PitchBook's Q2 2026 U.S. VC Valuations report shows that 87.5% of U.S. venture dollars in the first half of 2026 went to AI megadeals, with AI companies seeing median valuation step-ups of 2.2x versus…
Most companies overpay for AI inference by 2x to 5x, sometimes by an order of magnitude, because their inference stack doesn't match their use case priorities, said Meryem Arik, co-founder of inferenc…
A BBC report reveals that employees at AI labs like OpenAI and Anthropic work 50 to 90-hour weeks despite public promises that AI would enable shorter workweeks. OpenAI President Greg Brockman has doc…
NVIDIA has released SwitchYard, an open-source routing library that automatically directs AI agent tasks to the most appropriate local or frontier model, claiming it can deliver 50% faster responses a…
TileRT's persistent engine on NVIDIA GPUs achieves up to 500 tokens/s/user on the InferenceX GLM5 FP8 744B benchmark on a single B200 decode server, approximately 3× faster than GB300 NVL72 running tr…
AMD acquired Toronto-based Taalas on August 6, a chip startup that hardwires AI model weights into silicon using Mask ROM, eliminating the memory bottleneck. Taalas's HC1 chip, built on TSMC's 6nm pro…
A developer has detailed the architecture behind SARA, an open-source WhatsApp AI agent that has achieved 99.7% uptime over six months using a multi-provider AI chain with zero inference cost. The sys…
MarketDaily, a daily financial email digest serving 21 subscribers, uses a 9-model LLM council with a judge and a 31-check audit gate to prevent silent degradation and hallucinated price targets, but …
Andrew Feldman, cofounder and CEO of AI chip company Cerebras, which is valued at $49.5 billion, said on the 20VC podcast that it is 'mind-boggling' to think people can achieve greatness by working 38…
AMD agreed to acquire Taalas, a 25-person Toronto-based chip startup that builds model-specific silicon, for an undisclosed amount. Taalas's first chip, the HC1, runs only Llama 3.1 8B, with weights h…
AMD has acquired Taalas, the AI inference chip startup co-founded by former Tenstorrent CEO Ljubisa Bajic, who will rejoin AMD under Vamsi Boppana's AI organization. Taalas's first chip achieves over …
Lamb Labs, a Y Combinator-backed startup founded by Niki Kotecha and Thomas Lanning, announced on August 4th that its planned custom inference chips could generate more than 20,000 tokens per second a…
PitchBook data shows $1,786.8 billion in US VC-backed IPO value through July 28, 2026, but SpaceX's June 12 IPO at a $1.77 trillion valuation accounts for about $1,770 billion of that total, leaving t…
Olix Computing, a UK-based startup founded by 25-year-old Thiel Fellow James Dacombe, closed a $312 million Series B at a $3.3 billion valuation to scale its photonic AI chips, up from roughly $1 bill…
MarketDaily, a daily financial email digest serving 21 subscribers, uses a 9-model LLM council with a judge and a 31-check deterministic audit gate to generate personalized reports, but has experience…
OpenAI released GPT-5.6, a family of models including Sol, Terra, and Luna, on July 9 after a 12-day government-gated preview, with the flagship Sol achieving state-of-the-art scores on Terminal-Bench…
OpenAI has launched the GPT-5.6 family, comprising Sol, Terra, and Luna, with a focus on speed and multi-agent work. The models offer varying balances of capability, cost, and efficiency, with GPT-5.6…
TechCrunch Disrupt 2026, taking place October 13–15 at Moscone Center in San Francisco, will feature leaders from Amazon, Replit, Tether, Flock Safety, Benchmark, Harvey AI, and Cerebras on its main s…
OrionChat, a free, open-source web-based chat interface for local LLMs, offers a simpler alternative to Open WebUI for running Ollama models in a browser, with setup taking less than five minutes. The…
Cerebras and AMD have partnered to build what they claim will be the world's fastest disaggregated AI inference solution, combining AMD's Helios rack-scale architecture for the pre-fill phase with the…