AI labor, cloud spend, and space bets — three very different reasons venture money moved today. And for once, the leads have names attached. This is Startup Fundraising. After a week of unnamed leads and sovereign funds, today's slate is smaller and — — actually disclosed. Which means I get to underwrite instead of just yell. Two stories today where the money has somewhere real to go. Let's start with Stepful — fifty-five million, Series C, Oak HC/FT leading. Let's get into it. Oak HC/FT is a healthcare-and-fintech specialist, not a generalist chasing the AI label. And Y Combinator plus a hospital group were already on the cap table before this round. That hospital money is the part I like — it's a customer writing checks, not just a fund. The problem's even quantified: ninety-seven billion in annual agency staffing spend. Strategic customer validation ahead of the C — that's a cleaner signal than any valuation headline. My one worry: if AI accelerates certification that well, what stops a big hospital system from just building it in-house? Show me the retention before I believe the moat. PointFive — sixty million from Accel to manage cloud and AI costs. So we've now funded a company to clean up the spending the rest of the portfolio created. And it's one named lead, solo. After a week of three-co-lead structures where you can't tell who set the price, that's almost refreshing. Ninety-six million total since 2023. Accel leading the B here says cost-optimization may become permanent infrastructure instead of a recession reflex. That's exactly my question — what does the retention curve do once a CFO gets religion on discipline and just hires two analysts instead? Spending tools are sticky until the spending stops. Moonshot's back — seeking thirty billion, up seven-x in six months. But today we finally have a number: ARR topped two hundred million in April. Two hundred million ARR against a thirty-billion ask. That's a hundred-and-fifty-times multiple in the most commoditized AI market on the planet. At that multiple, you're pricing in dominance. Revenue barely moves the math. And what happens to the margin when every Chinese lab races to compress price? Revenue's nice — I want to see what's left after the price war. Smallest round on the board: NewOrbit, thirteen-point-nine million, Series A for satellites in Very Low Earth Orbit. Two hundred to three hundred kilometers — historically spy-satellite territory, no commercial playbook. Honestly the most honest risk profile here. Who's the anchor customer, and is this a product or a government contract in disguise? Capital-intensive hardware that takes years to validate off a thin early check — same structural worry I had on the deep-tech pre-seeds this week. The frontier's real; the timeline's brutal. Here's The Next Web:
America’s hospitals are spending around $97bn a year renting staff they cannot train fast enough. Stepful thinks AI can fix the supply side, and investors have just handed it $55mn to try. The New York startup has raised a Series C led by healthcare-and-fintech investor Oak HC/FT, with new backers Foresite Capital, Hearst Ventures, and the Citi Impact Fund joining existing investors including Y Combinator and the hospital group Intermountain Health.
Stepful's Series C — $55 million, led by Oak HC/FT. After a week of leads nobody would name, here's a fund with an actual thesis: healthcare and fintech, picking a healthcare company on purpose. And the number underneath it is one I can actually do something with — hospitals burning roughly $97 billion a year renting staff they can't train fast enough. That's a real wound; you can underwrite it. And look who was already on the cap table — Y Combinator and Intermountain Health. A hospital group writing checks before the C tells you the enterprise validation came before the round, not after. 32,000 graduates, 35-plus systems including Mount Sinai and Ochsner. My question is the retention curve — does AI accelerate certification fast enough to matter before a Providence just decides to build the pipeline in-house? When ICEYE puts together a €1 billion capital structure — equity, secondary, revolving credit — what are investors buying? A venture-scale software story, or basically a defense prime that still has 'startup' in the press release? Honestly, the capital structure basically answers it: ICEYE is a hybrid beast, and it looks deliberate. In December 2025, the equity piece was €150 million of new money led by General Catalyst at a €2.4 billion valuation, plus €50 million of secondary for early investors. That's the venture layer. Then right next to it, there's a €300 million revolving credit facility, arranged by a bank syndicate with Citi and Danske Bank as joint global coordinators, per Castrén & Snellman. That facility is there to backstop liquidity and issue guarantees for customer contracts — very defense-prime financing. And the operating numbers make the debt look conservative rather than speculative: per Seraphim Space's reporting on ICEYE's unaudited 2025 results, more than €250 million in revenue, more than €100 million in EBITDA, more than €350 million of cash on hand, and a contracted backlog of €1.5 billion. CEO Rafal Modrzewski even said the equity raise was 'not strictly necessary' because ICEYE had already reached profitability, which is not something you hear from a typical Series C founder. So yes, investors are underwriting a government-procurement-anchored revenue base — seven European nations, including Poland, Germany, and Finland, have already signed SAR satellite agreements — but wrapped in software-ish margins and a venture growth rate. Revenue more than doubled year over year and beat the company's own projections by 25%. If the backlog is €1.5 billion and they're already cash-flow positive, what's the actual risk here — satellite production execution, or customer concentration inside a single geopolitical moment? Both are real, but I'd watch the geopolitical concentration more closely. ICEYE's entire demand surge traces directly to the U.S. halting intelligence sharing with Ukraine in March 2025, which per Defense News reporting crystallized European anxiety about sovereign access to space-based data and turbocharged ICEYE's pipeline. Powerful tailwind, sure, but it's also a single thesis: if the political environment shifts and U.S. data sharing resumes at scale, or European defense budgets stall, the backlog converts slower than the model assumes. The revolving credit facility — structured around contract guarantees rather than pure liquidity — suggests the banks have thought carefully about exactly that execution-against-backlog risk. The Next Web, with Cristian Dina:
Moonshot AI is seeking a $30bn valuation in its third funding round in six months, up from $4bn in December. The Kimi developer’s ARR topped $200m in April as Chinese AI labs race to match American peers before IPO windows open.
We finally have a number to hang this on. Moonshot topped $200 million ARR in April, and they're chasing a $30 billion valuation — that's 150 times forward revenue. Up sevenfold in six months, and it's the third round in that window. The Meituan-led round at $20 billion isn't even closed yet, and they're already pitching the next one at thirty. Two hundred million ARR is real, I'll take it. But this is the Kimi chatbot — the single most commoditized corner of AI on the planet, where every Chinese lab is racing the others to zero on price. So what's the margin look like when DeepSeek's raising seven billion right next door at a fifty-nine billion mark? You're underwriting a price war and calling it dominance. And note who's setting this price — Bloomberg's got the number, but the syndicate's still a question. We've spent all week on unnamed leads. This one's no different yet. At $200 million ARR, what has to be true in two years to make thirty billion sane? Either they hold pricing in a knife fight, or somebody's betting on a Shanghai IPO window cracking open before the music stops. Ventureburn writes:
PointFive announced it has raised $60 million of funding to gain more control of soaring cloud and AI expenditure. The company’s round was led by Accel and saw participation from other venture capital firms including Index Ventures, Salesforce Ventures, Entre Capital, Perpetual Growth, Vesey Ventures and Sheva Ventures.
PointFive — $60 million Series B, Accel leading solo. After a week of co-leads and unnamed-lead parades, one named lead on a clean B is almost refreshing. So there's now a whole funding category built on cleaning up the bill that every other funded AI company ran up. Sixty million to tell CFOs their compute spend is out of control. This round more than doubles what they'd raised before — $96 million total since 2023, with $60 million of that today. Accel's betting cost-optimization becomes permanent infrastructure instead of a recession-era austerity tool. That's the retention question, though. A CFO gets cost discipline religion, signs PointFive, then realizes they can hire two analysts and a dashboard. What's the renewal curve look like once the panic wears off? The bet is that AI spend stays unpredictable enough that you never stop needing the visibility. Large model deployments, compute spikes — you don't solve that once. Fair. If the mess is structural, the broom is recurring revenue. Index, Salesforce Ventures in the syndicate too — Salesforce money usually means there's a real enterprise pipe behind it. Here's Kirstie Pickering at UK Tech News:
European satellite manufacturer NewOrbit has closed a $18.5m (£13.9m) Series A to open near-Earth orbits to commercial space flights. NewOrbit is building satellites for Very Low Earth Orbit (VLEO), a band of near-Earth space at 200-300km above ground historically reserved for spy satellites and the International Space Station. By flying closer to Earth, NewOrbit’s satellites deliver drone-resolution imagery and direct-to-phone connectivity at a fraction of the cost of conventional constellations.
Smallest round on the board today — eighteen and a half million bucks — and it's the one with the most honest risk profile. NewOrbit's flying satellites at 200 to 300 kilometers, the band that's been spy-satellite-and-ISS-only for sixty years. Voyager Ventures led, it was oversubscribed, and the cap table reads like a deep-tech angel reunion — former NVIDIA chief scientist, the TIER Mobility co-founder. Real names, real checks. Here's my problem. The complex opens in 2027, first commercial satellite flies in 2028. So this £13.9m funds a multi-year buildout before a single revenue-generating thing hits orbit. Which is exactly the thin-slice-into-hardware worry — capital-intensive deep tech that takes years to validate. Eighteen million doesn't get you to 2028 alone; this is the first of several rounds, guaranteed. And the question nobody's answering — who's the anchor customer? Drone-resolution imagery and direct-to-phone at VLEO sounds like a government contract wearing a commercial jacket. Is anyone actually signed? If Startup Fundraising helps you stay sharp, please subscribe and leave a quick review wherever you're listening. It only takes a moment, and it helps other founders and operators find the show.
If a funding round, investor note, or market shift from today's briefing caught your ear, you'll find links to every story in the show notes. They're there whenever you want to dig a little deeper.
That's Startup Fundraising for today. This is a Lantern Podcast.