OpenAI wants $30B more at a $1.4 trillion valuation

Good morning and happy Saturday, crowd. ☕

This was the week AI money stopped pretending to be normal. An invite-only assistant most people have never used raised $1 billion. AMD wrote an $8.2 billion check for a company that simulates the physical world. And OpenAI is reportedly asking for $30 billion more just to bridge it to its next raise.

Pour a second cup. Here's what got funded:

• 🤖 Instinct raises $1B at a $10B valuation from Sequoia, Benchmark and Coatue

• 🧠 AMD buys World Labs for $8.2B in an all-stock deal

• 💸 OpenAI launches Dots and reportedly seeks a $30B bridge at $1.4T

• 🛰️ Varda raises $251M to make medicine in orbit

• 🛡️ Armadin raises $255.5M to build AI hackers that attack you first

⏱️ Read time: 5 minutes

Read time: 5 mins

The Popcorn Brand Turning Heads and Raising Capital

What does it look like when a consumer brand starts to break out?

generated $5.8M in revenue in 2025 and is projecting $9M in 2026, with more than 100,000 customers and $20M+ in lifetime sales.

The brand has already earned national recognition through Oprah’s Favorite Things and partnerships with Costco, Sur La Table, and Williams Sonoma, where Popsmith is rolling out nationwide this holiday season.

Now is raising capital to accelerate its next stage of growth, and individual investors have the opportunity to become shareholders.

FUNDING ROUNDS

Gettty Images

Here's the strangest billion-dollar product of the year: there's no app.

Instinct is an invite-only personal AI assistant you talk to by text, voice, and email. You tell it to book a multi-city trip or grab a dinner reservation, and it just goes and does it across your digital life. This week it raised a $1 billion Series C at a $10 billion valuation, with Sequoia Capital, Benchmark, and Coatue all writing checks.

Read that investor list again. Sequoia and Benchmark rarely end up in the same round at this stage. When they do, it usually means the company had leverage and picked who it wanted.

The timing tells the real story. Instinct is going head-to-head with the giants at exactly the moment the giants are shipping their own agents. OpenAI just launched Dots, Meta has Muse, and Grok has its bots. A startup raising a billion dollars into that fight is a direct bet that distribution isn't destiny, and that the best assistant wins even without owning the phone or the social network.

The invite-only model is doing a lot of work here too. Scarcity creates demand, and demand creates a valuation. It's the oldest growth trick in consumer tech, and it still works when the product actually delivers.

But here's the thing. Agents are expensive to run, and every task Instinct completes costs real compute. A $1 billion round buys a lot of runway, but it also sets a bar. At $10 billion, Instinct needs to look less like a viral app and more like a habit for millions of people who pay.

For founders, the lesson is about the interface, not the model. Instinct didn't win this round by building a better LLM. It won by meeting users inside the channels they already use every day. The best product decisions are often about where you show up, not what you build.

AMD

AMD just made its loudest move yet to stop being "the other chip company."

On Monday, Advanced Micro Devices said it will acquire World Labs, the AI startup founded by Fei-Fei Li, in an all-stock deal valued at $8.2 billion. World Labs works on physical AI: systems that can understand and simulate the real world rather than just predict the next word.

Why would a chipmaker buy a research lab? Because the next big wave of compute demand probably won't come from chatbots. It will come from robots, autonomous vehicles, factories, and anything else that has to reason about three-dimensional space. Whoever owns the models that define those workloads gets to shape the hardware that runs them.

That's the playbook Nvidia has run for a decade: own the software layer, and the silicon sells itself. AMD is now buying its way into that same position instead of trying to build it from scratch.

The all-stock structure is the part private market investors should notice. AMD isn't spending cash. It's spending its own share price, and World Labs' backers are taking AMD stock as their exit. That's a vote of confidence in AMD, and it's also a very clean path to liquidity at a time when the IPO window still only fits the biggest names.

Expect more of this. Big public tech companies are sitting on expensive stock and want AI talent fast. Buying a world-class research team with paper is cheaper than competing for that talent one hire at a time.

For founders in deep tech, this is the clearest exit signal of the year. If your company owns a capability that a public giant can't easily build, your most likely buyer is someone paying in their own stock. Build your story around what you make possible for them, not just what you sell today.

Open Ai

OpenAI had two headlines this week, and together they explain the entire AI economy.

First, the product. OpenAI launched Dots, a persistent virtual assistant that can operate a computer, run workflows across platforms, and debug software on its own. It's OpenAI's clearest move yet from chatbot to agent, and it lands in the same week Instinct raised $1 billion to do something similar.

Second, the money. OpenAI is reportedly seeking a $30 billion bridge round at a $1.4 trillion pre-money valuation.

A bridge round. At $1.4 trillion. Let that sink in.

Bridge rounds are usually what startups raise when they need cash to get to the next real milestone. When the most valuable private company in the world needs $30 billion just to bridge, it tells you how much frontier AI actually costs to build. The race isn't won with great research alone anymore. It's won with whoever can keep raising.

Here's what's wild. Rival Anthropic's leaked IPO filing shows it plans to spend $518 billion on compute and infrastructure obligations in the coming years. OpenAI is playing the same game at the same scale. Both companies are essentially asking capital markets to finance the biggest infrastructure buildout since the railways.

For private market investors, this is the question that matters most: who's left to buy? At $1.4 trillion, the only checks big enough come from sovereign funds, hyperscalers, and eventually the public markets. That makes an OpenAI IPO less a matter of if and more a matter of when.

For founders building on top of these models, the takeaway is practical. The labs are now competing directly with their own customers. Dots does what plenty of agent startups pitched last year. If your product is a thin layer on top of someone else's model, assume they'll ship your feature eventually. Own the workflow, the data, or the customer relationship.

Varda / Traded VC

Not every big round this week was about chatbots.

Varda Space Industries raised a $251 million Series D co-led by Lux Capital and Natural Capital, with Founders Fund, Khosla Ventures, General Catalyst, Caffeinated Capital, and others joining. Varda manufactures pharmaceuticals in orbit and brings them home in reusable reentry capsules.

Why make drugs in space? Microgravity changes how certain molecules crystallize. Some of those crystal structures are hard or impossible to produce on Earth, and they can change how a drug behaves. Varda's bet is that orbit becomes a specialized factory floor, not just a place for satellites.

The timing is perfect. On Monday, SpaceX's Starship reached orbit for the first time and delivered its first real payload. Every time launch gets cheaper and more reliable, companies like Varda get a better business model. Cheap rides up and reusable capsules down are the two halves of the same equation.

The cap table is also worth studying. Lux, Founders Fund, and General Catalyst don't usually pile into science projects at Series D. A round this size says investors now see a credible path from experimental flights to repeatable, paid manufacturing runs.

For founders, Varda is a masterclass in riding someone else's wave. It didn't build rockets. It built the product that becomes viable because someone else built rockets. The best deep tech companies often look like this: they find the business that a falling cost curve suddenly makes possible, then move first.

If you're tracking the space economy, the money is shifting from launch to what you do once you're up there.

Armadin

The best defense is now an AI that attacks you all day long.

Armadin raised a $255.5 million Series B led by Andreessen Horowitz at a valuation above $2.5 billion. The company builds autonomous offensive security agents that continuously probe companies for vulnerabilities attackers could exploit.

Traditional penetration testing is a snapshot. You hire a team, they spend a few weeks trying to break in, you get a report, and by the time you fix everything, your systems have already changed. Armadin turns that into a permanent process: AI agents testing your defenses around the clock, the way real attackers increasingly do.

And attackers are absolutely using AI. That's why this category is exploding. When offense gets automated, defense has no choice but to automate too, and the companies that do it best get to name their price.

A $2.5 billion valuation at Series B tells you how seriously investors take this. Security buyers have budget, urgency, and very little patience for tools that don't work. If Armadin's agents find real holes, the product sells itself in every board meeting where a CISO has to explain the last breach.

For founders, Armadin shows a pattern worth copying. Take a service that companies already pay humans to do periodically, then turn it into software that runs continuously. You go from a project budget to a subscription, and from a vendor to infrastructure.

⚡ Quick Hits: More Money Moving

🏠 EliseAI raised $350M at a $4B valuation for AI that automates leasing, resident support, and admin work for housing operators and healthcare groups.

⚓ Quartermaster raised a $140M Series B led by Insight Partners plus a $40M debt facility from Stifel for maritime intelligence that fuses data from rugged ship sensors.

🔌 CScale emerged with $188M to build optical interconnects for data centers.

🛵 Simple Energy raised Rs 1,750 crore (about $180M) in a Series C, one of the largest rounds ever for an Indian electric two-wheeler startup.

🇧🇷 Monzo is reportedly in talks over a £10B sale to Nubank that could end its London IPO plans.

💬 Photon, backed by Vercel, raised $4.5M to put AI agents inside iMessage and WhatsApp.

💰 Get Access, a members-only club for startup investing, reportedly closed $5M at an $85M valuation.

🇵🇹 FRANK, a Lisbon startup, raised €2.9M from Armilar and Start Ventures.

📈 Moneyview listed in India at a 61.8% premium to its IPO price.

💬 What This Week Tells Us

Every big round this week shared one thing: the money went to companies doing the work, not just talking about it. Instinct books the trip. Armadin runs the attack. Varda makes the drug. AMD bought a team that simulates the physical world. The agent era has a price tag now, and investors are paying it.

Question for you: If you could get into one of these at this week's price, which would it be: Instinct at $10B, Armadin at $2.5B, or Varda before space manufacturing goes mainstream? Hit reply. We read every one.

Hunt accordingly. ☕

Jose, Alberto & the Founderscrowd Team