AI Furnace Newsroom: Google Turns Phone Photos Into Studio-Ready Product Ads, New Google’s Gemini 3.1 Pro, Former GitHub CEO Raises A $60M Seed, Nebius Buys Tavily To Build Agentic Search Into Its AI Cloud

Welcome to this week’s AI Furnace Newsroom

In today’s insights we cover:

1. Google Turns Phone Photos into Studio-Ready Product Ads

2. Google’s Gemini 3.1 Pro Launch, with Double the Reasoning Power

3. Former GitHub CEO Raises a $60M Seed Round for Developer Platform Startup

4. Nebius Acquires Tavily for up to $400M to Build Agentic Search into its AI Cloud

Read time: 5 mins

💡 Furnace Insights

Nebius

Nebius Acquires Tavily for up to $400M to Build Agentic Search into its AI Cloud

Nebius, a Netherland-based AI infrastructure company which spun out from the Russian internet giant Yandex, agreed to acquire Tavily, an “agentic search” startup. The deal was for $275M upfront, with the potential to rise to $400M based on milestones. Tavily builds real-time search infrastructure designed for AI agents, and Nebius is folding that capability into its AI cloud platform as enterprises push harder into agents that retrieve and act on behalf of the workforce.

Tavily surpassed one million monthly downloads and the enterprise-grade system already serves a number of Fortune 500 companies, helping financial institutions combat fraud, logistics providers manage operations, and academic organizations advance research.

According to Rotem Weiss, founder and CEO of Tavily, “We extract data in real time and feed it into AI agents…companies are looking for a unified search layer that can operate across a wide range of models…unlike traditional search companies, our approach is not structured data retrieval but dynamic information extraction.” Roman Chernin, co-founder and CBO of Nebius, said “This acquisition brings the search layer directly into our stack, so developers can focus on their applications instead of managing multiple vendors. Our strategy is clear: provide an open platform that serves everyone from startups to the largest enterprises, giving them the tools to own their AI destiny.”

Google

Google Turns Phone Photos into Studio-Ready Product Ads

Google Labs launched a Photoshoot feature inside the Pomelli marketing platform. This tool can take a basic product photo, often just a phone snap, and generate a polished, ad-ready image to match a brand. Pomelli can also infer a brand look via “Business DNA” from a website, and generate images using the Nano Banana image model. The feature is rolling out in the U.S., Canada, Australia, and New Zealand.

This is a quiet but meaningful shift for small businesses: the bottleneck isn’t ad targeting, it’s creative throughput. If one decent photo can quickly yield dozens of usable variations, teams can refresh catalogs faster, run more A/B tests, and keep campaigns moving without waiting on shoots or designers. The flip side is sameness - when everyone has access to “studio quality,” differentiation moves from production value to taste, brand voice, and what you choose to test.

Entire

Former GitHub CEO Raises a $60M Seed Round for Developer Platform Startup

Thomas Dohmke, GitHub’s former CEO, just raised a massive $60M seed round for Entire, reportedly at around a $300M valuation - a reminder that “seed” now sometimes means “Category launch with real capital.” His mission is to build a developer platform for the era where humans manage fleets of coding agents, including tooling to capture what agents did, why they did it, and how to review it.

It is clear that for some “special” founders with the right backgrounds, mega-sized seed rounds come across regularly and we don’t bat an eye. Fei-Fei Li’s company World Labs launched in 2024 with $230 million in initial funding. Ilya Sutskever’s Safe Super Intelligence raised $1 billion in funding out of the gate in 2024. And Mira Murati’s Thinking Machines launched with a $2 billion seed round. Recently, an AI research lab started by former Anthropic, xAI, and Google engineers, called humans&, launched with $480 million in funding at a $4.48 billion valuation.

Google

Google’s Gemini 3.1 Pro Launches, with Double the Reasoning Power

Google just shipped Gemini 3.1 Pro, claiming a big jump on harder reasoning tests, including a 77.1% verified score on ARC-AGI-2, more than double Gemini 3 Pro’s performance on that benchmark. Google is positioning it as a “do the hard stuff” model and is pushing it into places people actually use, such as the Gemini app and its broader product stack.

The key point here is whether reasoning upgrades translate into fewer “almost right” answers when you’re using AI for real work (analysis, planning, debugging, decision support). Benchmarks can move fast because models can think longer or optimize for specific tests, so most teams should treat this as a signal to re-evaluate reliability and cost in their own workflows, not as a winner-takes-all moment.

📈 AI Venture Deals of the Week

• Potpie AI, a foundational context layer service provider, raised its $2.2M Pre-seed funding.

• Cydelphi, an AI-native digital forensics and incident response platform, raised its $3M Seed.

• Rapidata, an AI infrastructure startup focused on human feedback for AI, raised its $8.5M Seed.

• Entire, an AI code collaboration and developer tools startup, raised its $60M Seed.

• Stacks, a agentic platform for enterprise finance, raised its $23M Series A.

• Braintrust, an AI observability and evaluation platform, raised its $80M Series B.

• xAI, raised its $3B Series E.

⚒️ New AI Product Launches You Don’t Want to Miss 

• Google released Gemini 3.1 Pro, with improved reasoning and problem-solving capabilities across text, image, audio, and multimodal tasks.

• Anthropic released Claude Code Security, an AI-powered vulnerability scanner that analyzes codebases for security issues and suggests fixes.

• xAI released Grok 4.20, an AI with a native multi-agent debate architecture where four specialized agents work in parallel and debate before producing a final answer.

• PaleBlueDot AI, a global AI compute platform for scalable GPU cloud infrastructure and intelligent AI Cloud Agents.

• Hibiki-Zero, a real-time multilingual speech translation model that translates between multiple languages with low latency and high audio quality.

• Migma.ai, a cutting-edge AI email creation platform that automatically generates on-brand, responsive emails from simple prompts.

Upcoming Events 📅

• AIF Founder Supper Club @ Private Room - Tuesday, March 24

Interested in meeting the Who’s Who of AI? Join the AI Hot 100 Summit in NYC

The AI Hot 100 Summit is back for its third edition on May 6-7th, 2026 in New York City. The only AI Summit built for real connections where AI Visionaries meet Industry Leaders. Join over 500 AI Executive Leaders, Founders and Investors for a jam packed 2 days of enterprise AI lightning talks, innovation showcases, and curated networking.

Get ready to meet and greet speakers and attendees from OpenAI, NVIDIA, Haleon plc, Legora, Anthropic, Cursor, Eleven Labs, Adobe, Walmart, Insight Partners & more.

🗓️ May 6-7, 2026

📍New York City