI workloads. The effort was driven by the need for more efficient AI hardware, as traditional chips require strung-together multiple units, which introduces latency and inefficiency. The breakthrough in 2019 marked a turning point, enabling the company to eventually secure high-profile customers and partnerships. The company’s IPO in 2026 and current valuation reflect its successful turnaround and industry impact.

“We were spending about $8 million a month, and we nearly burned through $200 million trying to solve one technical problem.”

— Andrew Feldman, CEO of Cerebras

“Watching a computer run is about as exciting as watching paint dry. But there we were, stunned that we’d solved the problem.”

— Andrew Feldman, CEO of Cerebras

What Remains Unclear

It is not yet clear how Cerebras plans to scale its manufacturing and sales to meet growing demand, or whether it will face new technical hurdles as it expands its product line.

What’s Next

Next steps include ramping up production capacity, securing additional high-profile clients, and expanding market share in the AI hardware sector. The company may also face new technical challenges as it develops next-generation chips and cooling solutions.

Key Questions

How did Cerebras manage to recover from near failure in 2019?

Cerebras overcame technical hurdles by inventing custom cooling and packaging solutions, ultimately enabling the production of their giant wafer-scale chips. This breakthrough allowed them to attract major clients and secure funding, leading to their current success.

What makes Cerebras’ chips different from traditional AI chips?

Cerebras’ chips are significantly larger, utilizing an entire wafer as one processor, which reduces latency and improves efficiency for AI workloads. They also require innovative cooling and packaging techniques.

What is the significance of Cerebras’ recent IPO and valuation?

The IPO and $60 billion valuation demonstrate the company’s successful turnaround and its importance in the AI hardware ecosystem, positioning it as a leader in AI acceleration technology.

Are there any ongoing technical challenges for Cerebras?

While the company has solved many initial problems, scaling manufacturing and developing next-generation chips may present new technical and logistical hurdles.