AMD Acquires AI Chip Startup Taalas to Boost Inference Workload Efficiency

AMD has acquired Taalas, a Canadian AI chip startup, in a move aimed at strengthening its position in the rapidly evolving AI inference market. The company said it plans to integrate Taalas' technology to "develop system-level solutions with AMD Instinct GPUs," signaling a broader push to deliver more efficient and scalable AI compute options for enterprise customers.

The acquisition reflects the intensifying competition among chipmakers to optimize inference workloads, which are becoming increasingly critical as AI models move from training to widespread deployment. AMD's statement emphasized that the deal will help customers run inference workloads more efficiently across a growing range of applications, a key differentiator as demand for low-latency, high-throughput AI services surges.

Financial terms of the transaction were not disclosed, and the companies did not provide a timeline for when the deal is expected to close. The acquisition adds to AMD's recent momentum in the AI hardware space, where it has been challenging Nvidia's dominance with its Instinct GPU lineup and a broader software ecosystem.

Taalas was founded in August 2023 by Ljubisa Bajic, a former architect at both AMD and Nvidia and a co-founder of Tenstorrent, alongside engineers Drago Ignjatovic and Lejla Bajic. Bajic's deep industry pedigree includes a job swap with Tenstorrent CTO Jim Keller in October 2022 before he stepped down from the company entirely in March 2023.

In February 2026, Taalas raised $169 million to support the development of chips tailored for specific AI models. The startup also unveiled its first product, the HC1 Technology Demonstrator, which is optimized to run the open-source Llama 3.1 8B language model. Fabricated by TSMC using its 6nm process node, Taalas claims its chips can generate more tokens per second per user than Nvidia's H200 and B200 offerings, as well as hardware from Groq, SambaNova, and Cerebras, while consuming just one tenth of the power.

"AMD is building a full-stack AI platform that gives customers the flexibility to deploy the right compute solutions for every AI workload," said Vamsi Boppana, Senior Vice President of the Artificial Intelligence Group at AMD. "Taalas' technology and world-class engineering team strengthen our AI portfolio by delivering differentiated inference performance and efficiency."

Ljubisa Bajic, co-founder and CEO of Taalas, added: "We founded Taalas to rethink AI inference from the ground up by building the hardware around the model. Our Canada-based team has combined deep technical expertise with a willingness to challenge conventional approaches. Joining AMD will give us the scale, engineering resources and global reach to accelerate our innovation."

The acquisition underscores a broader industry trend toward specialization in AI hardware, as companies seek to reduce power consumption and cost while improving performance for inference-heavy workloads. By bringing Taalas' model-specific chip expertise in-house, AMD is positioning itself to offer more tailored solutions that could appeal to hyperscale data center operators and enterprise users alike. With inference expected to dominate AI compute demand in the coming years, this deal could help AMD carve out a stronger foothold in a segment where efficiency is becoming the primary battleground.