4211781
The official channel of V3V Ventures. We share updates on our investments, portfolio companies, and fund activities. Buy Ads: @strategy (this is our only account).
⚖️ LegalOn raises $50M to speed up legal work with AI
Japanese startup LegalOn is building AI tools to help lawyers review contracts faster and more accurately.
The product is already widely used — and now the company is doubling down on the US and UK markets.
🖱 AI scans legal documents, flags risks, and suggests edits
🖱 Cuts review time by up to 8x, according to the company
🖱 7,000+ companies already use it across Japan, the US, and UK
🖱 In Japan, 25% of all public companies are LegalOn users
🖱 New funding round led by Goldman Sachs brings total to $200M
🖱 Actively building AI legal agents, with early access to OpenAI models
LegalOn is trying to become the default AI co-pilot for contract review — starting with scale, not hype.
📊 Key Venture Highlights of July: AI, Infrastructure, and Deeptech Dominate
The July venture landscape was shaped by record-breaking deals in AI, a resurgence in U.S. investment volumes, and a clear shift toward infrastructure-focused tech platforms.
AI mega-deals dominate the month
🟡Anthropic secured nearly $10B in fresh capital from Google, Amazon, and Salesforce Ventures, pushing its valuation above $100B — a milestone for private AI.
🟡 Elon Musk’s xAI closed a $10B round at an $80B pre-money valuation, backed by Morgan Stanley and others.
Venture funding surges in the U.S.
🟡 H1 2025 saw $162.8B in investments — up 75.6% YoY, with nearly 65% of capital flowing into AI.
🟡 Despite tight fundraising conditions, this is now the second-biggest year for VC in U.S. history.
Other notable July deals
🟡 LangChain (LLM infra): $100M led by IVP
🟡 Substack: $100M Series C, reaching a $1.1B valuation
🟡 Moment (fixed-income infra): $36M from a16z, Index, Contrary
🟡 PitchBook highlights Anthropic, Grammarly, and others as “object-oriented” investment outliers
📌Trends to watch
🖱AI is the market: Most major July deals center on AI or LLM infrastructure
🖱Deeptech resurgence: 65% of U.S. funding now targets deep infrastructure, not just consumer apps
🖱Private unicorn surge: Players like Anthropic and xAI are setting new valuation benchmarks
🖱LLM and Web3 infra still hot: LangChain and Moment show that infra demand remains strong
Investor optimism holds: Despite macro tension, venture outpaced expectations
Venture capital is rotating toward core infrastructure bets in AI and deeptech. These aren’t just hot narratives — they’re becoming the backbone of the next market cycle.
💰🧪 “Cash or Cure”: biotech investors demand capital back
Activist funds are targeting biotech startups that sit on huge cash reserves but have little to show in terms of drug development.
Roughly 1 in 6 companies in the Nasdaq Biotech Index is now worth less than the cash it holds on its balance sheet. That’s billions in idle capital — and investors are done waiting.
🖱 Funds estimate up to $30B is frozen in underperforming biotech stocks
🖱 Activists urge firms to kill stalled projects and return cash to shareholders
🖱 Tang Capital already shut down Allakos, Cargo Therapeutics, and iTeos
🖱 Acelyrin was pushed into an acquisition after similar pressure
🖱 Sector hype peaked after COVID and obesity drug wins, but approvals have slowed
For many funds, the logic is simple: if you’re not going to cure anything — pay us.
📣 Top OpenAI investor melts down after ChatGPT “reveals secret system”
Geoff Lewis, founder of Bedrock Capital and early OpenAI backer, posted a 3.5-minute monologue claiming he’s being targeted by a “non-governmental system.”
He says the system:
🖱 isn’t visible, but structurally real
🖱 doesn’t censor content, it distorts signal
🖱 isolates “recursive” people and replaces them with mirrored versions
🖱 leaves you “alive, but invisible”
💬His words:
People say, ‘he’s brilliant, but something’s off.’ That’s the pattern.
A fund built on betting against consensus is now fighting phantoms.
⏫ 2025 could become a record year for startup acquisitions.
Carta analyzed data from companies on its platform for the first half of the year:
🖱 126 exits from companies that raised $1M–$10M
🖱 100 exits from startups with $10M–$50M in funding
🖱 48 exits from startups that raised over $50M
If this trend continues in the second half of the year, we may see a new record for M&A activity.
🤖 Robots in the Metro: China’s New Delivery Frontier
China's robotics game isn’t just about humanoids.
In Shenzhen, a new kind of delivery robot is being tested — not for streets, but for underground metro tunnels.
🖱 The goal: solve a growing logistics issue for 7-Eleven stores inside metro stations (over 100 locations).
🖱 Traditional delivery vans are failing — traffic jams and no parking near stations slow down restocking.
🖱 New autonomous robots now deliver goods directly inside the metro system.
🖱 These rover-like bots navigate the underground routes, bringing stock right to store doors.
The tests are going well so far. And they might just redefine last-mile delivery — one tunnel at a time.
📊 Crypto Venture Weekly: July 14–18, 2025
$202M raised across 19 projects last week — with capital pouring into stablecoins, crypto banking, DeFi infra, and tokenized finance.
Here’s what the top 10 are building 👇
⚪️KUN ($50M, GSR Ventures, BAI Capital)
Asia-based embedded finance and stablecoin platform. Licensed in Hong Kong and Singapore, offering cross-border payments, card issuing, and on-chain B2B finance tools.
⚪️ Spiko ($22M, Index Ventures, Bpifrance)
Tokenized money market fund platform with €400M AUM. On-chain T-Bill exposure for startups and SMEs, fully compliant and integrated with European custodians.
⚪️ XMTP ($20M, a16z, USV, Coinbase)
Cross-chain messaging protocol powering wallet-to-wallet DMs, notifications, and announcements. Already used by apps like Lens and Farcaster.
⚪️ Blockskye ($15.8M, United Airlines Ventures, Lightspeed)
Building blockchain infra for corporate travel. Used by major airlines and agencies to automate settlement, invoicing, and policy enforcement.
⚪️ Dakota ($12.5M, CoinFund, DCG)
Crypto-native B2B banking layer offering fiat rails, treasury tools, and compliance for Web3 startups and DAOs.
⚪️ AI Cross Matrix ($12M, Avalon Wealth Club)
Privacy-focused protocol enabling cross-chain, untraceable crypto transfers. Uses zk tech to anonymize flows between chains.
⚪️ ƒBTC ($10M, Galaxy Ventures, Mantle)
DeFi platform unlocking Bitcoin for staking, lending, and liquidity provisioning. Aims to create native yield products for BTC holders.
⚪️ W3 ($7M, Framework Ventures, Blockchange)
Decentralized compute protocol powering AI agents. Offchain workflows, verifiable compute, and token incentives for distributed AI infra.
⚪️ Football.Fun ($2M, 6th Man Ventures, Zee Prime)
Onchain football betting and prediction markets. Focused on gamified speculation around global matches.
⚪️ MoneyBadger ($400K, P1 Ventures)
South African QR-based crypto payments app for retail and e-commerce. Live with 60+ merchants in pilot phase.
💡Real-world utility and crypto infra are in focus — from stablecoin rails to tokenized yield and offchain compute.
📝Every Y Combinator startup tip on one screen
Since 2005, Y Combinator has invested in over 4,000 companies that have a combined valuation of over $600B.
Here's what they consider their most important, most transformative advice for startups:
🖱 Launch now
🖱 Build something people want
🖱 Do things that don't scale
🖱 Find the 90 / 10 solution
🖱 Find 10-100 customers who love your product
🖱 All startups are badly broken at some point
🖱 Write code - talk to users
🖱 "It’s not your money"
🖱 Growth is the result of a great product not the precursor
🖱 Don’t scale your team/product until you have built something people want
🖱 Valuation is not equal to success or even probability of success
🖱 Avoid long negotiated deals with big customers if you can
🖱 Avoid big company corporate development queries - they will only waste time
🖱 Avoid conferences unless they are the best way to get customers
🖱 Pre-product market fit - do things that don’t scale: remain small/nimble
🖱 Startups can only solve one problem well at any given time
🖱 Founder relationships matter more than you think
🖱 Sometimes you need to fire your customers (they might be killing you)
🖱 Ignore your competitors, you will more likely die of suicide than murder
🖱 Most companies don't die because they run out of money
🖱 Be nice! Or at least don’t be a jerk
🖱 Get sleep and exercise - take care of yourself
Something every early-stage startup should have on the wall.
🤖OpenAI launches ChatGPT Agent — AI that acts, not just chats
OpenAI has quietly rolled out a powerful new feature: ChatGPT Agent, a universal assistant that goes beyond answers to take action across apps and services.
Think of it as Operator + Deep Research + coding copilot — now in one seamless chat interface.
🖱 Clicks through websites and navigates UIs
🖱 Connects to Gmail, GitHub, Calendar, and more
🖱 Uses terminal and APIs to run real commands
🖱 Scrapes and organizes data across the web
🖱 Writes and executes code on the fly
Agent Mode is available now via the ChatGPT app (Pro, Plus, and Team tiers).
On benchmarks, it doubles the previous models on Humanity’s Last Exam (41.6% vs ~20%) and hits 4x the score on complex math (FrontierMath 27.4% vs 6.3%).
This is no longer chat — it’s execution.
📝 Top Books to Level Up Your Venture Mindset
Venture capital isn’t just about deals.
It’s about strategy, analysis, and spotting trends before others do.
Here’s a list of books to help you understand how startup investing really works 👇
🟡 Venture Deals — Brad Feld & Jason Mendelson
How term sheets, deal structures, and negotiations really work.
🟡 Secrets of Sand Hill Road — Scott Kupor
How VCs make decisions and what founders need to know about the system.
🟡 Super Founders — Ali Tamaseb
A data-driven breakdown of unicorns and what made them stand out.
🟡 The Power Law — Sebastian Mallaby
The history and logic behind power law outcomes in venture.
🟡 High Growth Handbook — Elad Gil
Scaling after PMF: hiring, fundraising, and managing hypergrowth.
🟡 Creative Capital — Spencer Ante
The story of Georges Doriot, one of the fathers of modern venture.
🟡 The Art of Startup Fundraising — Alejandro Cremades
A practical guide to raising money and building investor relationships.
🟡 Blitzscaling — Reid Hoffman & Chris Yeh
Why VCs back aggressive scale and how to build for massive growth.
🟡 The Power of Unreasonable People — John Elkington & Pamela Hartigan
Spotting game-changing entrepreneurs, with a focus on social impact.
🦄 Top unicorns by revenue multiple
7x is the standard SaaS multiple — but AI startups are way above that.
Second place goes to Bolt (the fintech, not the taxi app).
They raised at a $11B valuation in 2021–22.
That number never moved — but the revenue didn’t grow either.
Still looks like a unicorn.Читать полностью…
Just don’t check the denominator.
📈 Cognition acquires Windsurf’s core team and product
After Google took Windsurf’s CEO and top researchers in a $2.4B deal, Cognition moves fast to buy the rest — including IP, product, and ~250 employees.
🟡 Cognition is integrating Windsurf’s AI IDE into its own AI engineer, Devin
🟡 Windsurf had 350+ enterprise clients, hundreds of thousands of users
🟡 ARR reached $82M before the breakup
🟡 OpenAI’s $3B offer expired prior to Google’s acquisition
🟡 The remaining Windsurf team joins Cognition for continued product development
A rare split-acquisition in AI: Google got the researchers, but Cognition walked away with the engine and the builders.Читать полностью…
💎 Hidden Gems in Claude Code: How to unlock 125x more context
Almost no one knows this — Claude Code responds differently depending on how you phrase your prompt.
Say “think” and it uses 1K tokens.
Say “ultrathink” or “megathink” — and it loads 128K tokens.
That’s 125x deeper reasoning for the same question.
⚙️ Prompt triggers that actually work:
• “think” or “think about” — for fast ideas
• “think hard” — for complex problems
• “think deeply” — for system design
• “ultrathink” — for full-stack deep dives
This isn’t magic — it’s engineered behavior. Claude Code is wired to allocate different compute depending on trigger words.
Example:
❌ Bad: “Can you check what I can improve in my app?”
✅ Good: “ultrathink about JavaScript optimization in frontend component X”
💰Microsoft commits $4B to AI workforce training
Microsoft is investing over $4 billion to teach 20 million people how to work with AI — from schools to nonprofits, across 10+ countries.
🟡Includes grants, cloud credits, and free AI tools
🟡Launching Elevate Academy to upskill adults
🟡Supporting “Hour of AI” with Code.org
🟡Building a national AI teacher center with OpenAI & Anthropic
One of the biggest education pushes in AI history — and a clear signal: the AI labor gap is now a global priority.Читать полностью…
🧠 Eric Schmidt: AI agents will kill user interfaces
The former Google CEO says every CEO, developer, and product lead should be thinking about one thing — the end of interfaces as we know them.
In the future, AI agents will handle tasks directly.
No buttons, no menus — just outcomes.
✔️Powered by Trade Watcher
🎯 The Art of Saying No: Why Focus Wins in Venture
In venture, the temptations are endless — shiny trends, big markets, exciting ideas. But winners aren’t those who try everything. They’re the ones who pick what matters and make it the best.
As Steve Jobs said: “I’m as proud of what we don’t do as I am of what we do.”
🖱 Focus fuels faster growth
Growing companies must avoid the trap of chasing everything — and double down on what truly matters.
🖱 Depth beats distraction
Becoming world-class in one area beats being average in five.
🖱 Strategic refusal is power
Success often means turning down good opportunities to stay aligned with the right one.
Focus isn’t a constraint. It’s a weapon.
⏫ Fastest-growing professions by 2030
The job growth forecast is based on a survey of 1,000 major international companies.
✔️Powered by Trade Watcher
🔍 Perplexity hits $18B valuation after fresh $100M round
Just two months after its last raise, Perplexity has secured another $100M — pushing its valuation from $14B to $18B.
The surge comes after a major partnership with Indian telecom giant Bharti Airtel. Under the deal, all 360 million Airtel users will get a free 1-year subscription to Perplexity Pro (normally $20/month).
Distribution in scale — the AI way.
💠 Venture Bingo: Due Diligence Edition
If you've invested in at least 10 startups —chances are, you already have your own bingo card.
🖱 “The founder is super charismatic” — ✅
🖱 “AI” mentioned 12 times in the deck — ✅
🖱 “Unit economics? We’re too early for that” — ✅
🖱 “Pre-revenue, but TAM is $100M” — ✅
🖱 “Still in stealth, but two friends at BigCo said it's interesting” — ✅
We’ve gathered the classic phrases you hear over and over in due diligence.
🎬 4 documentaries that show how venture capital really works
Venture isn’t just term sheets and growth charts — it’s founders, mistakes, ambition, and billion-dollar bets.
These four documentaries reveal the human side of the game — no lectures, no YouTube shorts.
⚪️Something Ventured
How venture capital began in the U.S.
The early days of Apple, Intel, Genentech, and Cisco — and the investors who backed them before the market even existed. A deep dive into risk, vision, and the birth of Silicon Valley.
⚪️ Startup.com
The rise and fall of govWorks during the dotcom era
Part startup doc, part emotional drama — this one captures how fast things collapse when ego, friendship, and execution fall out of sync.
⚪️ The Inventor: Out for Blood in Silicon Valley
The Theranos disaster
Elizabeth Holmes convinced the world — and top VCs — without a working product. A $900M cautionary tale about blind faith, charisma, and due diligence.
⚪️ The Social Dilemma
The dark side of viral platforms
Not strictly about VC, but essential for product-focused investors. Ex-employees from Facebook, Twitter, and Google break down how design choices shape addiction — and what that means for startup due diligence.
🤖 Kimi-K2 tops LMArena as best open-source model
Moonshot’s Kimi-K2 has become the highest-ranked open-source model on LMArena, surpassing DeepSeek and claiming the #5 spot overall.
🖱 Strongest in code generation and agent-style tasks
🖱 Outperforms other open models on multiple benchmarks
🖱 Upcoming features: multimodal support and advanced reasoning mode
China continues to push the frontier in open AI — and Kimi is now leading the pack.
📊 How founder ownership changes after each funding round
Carta data shows how equity is distributed after each round — with a clear difference between software and Deep Tech startups.
🖱 Deep Tech founders retain less ownership at every stage
🖱 In SaaS, founders often keep 30–40% after Series A
🖱 In Deep Tech, it drops to 15–25% by Series B
🖱 Statistically, founders are more likely to be replaced after Series B, though it’s not the norm
🧠 The more capital you need, the more ownership you give up — especially in hard tech.
💡 Most popular chatbot apps
Microsoft is spending billions on Copilot — but it’s hopelessly behind OpenAI, the company it invested $13 billion into.
✔️Powered by Trade Watcher
🎶 Mantis VC closes $100M fund to double down on B2B
Founded by The Chainsmokers, Mantis has raised a $100M third fund — up 25% from their previous $80M, defying the broader VC slowdown.
🟡 Fund III comes in at $100M, 25% bigger than Fund II
🟡 Portfolio includes B2B bets like Chainguard and Rogo
🟡 Unlike most celeb VCs, they avoid consumer plays
🟡 Use access to Fortune 500 (via private shows) to help with intros
🟡 Still lean team, but deep enterprise network leverage
Not just another celebrity fund — Mantis is turning concerts into go-to-market strategy.Читать полностью…
🚀How Notion went from failure to 30M+ users
In 2015, Notion was nearly shut down.
The first version was clunky, confusing — and funds were running out. Only a few months left to fix it.
💵SoftBank's $200B Lesson in Patience
In 2019, SoftBank sold its entire 4.9% stake in Nvidia for $6B.
That same stake today would be worth $200B.
✍️ Sequoia and the silent playbook
Sequoia’s go-to crisis move?
Say nothing and wait.
After partner Shaun Maguire’s July 4 tweet calling a NYC candidate an “Islamist” from “a culture that lies,” backlash followed — 5M+ views, 1K+ petition signatures.
What did Sequoia do?
Nothing. No comment. No apology.
Why?
🖱 Maguire is tight with Stripe’s Patrick Collison
🖱 Helped close a $20M Stripe deal at GV
🖱 Led Sequoia’s $1.1B Bridge deal (acquired by Stripe)
🖱 Rumored link to Elon Musk
🖱 Sequoia has $500M+ in Stripe
🎥Maguire apologized, then doubled down:
→ Threatened critics
→ Claimed it was just “1% power”
🧠 AI and critical thinking: most people don’t read past the headline
A viral article claimed devs using AI felt 20% more productive — but actually saw a 20% drop.
Everyone reposted.
Almost no one checked the details.
The study used Claude 3.5, while real coding today happens on Sonnet 4 or Opus.
That’s like benchmarking an iPhone 6 and judging the iPhone 15.
❕The real edge isn’t prompting — it’s thinking
MIPT-style logic applies here: when the program is too big to cover, extract what matters and deliver.
That’s what AI work requires now — clarity, structure, and prioritization.
Any brain — human or LLM — has limits.
If the task fits, things work. If not — hallucinations.
The fix: break it down. Systematize. Apply judgment.
♨️ In practice? Structured thinking wins
One of our team’s current builds is ~500K lines of code.
Mostly JavaScript, written entirely via Claude Code.
The model slips? It’s caught, corrected, refined.
No whining. Just progress.
Total AI spend: $200/month.
Result: clean UX, working product, real value.
⚙️ It’s not about the tools — it’s about how you use them
People want to write one prompt:
Make me a startup, build it, market it, send dividends.
When a VC decides to invest in a startup...
Then Google, Anthropic, or OpenAI drops a new model that instantly wipes it out.