AI Chip Startups Raise Over $1.5B
What happened
The AI hardware landscape is seeing significant new investment to challenge market leaders. MatX, a startup founded by Google alumni, has raised over $500 million. Separately, Cerebras raised $1 billion at a $23 billion valuation to target AI inference workloads, supported by a deal with OpenAI.
Why it matters
- MatX's $500M Series B funding was led by Jane Street and a fund from former OpenAI researcher Leopold Aschenbrenner, with strategic backing from supply chain players like Marvell Technology. The startup, founded by ex-Google TPU engineers, aims to produce its MatX One chip with TSMC, targeting initial shipments in 2027. - Cerebras's latest $1 billion funding round nearly tripled its previous valuation to $23 billion, with notable participation from AMD, a direct competitor in the hardware space. This capital fuels their partnership with OpenAI to supply 750 megawatts of computing power, directly challenging Nvidia's market dominance. - For hardware companies with 6-12 month sales cycles, opportunity stage forecasting is a common methodology; it assigns a probability percentage to each stage of the sales pipeline to create a weighted forecast. This approach provides a more realistic revenue prediction than simple, gut-feel estimates from reps. - To manage the complexity of multi-stakeholder hardware deals, sales operations teams often implement CRM automation to handle repetitive tasks like follow-up reminders, lead scoring, and updating deal stages. This frees up sales representatives' time to focus on high-value activities and relationship building. - Key leading indicators for deal health in long-cycle hardware sales include Sales Velocity, which measures how quickly deals move through the pipeline, and a Pipeline Coverage Ratio of 3-4x the quota. These metrics provide an early warning system for potential shortfalls in the forecast. - High-performing revenue operations (RevOps) teams in technical sales focus on process efficiency metrics like lead response time (targeting under 15 minutes) and the MQL-to-SQL conversion time to identify bottlenecks between marketing and sales handoffs. - A critical practice for sales ops leaders in complex hardware sales is mapping the entire revenue process, from lead generation to renewal, and clearly defining the handoffs between marketing, sales, and customer success to ensure alignment. - The overall AI chip market was valued at approximately $94.44 billion in 2025 and is projected to exceed $1.1 trillion by 2035, indicating a sustained, high-growth environment for companies in this sector.
Key numbers
- MatX, a startup founded by Google alumni, has raised over $500 million.
- Separately, Cerebras raised $1 billion at a $23 billion valuation to target AI inference workloads, supported by a deal with OpenAI.
- - MatX's $500M Series B funding was led by Jane Street and a fund from former OpenAI researcher Leopold Aschenbrenner, with strategic backing from supply chain players like Marvell Technology.
- The startup, founded by ex-Google TPU engineers, aims to produce its MatX One chip with TSMC, targeting initial shipments in 2027.
What happens next
- The startup, founded by ex-Google TPU engineers, aims to produce its MatX One chip with TSMC, targeting initial shipments in 2027.
- Separately, Cerebras raised $1 billion at a $23 billion valuation to target AI inference workloads, supported by a deal with OpenAI.
Sources
Quick answers
What happened in AI Chip Startups Raise Over $1.5B?
The AI hardware landscape is seeing significant new investment to challenge market leaders. MatX, a startup founded by Google alumni, has raised over $500 million. Separately, Cerebras raised $1 billion at a $23 billion valuation to target AI inference workloads, supported by a deal with OpenAI.
Why does AI Chip Startups Raise Over $1.5B matter?
MatX's $500M Series B funding was led by Jane Street and a fund from former OpenAI researcher Leopold Aschenbrenner, with strategic backing from supply chain players like Marvell Technology. The startup, founded by ex-Google TPU engineers, aims to produce its MatX One chip with TSMC, targeting initial shipments in 2027. Cerebras's latest $1 billion funding round nearly tripled its previous valuation to $23 billion, with notable participation from AMD, a direct competitor in the hardware space. This capital fuels their partnership with OpenAI to supply 750 megawatts of computing power, directly challenging Nvidia's market dominance. For hardware companies with 6-12 month sales cycles, opportunity stage forecasting is a common methodology; it assigns a probability percentage to each stage of the sales pipeline to create a weighted forecast. This approach provides a more realistic revenue prediction than simple, gut-feel estimates from reps. To manage the complexity of multi-stakeholder hardware deals, sales operations teams often implement CRM automation to handle repetitive tasks like follow-up reminders, lead scoring, and updating deal stages. This frees up sales representatives' time to focus on high-value activities and relationship building. Key leading indicators for deal health in long-cycle hardware sales include Sales Velocity, which measures how quickly deals move through the pipeline, and a Pipeline Coverage Ratio of 3-4x the quota. These metrics provide an early warning system for potential shortfalls in the forecast. High-performing revenue operations (RevOps) teams in technical sales focus on process efficiency metrics like lead response time (targeting under 15 minutes) and the MQL-to-SQL conversion time to identify bottlenecks between marketing and sales handoffs. A critical practice for sales ops leaders in complex hardware sales is mapping the entire revenue process, from lead generation to renewal, and clearly defining the handoffs between marketing, sales, and customer success to ensure alignment. The overall AI chip market was valued at approximately $94.44 billion in 2025 and is projected to exceed $1.1 trillion by 2035, indicating a sustained, high-growth environment for companies in this sector.
