Presented by
Good morning, AI enthusiasts. OpenAI may be the most talked-about AI company, but Anthropic is winning in enterprise. Now, OpenAI is going straight to Wall Street to close the gap.
The company is in advanced talks with TPG, Bain Capital, Brookfield, and Advent to form a joint venture valued at ~$10B, with the goal of putting its engineers directly inside corporate clients. If it lands, it's the most institutional enterprise play in AI history.
In today's recap:
OpenAI's $10B private equity venture with TPG and Bain
Mistral AI and NVIDIA co-develop open-source AI models
NVIDIA DLSS 5 brings AI-powered photorealism to games
4 new AI tools, and more
OPENAI
OpenAI bets on private equity for enterprise push
Reuters
Recaply: OpenAI just outlined a plan to form a joint venture with major private equity firms, targeting a valuation of ~$10B to push its AI tools into more enterprise clients.
Key details:
TPG, Bain Capital, Brookfield, and Advent would all get board seats and early access to Frontier, OpenAI's enterprise AI platform. TPG would anchor the deal with the largest commitment.
OpenAI's enterprise business hit $10B in annualized revenue out of $25B total by late February 2026. Codex has 2M+ weekly users, up 4x since the start of the year.
PE firms would put in ~$4B for preferred equity stakes, per Reuters sources. That gives them priority returns and limits downside compared to common equity.
No final deal has been signed yet. Fidji Simo, CEO of Applications at OpenAI, said the company is building a deployment arm and will share more details when ready.
Why it matters: OpenAI's enterprise revenue is at $10B, but Anthropic is still seen as the leader in corporate AI adoption. This JV is OpenAI's move to close that gap. By putting its engineers directly inside PE-backed companies, it's copying a playbook that Palantir made famous. Both AI giants are eyeing IPOs this year. Whoever wins enterprise first wins the race.
PRESENTED BY ATTIO
Here’s how I use Attio to run my day.
Attio is the AI CRM with conversational AI built directly into your workspace. Every morning, Ask Attio handles my prep:
Surfaces insights from calls and conversations across my entire CRM
Update records and create tasks without manual entry
Answers questions about deals, accounts, and customer signals that used to take hours to find
All in seconds. No searching, no switching tabs, no manual updates.
Ready to scale faster?
AI
Mistral AI and NVIDIA unite for open frontier AI
Nvidia
Recaply: Mistral AI just announced a partnership with NVIDIA as a founding member of the Nemotron Coalition, combining its model architecture with NVIDIA's compute to build open-source frontier AI together.
Key details:
Mistral brings its training techniques, multimodal tools, and fine-tuning expertise. NVIDIA contributes DGX Cloud compute, synthetic-data pipelines, and development tools. Together they'll build the coalition's first base model.
That first model will form the base of the Nemotron 4 family and be fully open-sourced. Mistral also released Mistral Small 4 on the same day, a 119B-parameter model that's 40% faster than its last.
Arthur Mensch, CEO of Mistral AI, said the two will take a leading role in training frontier models at scale, aiming to make open AI a true platform for developers.
The partnership builds on past work together, including the Mistral Nemo model. It was announced March 16, 2026, at NVIDIA GTC. Mistral Small 4 is out now under the Apache 2.0 license.
Why it matters: Open-source AI has long struggled to keep up with closed models. The Nemotron Coalition is built to fix that. Pooling Mistral's architecture skills with NVIDIA's compute gives this effort real firepower. If it works, developers get access to frontier models they can actually own and customize. For the open-source community, this could shift the balance. For the big closed labs, it's the first real challenge in a while.
TOOLS
Trending AI Tools
🔵 Mistral Small 4 - Mistral's new AI combining reasoning, coding, and vision
🔥 NemoClaw - Nvidia's open-source security layer for OpenClaw agents
🎨 Gamma Imagine - Gamma's AI-native visual design tool
🚀 Dispatch - Claude Cowork's new feature lets you assign tasks from your phone to Claude on your desktop
NEWS
What Matters in AI Right Now?
Manus just launched My Computer, a desktop app that takes its AI agent off the cloud and onto your local machine, where it can manage files, run commands, and build apps.
Adaptive just launched Adaptive Computer, an AI PC that runs software for you, with an Encoded Memory feature that learns your tools so it can repeat tasks on its own later.
Nebius just signed a five-year deal with Meta worth up to $27B, with $12B of that going to AI cloud capacity built on NVIDIA's Vera Rubin hardware, starting in early 2027.
Okara just launched what it calls the world's first AI CMO, with agents that run SEO, content, Reddit, and Hacker News for $99/month, versus $60K-$160K per year for a human team.
Perplexity just expanded Computer Agent to Android and its Comet browser, with Comet giving the agent full control of any logged-in app or website without needing extra setup.
OpenAI just launched Subagents in Codex, letting developers run multiple AI agents at once to split up a task and work on each part in parallel.
NVIDIA just unveiled DLSS 5, its biggest graphics leap since ray tracing in 2018, with an AI model that adds real-looking lighting and materials to game frames in real time at up to 4K.
Mistral AI just released Mistral Small 4, a 119B open-source model that handles reasoning, vision, and coding in one, with a 256k context window and 40% faster speeds, free under Apache 2.0.
xAI just released a Text-to-Speech API for Grok, letting developers build voice agents with tool calling, MCP support, and web search built in.
🧡 Enjoyed this issue?
🤝 Recommend our newsletter or leave a feedback.
How'd you like today's newsletter?
Cheers, Jason
