Indian AI Chip Startup Ceremorphic Enters Drug Development


Cerebras Systems successfully went public in 2026, raising $5.5 billion through its IPO, and its valuation surged to approximately $60 billion, making its founder a billionaire. The company was founded in 2016 and focuses on AI chip manufacturing, but it nearly went bankrupt in 2019, with monthly expenses of $8 million, almost exhausting its funds.
British semiconductor company Arm announced that its first self-developed AI chip is expected to generate $2 billion in sales next year, as market demand is strong. This move marks a significant breakthrough for Arm in the AI field. As a subsidiary of SoftBank, Arm holds a key position in the global semiconductor market, with its technology widely used in electronic devices. The launch of the new chip comes at a time when AI technology is rapidly developing and the demand for high-performance computing is increasing.
OpenAI launches GPT-Rosalind, an AI model for life sciences named after DNA pioneer, designed to accelerate drug discovery by analyzing biochemical data to aid in evidence synthesis, hypothesis generation, experimental planning, and protein engineering, enhancing lab efficiency and medical application.....
The Tsinghua Institute for Intelligent Technology Research collaborates with Shuimu Molecules to open source the general-purpose chemical molecule model BioMedGPT-Mol, marking a significant step in the deepening of domestic large models from general dialogue to specialized fields such as biomedicine. This model is specifically designed for chemical molecules, aiming to promote the intelligence and standardization of drug development.
Meta has signed an infrastructure agreement worth up to 27 billion dollars with cloud service provider Nebius, ensuring sufficient advanced computing power for the next five years to cope with global AI competition. The agreement includes 12 billion dollars in dedicated computing capacity and 15 billion dollars in optional expansion, highlighting Meta's urgent need for computing power reserves.