--- title: "Chip startup Taalas raises $169 million to develop AI chips to challenge NVIDIA" type: "News" locale: "en" url: "https://longbridge.com/en/news/276360968.md" description: "Toronto-based chip startup Taalas announced the completion of a $169 million financing round, developing an AI chip that is faster than traditional solutions and more cost-effective. The chip is customized for specific AI models, with a production cycle of about two months, significantly shorter than NVIDIA's six months. Taalas plans to launch processors supporting top large models such as GPT-5.2 by the end of the year" datetime: "2026-02-19T17:04:43.000Z" locales: - [zh-CN](https://longbridge.com/zh-CN/news/276360968.md) - [en](https://longbridge.com/en/news/276360968.md) - [zh-HK](https://longbridge.com/zh-HK/news/276360968.md) --- > Supported Languages: [简体中文](https://longbridge.com/zh-CN/news/276360968.md) | [繁體中文](https://longbridge.com/zh-HK/news/276360968.md) # Chip startup Taalas raises $169 million to develop AI chips to challenge NVIDIA Toronto-based chip startup **Taalas** announced on Thursday that it has completed a **$169 million** financing round and stated that it has developed a chip that runs artificial intelligence applications **faster** than traditional solutions and at a **lower** cost. Just weeks ago, NVIDIA acquired intellectual property rights from chip startup **Groq** for $20 billion on Christmas Eve, reigniting market interest in a batch of startups and technologies focused on specific aspects of **AI inference**. AI inference refers to the process by which AI models, like OpenAI's ChatGPT, respond to user queries. Taalas's chip design approach is to **directly engrave** part of the AI model's structure onto the silicon chip, creating custom chips optimized for specific models (such as Meta's lightweight Llama). These customized silicon chips will be paired with a large amount of high-speed but costly on-chip SRAM memory, similar to Groq's design approach. However, the true advantage of Taalas's chips lies in their **tailored design for each model**. "This hardware-level fixation is part of the source of our speed advantage," said CEO Lyubisha Baich in an interview with Reuters. Baich explained that the startup will first complete approximately 100 layers of structure, nearly a complete chip manufacturing, and then make **final customizations** on two of the metal layers. Taalas's chips are manufactured by TSMC, and custom chips for specific models can be produced in about **two months**. In contrast, the production cycle for NVIDIA's **Blackwell** AI processors is about **six months**. Taalas stated that it is currently able to produce chips that run **lightweight AI models** and plans to launch processors capable of supporting top models like **GPT-5.2** by the end of this year. Groq's first-generation processors also adopt a design approach with **high SRAM content**, and two other startups—**Cerebras**, which reached a cloud computing partnership with OpenAI in January this year, and **D-Matrix**—have also adopted similar solutions ### Related Stocks - [iShares Semiconductor ETF (SOXX.US)](https://longbridge.com/en/quote/SOXX.US.md) - [State Street® SPDR® S&P® Smcndctr ETF (XSD.US)](https://longbridge.com/en/quote/XSD.US.md) - [Invesco Semiconductors ETF (PSI.US)](https://longbridge.com/en/quote/PSI.US.md) - [Direxion Daily NVDA Bear 1X ETF (NVDD.US)](https://longbridge.com/en/quote/NVDD.US.md) - [T-REX 2X Inverse NVIDIA Daily Target ETF (NVDQ.US)](https://longbridge.com/en/quote/NVDQ.US.md) - [Tradr 1.5X Short NVDA Daily ETF (NVDS.US)](https://longbridge.com/en/quote/NVDS.US.md) - [GraniteShares 2x Short NVDA Daily ETF (NVD.US)](https://longbridge.com/en/quote/NVD.US.md) - [Direxion Daily Semicondct Bull 3X ETF (SOXL.US)](https://longbridge.com/en/quote/SOXL.US.md) - [NVIDIA Corporation (NVDA.US)](https://longbridge.com/en/quote/NVDA.US.md) - [T-REX 2X Long NVIDIA Daily Target ETF (NVDX.US)](https://longbridge.com/en/quote/NVDX.US.md) - [VanEck Semiconductor ETF (SMH.US)](https://longbridge.com/en/quote/SMH.US.md) - [Direxion Daily NVDA Bull 2X Shares (NVDU.US)](https://longbridge.com/en/quote/NVDU.US.md) - [YieldMax NVDA Option Income Strategy ETF (NVDY.US)](https://longbridge.com/en/quote/NVDY.US.md) - [GraniteShares 2x Long NVDA Daily ETF (NVDL.US)](https://longbridge.com/en/quote/NVDL.US.md) ## Related News & Research - [Fractile Seeks $200 Million as Nvidia Faces Inference Pressure](https://longbridge.com/en/news/281215985.md) - [Nvidia Stumbled in the First Quarter: Here are My Top 3 Nvidia Predictions for the Second Quarter](https://longbridge.com/en/news/281620093.md) - [Nvidia’s $2 Billion Investment in Marvell: What Kind of Company Is Marvell?](https://longbridge.com/en/news/281372238.md) - [Nvidia (NVDA) H100 Prices Surge 40% as New GPUs Fail to Meet Insatiable Demand](https://longbridge.com/en/news/281588202.md) - [Argentarii LLC Acquires 20,602 Shares of NVIDIA Corporation $NVDA](https://longbridge.com/en/news/281612604.md)