Seoul-based Datumo raises $15.5M

PLUS: Recent Developments In AI

Author

The AI Surface
August 12, 2025

👩‍💻 Greetings To All AI Learners. Datumo just secured $15.5 million, led by Salesforce Ventures, to advance its no-code AI safety and evaluation platform positioning itself as a challenger to incumbents like Scale AI with a unique dataset and enterprise-friendly approach.

✨ In today’s AI News we will cover:

🕴🏻 Seoul-based Datumo raises $15.5M to take on Scale AI, backed by Salesforce.

🕴🏻 Apple’s new Siri may allow users to operate apps just using voice.

🕴🏻 Former Googlers’ AI startup OpenArt now creates ‘brain rot’ videos in just one click.

🕴🏻 Nvidia unveils new Cosmos world models, infra for robotics and physical uses.

🕴🏻 4 Trending AI Tools

LATEST NEWS HIGHLIGHTS

✨ DATUMO’S NEW VENTURE

💰 Datumo, a Seoul-based AI evaluation and safety startup, has raised $15.5M led by Salesforce Ventures to expand globally and challenge Scale AI with its no-code AI safety testing platform. The company is betting on the growing enterprise demand for transparent, fair, and safe AI systems.

 💫 Highlights

🪞 Funding amount and investors: Datumo raised $15.5 million in its latest round of financing, bringing its total funding to approximately $28 million. Key investors include Salesforce Ventures, KB Investment, ACVC Partners, and SBI Investment.

🪞 Evolution and strategy: Originally focused on AI data labeling since its founding in 2018 by KAIST alumni, Datumo has pivoted toward providing enterprise-ready AI safety and evaluation tools.

🪞 Market positioning: The startup competes in evaluation and monitoring domains alongside firms like Scale AI, Galileo, and Arize AI. It differentiates through its proprietary, hard-to-clean datasets sourced from published books, and a comprehensive evaluation stack.

💫 Significance

📍 AI safety is a growing priority: As generative AI becomes more pervasive, businesses urgently need tools that ensure transparent, fair, and safe model behaviors.

📍 Low-code/no-code is key: Datumo’s platform opens up evaluation capabilities beyond technical teams empowering compliance, policy, and trust teams to act directly.

📍 Reading between the lines: The inclusion of Salesforce Ventures underscores strong strategic confidence in Datumo’s vision and growth potential.

✨ NEWEST VERSION OF SIRI

🍎 Apple is developing a major Siri overhaul, powered by a revamped App Intents framework, which will let users control apps entirely via voice commands no screen touches required. This includes tasks like editing photos, shopping, logging into services, and commenting on social media. The feature is being tested now and is expected to roll out in spring 2026.

💫 Highlights

🔊 Full in-app voice control: Siri will be able to manage complex actions finding, editing, sending content across apps using only voice.

🔊 Third-party integration: Apple is testing across its own apps and key third-party services like Uber, Amazon, YouTube, Facebook, WhatsApp, Threads, Temu, and AllTrails.

🔊 Launch timeline: Scheduled for spring 2026, potentially tied to iOS 26.

💫 Significance

🐝 Hands-free, frictionless experience - Enables users to perform complex multi-step tasks in apps without touching the device, ideal for accessibility, multitasking, and on-the-go use.

🐝 Shift toward a “voice-first” ecosystem - Marks Apple’s move from Siri as a simple assistant to a core control interface for iPhones, iPads, and potentially future devices like smart glasses or home hubs.

🐝 Accessibility game-changer - Empowers users with disabilities or mobility limitations to fully operate their devices independently.

✨ GOOGLER’S ‘BRAIN ROT’

🥀 OpenArt, founded by ex-Google engineers, has launched One-Click Story, an AI tool that turns a single sentence, script, or song into a surreal, one-minute “brain rot” video. With over 50 AI models powering it, the platform blends speed, accessibility, and meme-driven aesthetics positioning itself as a viral content factory while raising legal and cultural questions.

💫 Highlights

🌈 One-Click Story tool: Generates fully animated videos with a narrative arc from minimal input.

🌈 Target style: “Brain rot” - absurd, chaotic, meme-friendly visual storytelling perfect for TikTok and Reels.

🌈 User base: 3M–6M monthly active users; business already profitable.

🌈 Funding: $5M raised from Basis Set Ventures and DCM Ventures.

🌈 Pricing: Subscription starts at $14/month; credit-based model.

🌈 Revenue: Projected annual revenue > $20M.

💫 Significance

🍭 Democratizes video creation - Removes technical barriers, letting anyone produce share-ready, visually rich content in seconds.

🍭 Cultural accelerant - Fuels the rise of absurdist, ultra-fast-consumption media popular with Gen Z and meme culture.

🍭 Potential IP minefield - Highlights the tension between AI creativity and copyright laws.

🍭 Commercial opportunity - Early profitability and strong subscription traction make it a viable player in the AI-content economy.

✨ THE COSMOS MODEL

🏹 Nvidia’s Cosmos advancements redefine physical AI, marrying advanced reasoning models with scalable simulation and infrastructure. This not only speeds up robotics and autonomous development but opens the field to broader innovation and experimentation setting the stage for smarter, more autonomous machines in diverse industries.

💫 Highlights

👩🏼‍🏫 Cosmos Reason - a 7B-parameter reasoning VLM Empowers robots with memory and physics-aware planning capabilities, enabling them to predict and plan the next best action. Useful for data curation, robot planning, and video analytics.

👩🏼‍🏫 Cosmos Transfer-2 & its distilled counterpart: These models dramatically accelerate the generation of synthetic data from 3D simulations or spatial control inputs, with the distilled version optimized for faster execution.

👩🏼‍🏫 Neural Reconstruction Libraries & Omniverse Upgrades: Introduces rendering techniques enabling accurate 3D simulations from sensor data, now integrated into platforms like CARLA and updates to the Omniverse SDK.

💫 Significance

👉🏾 Empowering Physical AI with Foresight and Planning: The highlight is Cosmos Reason, a 7B-parameter vision-language model that equips robots and AI agents with memory and physics-informed reasoning enabling them to plan the next best action based on real-world understanding.

👉🏾 Mass-Scale Synthetic Reality for Faster Development: The new Cosmos Transfer-2 model, and its faster distilled version, streamline synthetic data generation from 3D scenes and control inputs making it dramatically easier to create training datasets.

👉🏾 Hardware & Cloud Ready for Robotics Developers: Nvidia’s RTX Pro Blackwell Servers and DGX Cloud infrastructure offer a unified and scalable stack for physical AI workloads, from simulation to model training and deployment.

QUICK HITS

✨ 4 TRENDING AI TOOLS

 🤖 Hatchful

❄️ Function: Creates professional logos with free logo maker.

 🤖 Miro AI

❄️ Function: Visual collaboration & idea generation.

❄️ Function: Academic writing enhancement and plagiarism detection.

 🤖 Portaly

❄️ Function: Creates personalized link-in-bio pages instantly.

✨ Other Recent Developments In AI

🎗 The ATOM Project (“American Truly Open Models”) aims to reestablish U.S. leadership in open-source AI development. Backed by high-powered GPUs and supported by key AI figures, it targets $100 million in funding to democratize advanced model accessibility.

🎗 Academic work reviews AI’s expanding role in energy management, from forecasting to creating autonomous "digital twin" systems that can simulate and act within smart grids.

🎗 A collaboration between the Cleveland Clinic and AI startup Piramidal is creating a real-time EEG-based AI system to detect neurological issues like seizures significantly reducing diagnosis time from hours to seconds. 

🤓 That’s A Wrap For Today!

Stay tuned for more updates.

Build AI awareness. Lead the future.

👉 Join the AI Revolution to make a difference.

— The AI Surface