Models know the internet. @EngramLab believes they should know you.
Congrats to @dan_biderman and the team. few have worked on continual learning longer.
We are quadrupling down on @AnthropicAI. This is the largest single investment in @generalcatalyst history.
We first invested four rounds ago (March '25 at $61.5b) because of the tech and the soul. It's been a privilege to watch Anthropic grow into one of the world's most important companies.
Long Lake, one of our very first AI roll ups, has agreed to acquire Amex GBT for $6.3 billion.
@alextaubman had a vision to take durable franchise businesses with decades of customer trust and make them better with AI. In two years: doubled free cash flow in HOA management businesses without reducing headcount. This is Long Lake's 31st acquisition and their first take-private. Congrats Alex and team
I've been posting less and writing more. I've found it helpful to slow everything down and take the time for proper reflection.
Today I'm putting out a review of the last quarter:
Growing AI Anxiety
Turning LLMs into Computers
Anthropic and the Department of War
Iranian Conflict
The Incoming Techlash
It's easy to feel like the response to this is for tech to "get tough"
But I want to encourage the next generation to do something simple: keep your soul
Thrilled to welcome @nikesharora as @generalcatalyst's first lead independent director. Nikesh brings the rare combination of investor thinking and operator execution and has led some of the most consequential tech orgs in the world. I've long admired the clarity and ambition he brings to everything he does. Nikesh joins me and our chairman @ChenaultKen on the board.
Look forward to our work together building GC and supporting the founders we back.
Awesome. 680% speed improvement. 97% cost reduction.
As the hardware/software loop condenses, intelligence is folding back on the physical substrate of its own existence, optimizing the lithography that prints the transistors that run the models that optimize the lithography. We're seeing this in energy too. The end state will be a full-stack intelligence loop, running all the way from raw materials to apps.
Over the past few months, we have integrated @googledeepmind's AlphaEvolve into our computational lithography. Enabled by AlphaEvolve's algorithmic leaps, we are now printing complex patterns in a single exposure that would otherwise require multiple. https://t.co/0ujvA4ALWz
Second mover advantage has always been powerful. AltaVista, Google. Myspace, Facebook. AppDynamics, Datadog. OpenAI, Anthropic. We're seeing this at GC with @WeAreLegora in legal and @parloa_ai in CX.
Every tech era has a different source of 2nd mover advantage. internet era = distribution and reaching users at scale. cloud era = design and building workflows that got people to use tech.
Today we are in the AI era. Intelligence is an ever improving input. Do you have the right models, infra, and capability to build products that deliver what AI unlocks for customers?
founders shouldn’t be afraid if someone is currently the leader. Most industries have not yet seen the dominant stack.
@ChristosTzamos Wait this is so awesome!! Both 1) the C compiler to LLM weights and 2) the logarithmic complexity hard-max attention and its potential generalizations. Inspiring!
This is not politics as usual. As investors, innovators, leaders, and citizens, we need to come together to preserve our democracy, the rights set forth in our country's Constitution, and our shared humanity. All of these serve as the foundation for our modern society, human progress, and innovation.
What we are seeing in Minnesota is a threat to those core tenets and to the promise of America. Without confidence in the fundamentals -- human rights, the legal process, and the privilege and responsibility of our government and its leaders to protect all people -- the future is at stake. That impacts all of us here in the U.S., and globally.
I’m incredibly energized to announce that I’ve joined @generalcatalyst as their Creative Director
It’s an honor to be trusted to hold this position by @htaneja, @jcfurstenberg, and the entire GC Famiglia
I had the opportunity to spend a lot of time with incredible folks across our industry during my break. But the ambition that General Catalyst has to tackle huge problems, is simply unnatural.
There’s so much work to do to show the story of the immense world being built here. But I couldn’t imagine doing anything else right now.
It would be absolutely foolish to not thank the many people that spent time supporting me during this time. So here’s a list of thank yous: @gaby_goldberg@seanxthielen@menemazarakis@thecpe@tamarawinter@marannelson@pythianism@janehk@cokiehasiotis@jacksondahl@nabeelqu@EmilyHerrera@DevinLewtan@bryce@soleio and I’m sure there’s others I’m forgetting who had a huge impact on me during this time
And most importantly, my very offline Wife whom I’m nothing without 💕
*bonus first day of work fit yesterday
A Decade of Reflection on Meesho
It’s hard to believe it’s been nearly 10 years since we, at Venture Highway, signed that first seed check for @Meesho_Official in 2016. With the news of their upcoming IPO, I’ve been looking back at the email threads, the early product demos, and the "aha" moments that started it all.
What that decade has really reinforced for me is that Meesho’s journey was anything but linear: the company evolved through multiple product iterations—from the early Fashnear experiment in late 2015, to reseller-led commerce, and ultimately to the direct-to-consumer marketplace Meesho is today—while the core insight and our conviction in the founders stayed constant.
For me, the conviction to partner with @viditaatrey and @barnwalSanjeev wasn't just about e-commerce. It was about a behavior I had witnessed firsthand while building WhatsApp.
The "WhatsApp" Insight
During my time at WhatsApp, I saw something fascinating happening in India. Small business owners—boutique owners, home chefs, local artisans—were "hacking" WhatsApp to run their businesses. They were creating groups, sharing photos of inventory, and closing deals via chat. There was no shopping cart, no payment gateway, just trust and conversation.
When I met Vidit and Sanjeev, they weren't trying to force a Western e-commerce model onto India. They had noticed the same behavior. They understood that for the next 500 million Indians, commerce wouldn't look like a catalog; it would look like a chat.
The "TAM" Trap & The Valley of Death
But the path wasn't a straight line. After Y Combinator, Meesho hit a wall. They struggled to raise capital. Investors were passing left and right.
I distinctly remember a call from Ravi at Elevation Capital. He asked me, "Neeraj, why is no one funding this company? These are good guys, they are executing, and they are making progress. What am I missing?"
I told him the truth: "Investors are getting stuck on TAM (Total Addressable Market). They are looking at this through a Western lens where transaction values need to be high to make sense. But that is the wrong way to look at this deal. This isn't about basket size; it's about frequency and behavior."
Most people saw a small market; we saw a massive, unorganized economy waiting to be digitized.
These were also the moments when founder conviction mattered most. Even as the product evolved, Vidit and Sanjeev stayed anchored to their core user insight, and our job as early investors was to lean in through that ambiguity rather than step back and “wait for the final form” of the company.
Shared Roots
Vidit and Sanjeev being from IIT Delhi (where I also studied) gave us a shared starting point, but what really built our conviction was their resilience during those tough months. They were technical and data-driven, yet remarkably humble and willing to pivot. They were building for India, not just in India.
The Thesis that Built Venture Highway
Investing in Meesho in 2016—and doubling down in subsequent rounds—was more than just a portfolio decision. In many ways, Meesho was the thesis that led me to scale Venture Highway.
Back then, the narrative was often that Indian startups were copycats. Meesho proved that India could produce category-creating, world-class companies that solve uniquely Indian problems at scale. That conviction drove us to build Venture Highway to support founders capable of building enduring institutions, a mission the entire VH team has worked hard to support.
Going Global with General Catalyst
It is fitting that as Meesho steps onto the public stage, Venture Highway has also evolved. Our conviction in Indian founders is what led us to join forces with General Catalyst. What drew us to GC was how closely their seed philosophy mirrors what we lived through with Meesho: seeing seed as the unveiling of a hidden truth, backing exceptional people early, and leaning in when the path is still murky. Meesho’s evolution—from its Fashnear beginnings to the scaled direct-to-consumer marketplace it is today—is exactly the kind of non-linear founder journey that GC is set up to support: where others see risk in change, GC sees learning, resilience, and conviction. By merging our local expertise with GC’s global platform, we are now even better equipped to support the next generation of Vidits and Sanjeevs—helping them build world-class companies from India for the world.
To the IPO and Beyond
Watching this journey from a seed-stage pitch deck to a public listing has been one of the greatest privileges of my investing career. We have been there since the seed round and are proud to remain shareholders today, as the company enters this new chapter.
Vidit and Sanjeev, you have validated our belief that India is a cradle for world-class innovation. You’ve also reminded us that the most enduring companies are rarely built in a straight line - and that the real test of partnership is in those lows (and not the highs). Here’s to the next decade of building.
General Catalyst ran a screener of 70 industries for their AI Roll-Up strategy.. they only selected 10.
That discipline is paying off:
even 20% revenue growth with a flat cost base + 20–30% task automation turns a 10–15% EBITDA business into ~30% margins.. doubling the FCF.
Marc Bhargava (@marcbhargava) , Managing Director at General Catalyst explains:
"We're already seeing in many of the case studies that if you look at the first couple companies they acquired, even if they've only grown revenue 20%, if they've kept the cost basis flat using AI, but then bring up people to kind of do 20% more tasks, around 10-15% EBITDA-margin businesses can become 30% EBITDA-margin businesses. You can roughly double the free cash flow, and we're seeing that happen in many of our businesses in a one-year period.
So for us, at the heart of that is: is the AI automation enough?
And obviously we decided not to go forward with 60 of the 70, so we held a really high bar to: is this an industry where AI automation actually can make a meaningful difference?
Then, can we actually go and buy and roll up these companies, get their data, improve the models, improve our software, automate 20-plus percent of tasks, free people up to cross-sell, get more revenue?
And the result is a higher revenue base with a very similar cost basis, which generally translates to a really large improvement in EBITDA margin."
@generalcatalyst
Percepta is hiring for research (RL, modeling, OR), engineering, and product roles in Europe and on the US East Coast. A few of us are at NeurIPS this week. Reach out to me, @EugeneVinitsky, @ChristosTzamos, @justintchiu, or @zzzzgq to learn more!