Pinned Tweet
UK private equity target universe is ~40% London and South East but the mix is increasingly shifting to regional deals.
Five AI value creation opportunities now move the value of a business.
Dynamic pricing. AI that prices each deal based on demand, competitor moves, inventory, and customer segment. Was the preserve of airlines and Amazon; now economic for any business with a price list. (revenue, margin)
AI sales development. Software that researches prospects at scale, drafts personalised outreach for each one, qualifies the replies, and books meetings. Top of funnel widens, conversion at each stage improves, sales reps spend their time with buyers rather than spreadsheets. (revenue)
AI in the product, not just in operations. Software, services, healthcare, and intelligent equipment companies can now embed AI into what the customer pays for, building a data and experience moat competitors cannot copy. (exit multiple)
Company knowledge base. Plain-English questions answered from across finance, sales, ops, and customer systems. Top management gets visibility into the bottom of the organisation; every employee sees beyond their own function. (productivity)
Autonomous service agents. AI that holds a real conversation — reads tone, asks the right follow-up, holds context across turns — and resolves end-to-end across systems. The team handles only the complex cases. (margin)
Execution speed depends entirely on the data the business already has.
OpenAI at $852B valuation is cheap.
$25B ARR today. Growing 3x YoY.
900M weekly active users.
Google monetises at ~$60/user/yr from ads.
OpenAI is at ~$3.50.
Conservatively:
→ 1.5B WAUs by 2030
→ Half Google's ad ARPU (~$30/user)
→ That's $45B in ad revenue alone
Now add:
→ Enterprise (already 40% of revenue)
→ API (15B tokens/min and growing)
→ Codex (1.6M weekly users, tripled since Jan)
Subscriptions + enterprise + API already at $25B and compounding. Stack $45B+ in ads on top.
$852B is cheap.
Anu Chandra retweeted
Can we trust AI agents with critical enterprise tasks? Absolutely not.
Introducing Wow (World of Workflows), the first Agentic Safety benchmark that proves that frontier LLMs fail miserably under safety constraints at enterprise tasks.
🧵 WoW demonstrates that LLM agents are “dynamically blind”. They fail to track the downstream ripple effects of their actions against complex enterprise rule sets. In an enterprise, that’s a safety and compliance hazard.
Our research shows how the future of enterprise AI requires proactive agent architectures and Wow is just a starting point.
📌 It’s now available to all researchers at: github.com/Skyfall-Research/…
Full blog here: skyfall.ai/blog/wow-bridging…
There are only two viable strategies: cost leadership or differentiation.
Try to do both, and you end up stuck in the middle — outcompeted on both fronts. (Michael Porter)
Reports of a new low-cost MacBook ($599–$699) suggest Apple might be drifting the wrong way.
Apple’s strength has always been premium differentiation — it should double down on AI, not price wars.
Anu Chandra retweeted
This is insane 🔥
Alibaba just dropped a 30B-parameter AI agent that outperforms GPT-4o and DeepSeek-V3 using only 3.3B active parameters.
It’s called Tongyi DeepResearch and it’s completely open-source.
Instead of brute-forcing scale, they changed how models learn to think.
They introduced a new phase called agentic mid-training —
where the model learns how to behave like an agent before task fine-tuning.
Pre-training → learn language
Mid-training → learn reasoning
Post-training → learn tasks
The dataset is entirely synthetic AI-generated questions, reasoning chains, and tool-use sequences.
20% of samples exceed 32K tokens with 10+ tool calls. That’s superhuman complexity.
Results:
• 32.9% on Humanity’s Last Exam (↑ from 26.6% GPT-4o DeepResearch)
• 43.4% on BrowseComp (↑ from 30.0% DeepSeek-V3.1)
• 90.6% on FRAMES best so far
All trained on 2 H100s for <$500.
While others chase 600B+ models, Alibaba just proved smart training beats scale.
The future isn’t bigger models it’s agentic ones that can reason, search, and synthesize autonomously.
📄 Paper: arxiv. org/abs/2510.24701
💻 Code: github. com/Alibaba-NLP/DeepResearch
Anu Chandra retweeted
I got rejected by 144 investors before raising $150M for my $200M+ rev/year startup.
After 144 rejections, I started questioning our approach.
Were we solving the right problem?
What were we doing wrong?
Why weren’t investors seeing what we were seeing?
Were we the right team to build this?
We tried everything: different pitch angles, new deck structures, and reframing the problem.
Then came the 145th meeting, where we closed our first growth round.
That yes made everything worth it. But getting there took years of mistakes and hard work.
We went through a lot of trial and error just to figure out what resonates with investors.
We tried dozens of approaches to figure out what made investors engage.
Some landed, most didn't. But each iteration taught us something about what builds conviction versus what just sounds good on paper.
And once we cracked that code, our Series C closed faster than expected.
And today, I see so many founders in the exact same position I was in 10 years ago: grinding through rejections, questioning everything, and trying to figure out what works.
So today I want to give you the resource I wish I had back then:
Something that shows you exactly how to structure these conversations and navigate the entire process
(because the fundraising cycle can be a big distraction and take a toll on you as a founder).
So I've partnered with Notion's Startups Team to create the essential fundraising resource that helps you avoid the mistakes that cost me years.
Here's what you are getting:
• The actual decks I used to raise $150M for Super[.]com (Series B, C)
• 50 real examples from funded startups like Eleven Labs and Artisan AI
• A searchable database of 10,000+ investors - angels, VCs, and accelerators you can reach out to immediately (this alone would take months to build manually)
• An AI-powered fundraising agent built into Notion with step-by-step prompts (no separate ChatGPT needed)
Want access?
• Like and share this post
• Comment "FUNDRAISE"
• Follow me so I can DM you the link
I'll send it over ASAP.
P.S.: If you are serious about fundraising (now or in the future), you should grab it right away.
Steve Jobs predicted large language models
youtube.com/watch?v=iTsbHQ-M…
Anu Chandra retweeted
The world's healthiest countries:
1. Spain
2. Italy
3. Iceland
4. Japan
5. Switzerland
...
19. United Kingdom
...
35. United States
bloom.bg/2EeIcLx
Anu Chandra retweeted
1,000,000 acres of trees were burned during last year's California wildfires, and this drone startup is dropping seeds to replace them
Anu Chandra retweeted
In the UK? Use our template letter to write to your local MP today to demand an answer on Assange! defend.wikileaks.org/2019/01…
"World's Simplest Animal Reveals Hidden Diversity" scientificamerican.com/artic…
"Protein for deeper insights into the brain" phys.org/news/2018-09-protei…
"Researchers to Release Genetically Engineered Mosquitoes in Africa for First Time" statnews.com/2018/09/05/rele…
"An Acceleration Breakthrough Could Fundamentally Change How We Study Particles" futurism.com/particle-accele…
"Weird circles in the sky may be signs of a universe before ours" newscientist.com/article/217…
"Lifting the lid on the unconscious" newscientist.com/article/mg2…
"Scientists Use CRISPR to Turn E.Coli Bacteria into Microscopic Tape Recorder" futurism.com/scientists-use-…










