由于数据需求旺盛，Snorkel AI 将估值提高至 35 亿美元

Snorkel AI, a startup that helps AI labs and corporations create training datasets and simulated environments, has raised $350 million in a Series E round at a valuation of $3.5 billion, according to TechCrunch. The new round, led by Insight Partners and S32, values ​​the seven-year-old startup at nearly three times the $1.3 billion it raised in its $100 million Series D round 17 months ago. Existing investors including Addition, Lightspeed, Greylock, GV and Wells Fargo also participated in the round. Snorkel initially provided software to automate data tagging, but last year the company changed strategy and began offering clients ready-made data sets, which it calls data as a service. Rather than operating solely as a marketplace of human experts, Snorkel takes a hybrid approach, using its software and models to synthetically generate data alongside subject matter experts. According to the company, Snorkel's current annual revenue is $375 million, up 18 times from last year. This growth is driven by insatiable demand

AI labs on high-quality training data. Other data companies that position themselves as AI data labs are also seeing similar explosive growth. Mercor's annual revenue reached $2 billion, Handshake passed the $1 billion mark earlier this year, and TechCrunch reported that Micro1 increased its revenue to $500 million. It's important to note that since these companies pay approximately 60% to 70% of their income directly to the professionals doing the work, their actual net annual income is significantly lower than the numbers they report. Given that Snorkel sells reinforcement learning environments and complete data sets rather than human labor, payments to its experts are counted in cost of goods sold rather than headline annual revenues, according to the company. Snorkel began commercial operations in 2019 after four years of research conducted by co-founder and CEO Alex Ratner and his team at the Stanford AI Lab. The material was prepared by a neural network based on the original publication. AI may be wrong in details (numbers, names, dates, quotes) - for important information, check with the original source

full-time student Rabbit introduced a new AI agent that works without R1 Key-value caching improves the interaction of AI models SpeakON introduced a magnetic button with voice input for the iPhone. Rabbit announced OS3, a new software platform for AI agents. Trane Technologies reduced HVAC diagnostic time from 20 minutes to 20 seconds using AI.