🚫Tesla shuts down Dojo supercomputer project

Tesla has officially pulled the plug on its Dojo AI supercomputer, a project once described by Elon Musk as the key to achieving full self-driving.

After years of development, the team is being disbanded and its leader, Peter Bannon, is exiting the company.

🖱 Musk had positioned Dojo as Tesla’s in-house solution to train video-based AI models at scale
🖱 The move follows the exit of ~20 employees who formed a new AI chip startup, DensityAI
🖱 Tesla will now rely more heavily on external partners like Nvidia, AMD, and Samsung
🖱 A $16.5B chip production deal with Samsung will help power both FSD and Tesla’s Optimus robot

Dojo was once expected to add $500B in value, according to Morgan Stanley. Now it’s just another pivot in Tesla’s AI story.

✔️Powered by Trade Watcher

🚀Firefly Aerospace takes off with 34% IPO surge

Spacetech startup Firefly Aerospace jumped 34% on its Nasdaq debut, closing its first day of trading under the ticker $FLY with strong investor interest despite ongoing losses.

🖱 The company raised $868M, pricing at $45/share — above its raised range
🖱 Q1 2025 revenue hit $55.9M, up from $8.3M YoY, but net loss still at $60M
🖱 Firefly offers launch, land, and in-orbit services for national security & commercial clients
🖱 Largest pre-IPO shareholder: AE Industrial Partners (47%)
🖱 The IPO comes amid a strong run for spacetech, with multiple $100M+ raises this year
🖱 Investors are especially excited about spacetech x AI crossovers

Firefly joins a wave of hot 2025 debuts including CoreWeave, Circle, and Figma — another sign that the IPO window is wide open again.

✔️Powered by Trade Watcher

🚀 Rocket Lab goes deeper into defense with optical systems play

Rocket Lab is pivoting harder into defense, using its latest acquisition to move beyond launch and strengthen its position in high-value government contracts.

In Q2, 68% of its record $144.5M revenue came from its space systems segment — not rockets.

🖱 Rocket Lab is acquiring Geost, an optical payload company, for $275M
🖱 This unlocks a new Optical Systems division focused on EO/IR sensors for missile tracking and space surveillance
🖱 The move targets large DoD contracts like the multi-billion dollar Golden Dome initiative
🖱 Already secured: an $515M contract to build 18 missile-tracking satellites for the Space Development Agency
🖱 First Neutron rocket launch still on track for end of 2025, with its Virginia complex nearing completion

With defense spending accelerating, Rocket Lab is positioning itself not just as a launcher — but as a key tech supplier for space-based military systems.

✔️Powered by Trade Watcher

💡Rethinking MVP: From Minimum Viable to Most Valuable Product

It’s time to rewrite one of startup culture’s core ideas.

The old MVP mindset tricks founders into thinking “minimum” means cheap or free.
But that logic is flawed and often fatal.

🖱 Most teams build the core 20% that solves the real user pain, label it MVP, and give it away
🖱 Then they waste years adding 80% of features nobody asked for, just to justify charging
🖱 The paradox: what users would pay for is given for free, while useless extras are monetized

⚡️ The shift: MVP = Most Valuable Product.

Your MVP shouldn’t be the smallest thing you can build. It should be the highest-leverage slice of your product—something that:

🖱 Solves the key pain point
🖱 Creates a clear “aha” moment
🖱 Justifies payment from Day 1

👇 How to offer value without giving it away:

🖱 Give 3 free uses, then charge
🖱 Offer a 7-day trial with full features
🖱 Use credits (e.g., $50 worth) for testing

If your MVP truly solves a problem, people will pay after trying it. If not—they won’t. That’s your signal to rethink.

Example:

Skip the limited free tier. Instead, build a premium tool that does one job incredibly well. Let people try it 3 times. Then charge.

Add the rest only once you’ve validated demand through real feedback and real payments.

✔️Powered by Trade Watcher

📊 Crypto Venture Weekly: August 4–8, 2025

$207M raised across 24 projects this week, with capital flowing into AI infra, gaming, Bitcoin L2s, and DeFi tooling.

Here’s what the top 9 are building 👇

🖱 OpenMind ($20M, Pantera, Coinbase, DCG)
Open-source OS for AI agents and robots, plus FABRIC — a decentralized network for coordination and trust.

🖱 SuperGaming ($15M, Skycatcher, Steadview, a16z crypto)
India-based game studio behind Indus and MaskGun, building Web3-enabled multiplayer titles.

🖱 BOB ($9.5M, CMS, IOSG, Amber)
Hybrid L2 for building Bitcoin-native tokens and applications.

🖱 Perle ($9M, Framework Ventures)
Modular AI training data solutions for large-scale model development.

🖱 Euphoria ($5M, Karatage, Robot Ventures, Bankless)
MegaETH-based mobile app simplifying crypto derivatives trading.

🖱 BlindPay ($3.3M, Y Combinator, 468 Capital)
Stablecoin payment APIs for compliant, instant global transfers.

🖱 Levr.Bet ($3M, Blockchain Capital, Maven 11)
Decentralized leveraged sports betting exchange.

🖱 Doppler Finance ($3M, Reforge, DCG, Hashkey)
DeFi protocol on XRP Ledger for on-chain trading and lending.

🖱 Hylo ($1.5M, Robot Ventures, Solana Ventures)
Solana-based DeFi platform offering LST-backed stablecoin hyUSD and leveraged SOL exposure.

Investor focus this week leaned toward AI-driven infrastructure, Bitcoin scalability, and DeFi payment rails.

✔️Powered by Trade Watcher

🚁 Harmattan aims to be the “Anduril of Europe” with $200M raise

Paris-based Harmattan AI is seeking $200M in early-stage funding to scale its AI-powered military drones. The startup, launched in April 2024, has already secured $30M from FirstMark and Atlantic Labs, per Bloomberg.

🖱 Building three drone types: training, surveillance, and weapon-interception systems.
🖱 Plans to mass-produce 10k drones/month in France by 2026, with facilities in Germany and the US.
🖱 Signed a multi-million-dollar NATO deal for small AI drones, delivery set for October.
🖱 French defence ministry ordered 1,000 drones for delivery by end of 2025.

With European defence tech funding up 26% YoY, Harmattan is chasing long-term government “programmes of record” and aims for its first $1B contract by 2030.

✔️Powered by Trade Watcher

🦄 June hits 3-year high for new unicorns

Twenty companies joined the Crunchbase Unicorn Board in June - the biggest monthly jump since July 2022.

AI and robotics led the way, with Thinking Machines Lab topping the list after a $2B seed round at a $12B valuation.

🖱 The US led with 11 new unicorns, followed by China (4), and one each in Israel, India, UAE, Switzerland, and New Zealand.

🖱 Six unicorns went public, including neobank Chime and Circle, while two were acquired.

🖱 Notable sectors: AI (Thinking Machines Lab, Decagon, Fireflies.ai), robotics (Unitree, Gecko Robotics), and Web3 (Zama, The Open Platform).

The surge shows capital is flowing again into high-growth tech, with AI continuing to dominate, but robotics, defense, and niche SaaS also getting their share of billion-dollar valuations.

✔️Powered by Trade Watcher

💰 Inside the Windsurf–Google deal fallout

Weeks after it emerged that Google paid $2.4B to license Windsurf’s tech and hire its CEO plus 40 key staff, more details reveal why the deal stirred controversy in Silicon Valley.

🖱 Half of Google’s payment ($1.2B) went to investors, including Greenoaks, Kleiner Perkins, and General Catalyst — Greenoaks turned $65M into ~$500M.

🖱 The other half was compensation for the 40 hires, with a large share going to co-founders Varun Mohan and Douglas Chen.

🖱 ~200 remaining employees saw no payout despite hopes from an earlier $3B OpenAI acquisition attempt.

🖱 Investors and founders left Windsurf with $100M+ in capital, but opinions differ on whether it could’ve been used to pay all employees without sinking the company.

🖱 Some Google hires had stock grants revoked and vesting restarted, delaying full payouts by 4 years.

🖱 Vinod Khosla called the move “a bad example of founders leaving their teams behind.”

🖱 Eventually, Cognition acquired Windsurf’s remaining entity and staff for an estimated $250M, ensuring those employees got a financial benefit.

A lucrative win for VCs and founders — but a cautionary tale for startup teams betting on a big exit.

✔️Powered by Trade Watcher

🚀 What Startups Y Combinator Wants — Fall 2025

Y Combinator has updated its Requests for Startups, highlighting the areas it’s most eager to back. Almost all of them are now AI-focused. Here’s a condensed look at the fresh list.

🖱 Retraining Workers for the AI Economy – AI tools (possibly with AR/VR) for quickly training electricians, technicians, welders, and other skilled workers to meet infrastructure needs.

🖱 Video Generation as a Primitive – Building new applications and infrastructure for low-latency, unlimited AI-generated video as a core creative tool.
• The First 10-person, $100B Company – Leveraging AI agents to create ultra-efficient teams with unprecedented revenue per employee.

🖱 Infrastructure for Multi-Agent Systems – Developer tools for building, scaling, and maintaining distributed AI-agent architectures.

🖱 AI-Native Enterprise Software – Next-gen enterprise platforms deeply integrated with AI, replacing traditional “systems of record.”

🖱 LLMs Instead of Government Consulting – AI services that replace expensive government consulting contracts from firms like Deloitte and Accenture.

YC’s message is clear: if you’re building for AI-first markets with high leverage and big vision, they’re listening.

✔️Powered by Trade Watcher

👆 Round sizes for software startups over the past 12 months.

✔️Powered by Trade Watcher

🖥 Perplexity makes $34.5B bid to buy Chrome from Google

AI search startup Perplexity has made an unsolicited $34.5B cash offer to acquire Chrome, pledging to keep Chromium open source, invest $3B into it, and maintain Google as the default search engine.

🖱 Offer follows DOJ’s proposal that Google be forced to sell Chrome after monopoly ruling
🖱 Chrome holds ~68% global market share, making it the dominant browser
🖱 Perplexity’s bid exceeds its own $18B valuation and $1.5B total funding
🖱 DOJ could set divestment terms as early as this month, attracting multiple bidders

If approved, the move would give Perplexity one of the most valuable browser assets in the world - without touching its search defaults.

✔️Powered by Trade Watcher

🚀 Risk as the gateway to high returns

History shows that the most rewarding deals are rarely the safest ones. True success often comes from taking calculated risks in opportunities with strong growth potential. While most investors stick to the comfortable path, those willing to step into uncertainty often reap outsized rewards.

Examples:

🖱IPOs can be volatile, but some deliver rapid growth — Twilio’s stock more than doubled shortly after going public.

🖱 Cryptocurrencies are highly risky, yet early Bitcoin backers saw life-changing gains.

🖱 Venture capital comes with a high failure rate, but a winning startup can return multiples on the initial investment.

In investing, discomfort often signals the possibility of extraordinary upside.

✔️Powered by Trade Watcher

📊 Which VCs got in early on today’s $5B+ startups

Crunchbase ranked funds by how often they entered Series A or B rounds of companies now valued at $5B or more.

🖱 Accel leads with 12 Series A and 13 Series B bets

🖱 Index Ventures, IDG Capital, Lightspeed, and General Catalyst follow closely

🖱 Data excludes cases where the company was already a unicorn at its first round — rare, but common in AI

A clean snapshot of who’s spotting mega-winners before they blow up.

✔️Powered by Trade Watcher

📱 Apple is not having its BlackBerry moment

Business Insider claims Apple is missing the AI wave just like BlackBerry missed touchscreens. The comparison doesn’t hold up, BlackBerry lost the platform war, not the hardware war.

🖱 BlackBerry’s real killer was the iOS/Android app ecosystem - endless apps for users, endless users for devs.

🖱 Touchscreen vs. buttons was secondary. Big companies can change form factors — ecosystems are harder.

🖱 For Apple to be at risk, devs would need to skip mobile apps entirely and build straight for GPT or other AI-native platforms.

🖱 We’re far from that, and Apple could acquire an AI leader long before it becomes a threat.

Until AI platforms replace the App Store as the go-to launchpad, Apple’s not in danger of a BlackBerry-style collapse.

✔️Powered by Trade Watcher

🔎 Due Diligence hacks they don’t teach in VC handbooks

You can spin a perfect pitch deck with AI, but you can’t spin the energy of a team, the loyalty of employees, or the founder’s true mindset.

Early-stage due diligence is shifting away from just reading metrics and projections. The most revealing insights often come from creative, off-the-record moves that catch founders in real life, not in presentation mode.

Here’s how the savvier VCs do it 👇

🖱 Talk to ex-employees - the ghosts of startups past. Reach out on LinkedIn or via mutuals to ask why they left. Sometimes the tone says more than the words. Silence can be the loudest answer of all.

🖱 Chat with office neighbors. Other tenants notice patterns like when lights go off and laptops close. If that’s consistently 5 p.m., it raises questions about commitment and pace.

🖱 Probe by offering to buy shares from insiders. If early investors, founders, or key staff are quick to sell (even at a discount), it’s often a sign they’ve lost conviction in the company’s future.

🖱 Test loyalty by trying to poach a key hire. A friendly “job offer” via a portfolio company can reveal whether a star engineer is rooted in the mission or already halfway out the door.

🖱 Schedule a weekend meeting. Founders in the heat of building usually make it happen. A flat “no” without context might hint at a mismatch between words and hustle.

🖱 Cold-call customers, not the ones they hand-pick. Ask bluntly, “What would you change about the product?” The honest answers tell you the real product-market fit.

🖱 Audit personal social media. Instagram, X, or even TikTok can reveal more than polished LinkedIn profiles. Lavish trips during a cash crunch? Public industry rants? It’s all signal.

🖱 Work from their office for a day. The atmosphere doesn’t lie. Is there energy, urgency, and collaboration — or silence, clutter, and drift?

Spreadsheets show the story founders want to tell.
The vibe shows the story you actually need to know.

✔️Powered by Trade Watcher

📊 How co-founders split equity in startups

Only 1 in 10 startups with four founders split equity evenly. According to Carta, whether shares are split equally or not has no measurable impact on a startup’s success.

The time until hiring the first employee has also grown - now often reaching up to three years. And in recent years, the number of solo founders has been steadily rising.

✔️Powered by Trade Watcher