Index Ventures closes a three-and-a-half-billion-dollar war chest, then leads a two-hundred-and-twenty-seven-million-dollar round in the same news cycle. Same firm, both sides of the page. New to this story? Here's where it stands. AI venture money keeps piling into companies staking out foundational turf — model platforms, infrastructure, applied tooling, physical AI. Lately, Generalist AI has been chasing a three-billion-dollar valuation. An Israeli cyber startup is raising a hundred and thirteen million dollars to secure autonomous agents. Flapping Airplanes, meanwhile, is reportedly talking about hundreds of millions at a five-billion-dollar valuation. This is Startup Fundraising. Today — a six-month-old Shanghai brain-chip startup breaking angel records, an eight-million-dollar seed betting it can outrun CrowdStrike, and one fund I've got questions about. Augment's up first. Ship or Skip writes:
Augment Code announced a $227 million Series C round led by Index Ventures, pushing its valuation past $2 billion. The enterprise-focused AI coding assistant has positioned itself against GitHub Copilot and Cursor by emphasizing deep codebase indexing — the ability to understand context across millions of lines of code rather than just the file currently open.
Two hundred twenty-seven million into a company fighting Copilot, Cursor, Codeium, and Tabnine for the same enterprise buyer. The pitch is codebase-wide context indexing. Great — that's a real engineering problem. But at two billion, they have two years to turn legacy-code understanding into a moat. Everyone else shipping it as a feature by Q3 blows up that bet. And here's the cap-table piece — Index led this. Same Index that's all over today's rundown on the fund side. Nobody's seed sleeve can write a check this size at a two-billion-dollar valuation; this is the growth fund's first visible move. There's your AI venture concentration update: Augment Code just joined the two-billion-plus AI coding club, and Index is on the term sheet. Which is the part that bugs me. If the same GP writes this Series C and runs the growth fund that has to lead the Series D, where does price discipline come from? Who's negotiating against whom? Right — the mark on the next round gets set by the person holding the last one. The headline valuation is one number. Before I call this a clean win, I want to see the liquidation stack and the option pool. The money's going to agentic code review and large-scale understanding. Fine. Those are product bets, not revenue. Show me how many engineering orgs switched off Copilot because Augment held more of the repo in context. That's the number I need to justify the round. Here's TechNode:
Shanghai-based brain-computer interface startup Active Technology (our translation of 主动科技, the company’s Chinese name) has completed a RMB330 million angel round, setting a record for angel financing in China’s brain-computer interface sector. The company focuses on invasive brain-computer interface technology and was established in February 2026.
A company founded in February just raised three hundred thirty million yuan — call it forty-five million dollars — in an angel round. Five months old. No product, no clinical data. Angel round in name only. With Lenovo Star and Zhongke Chuangxing writing the lead check, nobody's uncle is betting on a garage here. And it's invasive BCI. So what has to be true in two years? They'd need regulatory clearance to put electrodes in a human skull in China. You're looking at a decade, not five months. The money's going to a GMP workshop and a ten-thousand-square-meter headquarters by year-end. A chunk of this round is concrete and clean rooms, alongside the decoding algorithms. Right — you're funding a building and a team. That record-setting label is selling the whole story. This one's from RuntimeWire:
AI agents turn software permissions into operational authority. Arrakis is competing to own the enforcement layer before identity, endpoint and cloud vendors fold it into existing platforms. Tal Baron, Omer Efrat and Ron Shani disclosed an $8 million seed round for Arrakis Security on August 2, giving the Torq and Palantir veterans capital to build runtime controls for enterprise AI agents, CTech reported.
Arrakis has an eight-million-dollar seed and a founding team out of Torq and Palantir. The whole bet is that they own agent runtime governance before CrowdStrike or Palo Alto ships it as a checkbox. That buys maybe eighteen months to prove it. Is that enough runway to stay independent? Credit to CTech for the break. And look at the lead — Hetz Ventures, not Bessemer. That's the third distinct agent-security bet on this desk in three days, each from a different investor base. Which tells me nobody's decided who owns this layer yet. Everyone's planting a flag with barely a credential check between them. Their own framing gives it away. The 'why it matters' line names the threat out loud: identity, endpoint, and cloud vendors folding this in. They pitch the bull case and print the bear case in the same paragraph. Right, so the pitch is 'buy us before the platforms eat us.' Great — you're pitching a new category and an acquisition in the same breath. Show me an enterprise actually paying to govern non-human workers today, not the ElevenLabs and Pentera CEOs on the cap table. Twenty people, and they're hiring R&D in Israel. Real team, real problem. The seed round can't tell us whether this becomes a standalone company or ends up as a feature. Here's CTech:
Index Ventures has raised $3.5 billion in new funds that will be used to invest across all stages of company building, including a significant focus on artificial intelligence, cybersecurity and infrastructure startups. The new capital includes a $400 million seed fund, a $900 million venture fund, and a $2.2 billion growth fund, expanded from the firm’s previous $1.5 billion growth vehicle.
So the fund we heard whispers about on Friday is real, and then some. Index closed $3.5 billion — a $400 million seed vehicle, a $900 million venture fund, and a $2.2 billion growth fund, up from $1.5 billion last time. Yeah, and I owe you a correction. On the first, I called this firm disciplined — smaller than the last fund, no need to show off. That growth sleeve actually grew by seven hundred million dollars, Hope. And in the same news cycle, that growth fund makes its first visible move — leading Augment Code's $227 million Series C, which we just hit. That $2.2 billion sleeve is already writing checks. So say it plainly — the same GP raises the growth fund, then leads a Series C in that exact category. If Index is still on both sides of the table for the Series D two years from now, where does price discipline come from? What jumps out to me is the $400 million seed fund underneath all that growth capital. You only build that kind of capacity at the earliest stage in AI and cyber if you plan to own the whole cap table from day one — the playbook behind Wiz, Gong, and Fireblocks. Have feedback, a story idea, or a correction? Send us a note at startupfundraising at lantern podcasts dot com. We’d love to hear from you.
One thing we’re watching: Active Technology says it plans to build a headquarters of more than 10,000 square meters by the end of 2026.
Links to every story are in today’s show notes if you want to dig into anything that caught your attention.
That’s Startup Fundraising for today. This is a Lantern Podcast.