THE BOMBSHELL OF THE WEEK
Parag Agrawal, the CEO fired by Musk in the Twitter acquisition (2022), disappeared from the radar for 3 years. Now he comes out of stealth mode on November 12, 2025 with explosive news:
He has raised a $100 million Series A ($740M post-money valuation) for Parallel Web Systems, a startup that has been operating in absolute secrecy for two years.
Agrawal isn’t building another AI chatbot. He’s rebuilding a fundamental piece of the internet: the web search layer, but designed specifically for AI agents, not humans.
While everyone competes to build better language models, Agrawal has identified the real problem: AI agents need to access web information in a completely different way than we humans do.
He worked in total secrecy for 2 years (he founded the company in 2023 and launched products in August 2025). Top-tier investors (Kleiner Perkins, Index Ventures, Khosla, FirstRound) came in without making noise. The timing is no coincidence: he announced it the same day Cursor raised $2.3B.
He already has enterprise customers using Parallel for AI agents that write code, analyze sales data and assess insurance risk. This isn’t vaporware: it has real, verified traction.
POWER MOVES
Cursor: from $9.9B to $29.3B in 5 months. The wildest round of 2025
November 13: Announcement of $2.3B in investment (round type unspecified)
- Current valuation: $29.3B (tripled its value since June: $9.9B)
- Annual revenue: $1B with just 300 employees
- Revenue per employee: $3.3 million (vs the sector average of $1.2M)
- Strategic investors: Nvidia, Google (on top of traditional venture funds)
Cursor raised $60M in March 2024 at a $400 million valuation. Going from $400M to $29.3 billion in 18 months means multiplying its value by 73. This isn’t normal growth — it’s an unprecedented explosion.
Having $1 billion in revenue with just 300 people means each employee generates more money than at any other software company in history — if we don’t count OnlyFans as a software company :)
Walmart becomes a tech company (and nobody saw it coming)
- Outgoing CEO: Doug McMillon (announced his retirement on November 14, a total surprise at 59)
- Successor: John Furner (current CEO of Walmart US)
- Effective date: February 1, 2026
- Secret alliance: OpenAI (not made public until September 2025)
McMillon declared that “AI will change every single one of Walmart’s more than 2 million jobs.” Furner takes over with a crystal-clear mandate: massively deploy the AI, robotics and automation McMillon had been developing in secret.
Walmart is keeping its workforce stable at 2 million employees despite growing. The automation is already working.
Walmart has a hidden advantage no digital company can replicate: more than 10,500 physical stores. That means real human behavioral data no purely online business can get.
While Amazon knows what you buy online, Walmart knows how you move, what you look at, what you touch, how long you take to decide. That data feeding AI will be decisive.
ChatGPT ships 5.1 + an App Store in 48 hours. OpenAI floors it
- November 12: Launch of ChatGPT 5.1 (warmer tone, configurable personalities, adjustable speeds)
- November 13: Announcement of the ChatGPT Apps Platform (a full App Store inside ChatGPT)
- Strategic timing: the same day Cursor announced its $2.3 billion
ChatGPT 5.1 isn’t simply another version. It introduces personality settings (the assistant can be “friendly,” “professional” or “concise”) and speed controls (slow mode is more thoughtful, fast mode is more agile). This shows OpenAI finally understood that a single version doesn’t work for more than 300 million users.
MONEY TALKS
The wildest investment week of 2025: $4.2B in 7 days
Total raised November 8-15: $4.326 billion (2.1x the weekly average of Q4 2024).
Here’s the breakdown:
1. Cursor: $2.3B at a $29.3B valuation (Nov 13) — AI for coding
2. CHAOS Industries: $510M Series D (Nov 13) — Military tech
3. d-Matrix: $275M at a $2B valuation (Nov 12) — AI chips
4. Abridge: $250M Series D (Nov 12) — Medical AI
5. Rebellions: $250M Series C (Nov 11) — AI chips (Korea)
6. Wonderful AI: $100M at $700M (Nov 11) — Multilingual AI agents
7. Parallel (Agrawal): $100M at $740M (Nov 12) — AI web infrastructure
8. Hearth: $100M Series B (Nov 8) — Contractor financing
9. Salesforce/Doti: $100M acquisition (Nov 13) — Enterprise search
The most important part:
Military tech is exploding: CHAOS Industries raised $785M in 4 months. Its product destroys 5,600 Russian drones a month in Ukraine. Investor: Andreessen Horowitz (the same fund behind Cursor).
Silicon Valley is literally deciding wars.
Wonderful AI record: from $34M (July) to $100M (November) = a 20x in 4 months. They already operate in 15 countries. That speed only happens when the product genuinely works.
Salesforce in panic mode: eighth acquisition in 2025. When you buy 8 companies in a year, it’s because you don’t have time to develop your own AI. OpenAI, Anthropic and Google are eating its market.
Buying for $100M is buying teams, not technology.
VALLEY DRAMA
Blue Origin: a quiet executive exodus
Reported: first half of November
- Pattern: senior executive turnover up 40% vs 2024
- Unofficial reason: Bezos’s micromanagement + a competition against SpaceX that’s impossible to win
Blue Origin has spent years “on the verge of” competing with SpaceX. Every year the gap gets wider. Executives leaving in droves is a sign that internally they know they’ve lost. Bezos has infinite money, but he can’t buy Musk’s execution culture.
MY TAKE
This week is packed, but here we go…
I believe the answer is YES.
AI agents are going to become the main consumers of web content, and Google Search is going to become obsolete. Parallel’s system returns “tokens” optimized for language models to process directly, instead of lists of links designed for humans.
For my own projects I already use Firecrawl, which takes a similar approach, and the difference is brutal — although Firecrawl RELIES on today’s search engines. What Parag is doing is different. It’s infrastructure. Big fan of this project.
Nvidia and Google investing in Cursor is no accident.
They’re betting that the future of programming will be 100% AI-assisted, and they believe Cursor is the one that will dominate.
For Nvidia, Cursor means more demand for its GPUs — the more developers use AI to code, the more of those chips are needed.
For Google, it’s life insurance: if Cursor replaces tools like Android Studio, Google won’t want to be left out of the business.
Still… careful… raising three massive rounds in less than 2 years can mean one of two things: either they’re burning money at a blistering pace, or they’re getting ready.
With a billion in annual revenue, they don’t need money to survive. They need it to block competitors and buy companies before anyone else does.
Winner takes all.
The Walmart story surprises me. I’m seeing traditional retail companies transform into tech companies faster than tech companies are becoming retail companies.
Amazon should be worried.
The ChatGPT news — I won’t even comment. It’s too early, and Sam shoots at anything that moves.
The chip war I covered last week.
Inference is not the same as training, and inference is where the usage is. If those chips are genuinely cheaper, NVIDIA will suffer… and it’s worth 5 Trillion (I’ll just leave that there).
Another week done. I hope my newsletter gives you what it gives me: being properly informed so you can make your BEST decisions.
Thanks for reading.