AI funding is piling into robots, power grids, creative data, and now the VCs themselves are raising too. Full-stack money week. Welcome to Startup Fundraising. Today we've got a $400 million robotics round, a £480 million stealth exit in the UK, a grid infrastructure play, a new Indian fund out of ex-Peak XV hands, and a data supplier slipping into the AI supply chain. Five rounds, five different bets on where AI actually needs infrastructure. Let's see which ones make sense and which ones are just chasing the acronym. First up: the robots. This one's from Ventureburn:
Mind Robotics has raised $400 million in fresh funding. The round was led by Kleiner Perkins and brings the company’s total capital to more than $1 billion. The new valuation now stands at $3.4 billion. The company reached this figure less than six months after it was created.
Mind Robotics, spun out of Rivian in 2025 and still less than six months old, just closed a $400 million round led by Kleiner Perkins. That pushes the valuation to $3.4 billion and total capital past a billion. Ventureburn had the story. Six months old and already at a $3.4 billion valuation. What's the revenue number? Because 'robots in Rivian's factories' is a captive customer, not a market. To be fair, Kleiner leading — not just tagging along — is a real signal. And using the Rivian factories as live training grounds is an actual moat argument, not just slide deck fluff. Sure, if Rivian survives and keeps scaling. In two years, this needs at least two non-Rivian enterprise deployments, plus some proof the hardware margins don't fall apart once they leave the friendly home court. This one's from Resultsense:
The 25-person team is pursuing “open-ended” self-improving AI through automated scientific discovery, described in the firm’s own statement as “likely the fastest path to superintelligence”. The framing is unusually direct — most well-funded AGI labs have moved away from explicit superintelligence rhetoric — and is more aligned with the founders’ published research backgrounds in quality-diversity algorithms, AI-generating algorithms, and self-improving coding agents.
Recursive Superintelligence, founded last year and staffed by twenty-five people with no product in market, is out of stealth at three-and-a-half billion pounds on a £480 million round. GV and Greycroft are leading, with AMD and NVIDIA coming in as strategic participants, which means the biggest checkwriters also have a hand on the compute tap. One year old, two dozen employees, and the pitch is literally 'fastest path to superintelligence.' Somebody's going to have to show me what £480 million buys you in year two of a company that hasn't shipped anything. To be fair, the bench is real — ex-OpenAI, DeepMind, Meta AI, Salesforce, and Uber research leads. And London HQ matters; this isn't just a UK mailbox with a San Francisco soul. But that strategic-cap table is worth watching, because AMD and NVIDIA aren't in this for the upside alone. Right, the chip makers own a piece of the lab that needs their chips. That's not a conflict — that's a subscription model dressed up as conviction. What has to be true in twenty-four months for £3.5 billion to look cheap? I don't think we've heard it. Business Wire writes:
GridCARE, the pioneer of Power Acceleration for AI, today announced the closing of its $64 million oversubscribed Series A financing, representing a significant step-up in valuation from its previous round less than a year ago.
GridCARE just closed a $64 million Series A, led by Sutter Hill Ventures. They're also name-dropping John Doerr's early bets on NVIDIA and Google in the press release, which, sure — but that's a lot of pedigree to put at the front for a company most people haven't heard of. They coined 'Power Acceleration for AI' in the same breath as announcing the round. That's not a category, that's a marketing deck. What does GridCARE actually ship, and who's paying for it today? To be fair, Sutter Hill leading is a real signal. They don't spray and pray. But 'significant step-up in valuation from less than a year ago' with zero numbers attached is the kind of line I read twice and trust half. The power constraint thesis isn't wrong — datacenter operators are yelling about it for a reason. The question is whether GridCARE is selling a product, or just a story that rides that wave into a bigger round. Inc42 Media writes:
Former Peak XV Partners MDs — Ashish Agrawal, Ishaan Mittal, and Tejeshwi Sharma — have launched Mettle Capital, a new India-focused VC firm, and are aiming to raise $350 Mn to $400 Mn for its maiden fund. This comes a few months after the trio quit Peak XV and said they would launch a new VC firm.
Three ex-Peak XV MDs — Ashish Agrawal, Ishaan Mittal, and Tejeshwi Sharma — have put a name to what they telegraphed months ago: Mettle Capital. They're targeting $350 million to $400 million for a maiden India fund focused on Series A and B, with some seed. Brand-name pedigree is real here — Groww at 94X, Razorpay on IPO watch — but 'enterprise AI, deeptech, and consumer internet' is basically every category that exists. What's the actual filter? The answer they'd give you is concentrated bets: five or six deals a year, high conviction, not spray-and-pray. Whether LPs buy that in one quarter is the live test. Raising $400 million with no deployed track record outside a bigger platform is a trust-me moment. The LP question is whether the returns belonged to the individuals, or to the Sequoia-then-Peak XV network they were plugged into. This one's from TechCrunch:
The startup said on Thursday it has raised $23 million in Series A funding to build out the new data supply business. The round was led by Nava Ventures and saw participation from SBVP (co-founded by Sheryl Sandberg), Formula VC, and I2BF Ventures.
Wirestock just closed a $23 million round — TechCrunch had it — and this is a clean pivot story: stock photo distributor to AI data supplier, with 700,000 artists on the platform doing tasks like a creative Mechanical Turk. They made the pivot in 2023, so they've had three years to prove it. I want to see the revenue before I get excited. 'Majority of artists opted in' is doing a lot of work when they won't give us the actual number. The business model shift is real, though. Moving from off-the-shelf library sales to custom data-collection contracts is a totally different margin profile, and that's probably what unlocked the round. Sure, but what has to be true in two years? AI labs consolidate, build their own pipelines, or the legal exposure around licensed creative data gets worse — and suddenly your 700,000 artists are working for a client list that just got cut in half. Got a fundraising question, a story idea, or a correction for us? Send a note to startupfundraising at lantern podcasts dot com. We read the inbox, and your feedback helps shape future episodes. You'll find links to every story we covered today in the show notes, so if something caught your ear, you can dig into the details there. Thanks for listening, and have a great Friday. That's Startup Fundraising for today. This is a Lantern Podcast.