What’s New Today: Cerebras has launched a new AI server system designed to accelerate chatbot responses and improve inference performance. AI property-management company EliseAI is reportedly targeting a $3.7 billion valuation through a proposed $300 million funding round. Meanwhile, one in five fresh Singapore university graduates remains jobless, US regulators are advancing separate crypto proposals, and India’s SraVaani AI model now supports speech recognition across 65 languages and dialects.
Fast-Track Insights
Cerebras launches a new system to accelerate AI chatbots.
EliseAI reportedly targets a $3.7 billion valuation with funding.
One in five Singapore graduates remains seeking jobs.
SEC and FASB advance separate proposals for crypto.
SraVaani brings AI speech recognition to 65 languages.
Cerebras Systems has launched its new CS-4 server system, designed to accelerate AI inference and chatbot responses. The system uses three large Cerebras chips, including the new WSE-3 Turbo, alongside upgraded networking components. Built on the company’s Nexus architecture, the system uses 50% fewer components and is expected to be available in Q3 2026.
AI property management company EliseAI is reportedly in talks to raise $300 million at a $3.7 billion valuation. Andreessen Horowitz and Bessemer Venture Partners are discussing leading the round. EliseAI was valued at $2.2 billion after its 2025 funding round, meaning the proposed deal could significantly increase its valuation within a year.
One in five fresh graduates from Singapore’s six autonomous universities were still seeking employment in June, amid cautious hiring, economic uncertainty, and increasing AI adoption. About 3,600 of 18,000 graduates remained unemployed, while 9,100 had secured jobs. Singapore is expanding work-based learning, apprenticeships, and skills programs to improve graduate transitions into employment.
The US Securities and Exchange Commission and Financial Accounting Standards Board have advanced separate proposals affecting the crypto industry. The SEC proposed exemptions for certain crypto asset investment contracts, while FASB proposed allowing some stablecoins to qualify as cash equivalents. The developments come as the broader Clarity Act remains stalled in Congress.
SraVaani-1.0, an open-source speech recognition model developed by IISc’s SPIRE Lab and ARTPARK, covers 65 Indian languages and dialects, focusing strongly on low-resource and tribal languages. The system was trained using extensive Indian speech datasets and shows competitive results against other models. Its released checkpoint is available through Hugging Face.