We invest at the edge of large markets being transformed by technological and societal pressures
New York, NY
- Tweets 2,103
- Following 494
- Followers 136,515
- Likes 4,167
USV retweeted
"To be great at seed investing, you have to put your ideas into the world.
Sharing your thinking signals to founders what you believe in and what kinds of companies you're looking to back.
Content isn't just marketing, it's a way for the best founders to find you before everyone else does." @mignano
Love to hear your thoughts on this @packym @jasonlk @honam @TurnerNovak
I missed out on $100M+ by passing on the seed rounds of Suno and Granola.
Mike @mignano did both. He is the single best example in the last 5 years of an operator successfully turning into a GP at a large fund. It is so hard to do. He crushed it.
He has been a friend for 10 years and we shoot the shit today on the biggest questions of the day.
I summarized my notes from the chat below.
1. Do You Have to Produce Content to Do Seed Well?
To excel at seed, you need to put your ideas into the market so founders know exactly what you are looking for. Publishing your theses acts as a bat signal for early-stage founders before they even launch a product. Success comes down to building a great network, sharing your point of view, and betting on people.
2. Single Biggest Lesson About Venture
The biggest lesson in venture is to avoid projecting your own operator ideas onto a founder. It is their company, not yours. Making investment decisions based on your personal roadmap for them leads to bad bets. Venture success requires trusting the team’s independent judgment and execution plan.
3. What Percent of Enterprise Workflows Actually Require Frontier Models Versus Open Source?
Roughly 80% of non-coding enterprise tasks can run well on non-frontier models. Standard workflows like summarization and document generation are well suited to open source, which is improving quickly. This lets enterprises optimize token budgets without sacrificing meaningful performance.
4. Why We Believe Suno Has Unlimited Upside
Suno has unlimited upside because it unlocks a new consumer behavior: creative entertainment. Users generate music simply for the joy of creation, with no need for distribution or monetization. Like YouTube or TikTok, platforms that democratize a creative medium can create immense, generational value.
5. Obliterate, Don’t Automate
Avoid backing software that merely automates existing workflows to make old businesses slightly faster. The true generational opportunities come from completely reinventing industries, like putting an AI doctor in everyone’s pocket, rather than selling incremental efficiency to a corporate middleman.
6. The Shift From Infrastructure to Applications
As the capital-intensive AI infrastructure buildout matures, massive value creation is shifting to the application layer. Similar to the internet after the broadband boom, we are entering an explosion of specialized software. Winning requires an opinionated thesis, early conviction, and exceptional speed.
7. Why Startups Must Maximize Token Spend
While incumbents restrict budgets, startup CEOs should maximize token spend. Frontier models give lean teams a critical advantage over slower giants. Handing low-level tasks to agents preserves small, hyper-efficient engineering teams and keeps them moving at winning speed.
(links below)
7
7
39
14,750
Frontier models are a luxury good most enterprise workflows can't tell they're buying.
Summarization and drafting don't reward intelligence past a floor. Coding still does. The rest is knowing where that floor is.
"Most enterprise workflows don't require frontier AI models.
Outside of coding, tasks like summarization, document generation, and briefing can often be handled effectively by lower-cost open-source models.
The real value is knowing when you need frontier performance and when cheaper models are more than good enough." @mignano
What percent of enterprise workflows requires frontier models today @nikesharora @matansf @ceo_clickhouse @lqiao
2
3
11
2,249
USV retweeted
"The biggest lesson in venture is simple: don't project your own ideas onto the founder.
Even if you think you know the right strategy, it's the founder's company, not yours.
The best investors back teams because they believe in their judgment and ability to execute, not because they expect them to build the company the way they would." @mignano
What is your single biggest lesson from 10 years investing @kirbyman01 @rabois @rebeccakaden @infoarbitrage
I missed out on $100M+ by passing on the seed rounds of Suno and Granola.
Mike @mignano did both. He is the single best example in the last 5 years of an operator successfully turning into a GP at a large fund. It is so hard to do. He crushed it.
He has been a friend for 10 years and we shoot the shit today on the biggest questions of the day.
I summarized my notes from the chat below.
1. Do You Have to Produce Content to Do Seed Well?
To excel at seed, you need to put your ideas into the market so founders know exactly what you are looking for. Publishing your theses acts as a bat signal for early-stage founders before they even launch a product. Success comes down to building a great network, sharing your point of view, and betting on people.
2. Single Biggest Lesson About Venture
The biggest lesson in venture is to avoid projecting your own operator ideas onto a founder. It is their company, not yours. Making investment decisions based on your personal roadmap for them leads to bad bets. Venture success requires trusting the team’s independent judgment and execution plan.
3. What Percent of Enterprise Workflows Actually Require Frontier Models Versus Open Source?
Roughly 80% of non-coding enterprise tasks can run well on non-frontier models. Standard workflows like summarization and document generation are well suited to open source, which is improving quickly. This lets enterprises optimize token budgets without sacrificing meaningful performance.
4. Why We Believe Suno Has Unlimited Upside
Suno has unlimited upside because it unlocks a new consumer behavior: creative entertainment. Users generate music simply for the joy of creation, with no need for distribution or monetization. Like YouTube or TikTok, platforms that democratize a creative medium can create immense, generational value.
5. Obliterate, Don’t Automate
Avoid backing software that merely automates existing workflows to make old businesses slightly faster. The true generational opportunities come from completely reinventing industries, like putting an AI doctor in everyone’s pocket, rather than selling incremental efficiency to a corporate middleman.
6. The Shift From Infrastructure to Applications
As the capital-intensive AI infrastructure buildout matures, massive value creation is shifting to the application layer. Similar to the internet after the broadband boom, we are entering an explosion of specialized software. Winning requires an opinionated thesis, early conviction, and exceptional speed.
7. Why Startups Must Maximize Token Spend
While incumbents restrict budgets, startup CEOs should maximize token spend. Frontier models give lean teams a critical advantage over slower giants. Handing low-level tasks to agents preserves small, hyper-efficient engineering teams and keeps them moving at winning speed.
(links below)
10
2
68
23,544
USV retweeted
Excited to announce our partnership with @tryramp to offer day-one incorporation for startups directly in Cofounder.
You can now apply for incorporation, an EIN, and bank account without leaving the roadmap!
30
13
364
99,105
USV retweeted
What makes Alqem really cool imo isn't just that they're using AI to discover previously unknown compounds. They're also synthesizing and testing these materials in their lab. It's an alchemist's magical workshop that can lessen our dependence on rare earth mining.
We think the biggest unlock in materials science over the next decade won't come from a lab breakthrough but from AI finally being able to search a space humans never could.
That's the thinking behind our newest investment - Alqem. Here's why:
2
2
9
2,580
USV retweeted
I missed out on $100M+ by passing on the seed rounds of Suno and Granola.
Mike @mignano did both. He is the single best example in the last 5 years of an operator successfully turning into a GP at a large fund. It is so hard to do. He crushed it.
He has been a friend for 10 years and we shoot the shit today on the biggest questions of the day.
I summarized my notes from the chat below.
1. Do You Have to Produce Content to Do Seed Well?
To excel at seed, you need to put your ideas into the market so founders know exactly what you are looking for. Publishing your theses acts as a bat signal for early-stage founders before they even launch a product. Success comes down to building a great network, sharing your point of view, and betting on people.
2. Single Biggest Lesson About Venture
The biggest lesson in venture is to avoid projecting your own operator ideas onto a founder. It is their company, not yours. Making investment decisions based on your personal roadmap for them leads to bad bets. Venture success requires trusting the team’s independent judgment and execution plan.
3. What Percent of Enterprise Workflows Actually Require Frontier Models Versus Open Source?
Roughly 80% of non-coding enterprise tasks can run well on non-frontier models. Standard workflows like summarization and document generation are well suited to open source, which is improving quickly. This lets enterprises optimize token budgets without sacrificing meaningful performance.
4. Why We Believe Suno Has Unlimited Upside
Suno has unlimited upside because it unlocks a new consumer behavior: creative entertainment. Users generate music simply for the joy of creation, with no need for distribution or monetization. Like YouTube or TikTok, platforms that democratize a creative medium can create immense, generational value.
5. Obliterate, Don’t Automate
Avoid backing software that merely automates existing workflows to make old businesses slightly faster. The true generational opportunities come from completely reinventing industries, like putting an AI doctor in everyone’s pocket, rather than selling incremental efficiency to a corporate middleman.
6. The Shift From Infrastructure to Applications
As the capital-intensive AI infrastructure buildout matures, massive value creation is shifting to the application layer. Similar to the internet after the broadband boom, we are entering an explosion of specialized software. Winning requires an opinionated thesis, early conviction, and exceptional speed.
7. Why Startups Must Maximize Token Spend
While incumbents restrict budgets, startup CEOs should maximize token spend. Frontier models give lean teams a critical advantage over slower giants. Handing low-level tasks to agents preserves small, hyper-efficient engineering teams and keeps them moving at winning speed.
(links below)
26
8
241
119,644
.@IsembardGroup is betting the future of manufacturing looks nothing like the vertically-integrated primes of the last century. Instead, it's a software-driven network stitching hundreds of small machine shops into a decentralized production model.
"A software-driven network linking hundreds of small machine shops into a decentralized production model."
@IsembardGroup reinventing what scaled localized manufacturing looks like 🔥
1
1
14
5,323
USV retweeted
"A software-driven network linking hundreds of small machine shops into a decentralized production model."
@IsembardGroup reinventing what scaled localized manufacturing looks like 🔥
4
18
8,943
USV retweeted
In 2016, the week after we launched Anchor, I went on a little podcast called 20VC. It was my first podcast interview ever. Harry was maybe...18?
Now, 10 years later, we run it back IRL in London and talk about what we've both learned since, AI, the Rebel Alliance, & @USV.
I missed out on $100M+ by passing on the seed rounds of Suno and Granola.
Mike @mignano did both. He is the single best example in the last 5 years of an operator successfully turning into a GP at a large fund. It is so hard to do. He crushed it.
He has been a friend for 10 years and we shoot the shit today on the biggest questions of the day.
I summarized my notes from the chat below.
1. Do You Have to Produce Content to Do Seed Well?
To excel at seed, you need to put your ideas into the market so founders know exactly what you are looking for. Publishing your theses acts as a bat signal for early-stage founders before they even launch a product. Success comes down to building a great network, sharing your point of view, and betting on people.
2. Single Biggest Lesson About Venture
The biggest lesson in venture is to avoid projecting your own operator ideas onto a founder. It is their company, not yours. Making investment decisions based on your personal roadmap for them leads to bad bets. Venture success requires trusting the team’s independent judgment and execution plan.
3. What Percent of Enterprise Workflows Actually Require Frontier Models Versus Open Source?
Roughly 80% of non-coding enterprise tasks can run well on non-frontier models. Standard workflows like summarization and document generation are well suited to open source, which is improving quickly. This lets enterprises optimize token budgets without sacrificing meaningful performance.
4. Why We Believe Suno Has Unlimited Upside
Suno has unlimited upside because it unlocks a new consumer behavior: creative entertainment. Users generate music simply for the joy of creation, with no need for distribution or monetization. Like YouTube or TikTok, platforms that democratize a creative medium can create immense, generational value.
5. Obliterate, Don’t Automate
Avoid backing software that merely automates existing workflows to make old businesses slightly faster. The true generational opportunities come from completely reinventing industries, like putting an AI doctor in everyone’s pocket, rather than selling incremental efficiency to a corporate middleman.
6. The Shift From Infrastructure to Applications
As the capital-intensive AI infrastructure buildout matures, massive value creation is shifting to the application layer. Similar to the internet after the broadband boom, we are entering an explosion of specialized software. Winning requires an opinionated thesis, early conviction, and exceptional speed.
7. Why Startups Must Maximize Token Spend
While incumbents restrict budgets, startup CEOs should maximize token spend. Frontier models give lean teams a critical advantage over slower giants. Handing low-level tasks to agents preserves small, hyper-efficient engineering teams and keeps them moving at winning speed.
(links below)
3
2
62
13,880
One thing worth underscoring: most AI advantages get competed away quickly. alqem's doesn't.
Every material they synthesize produces data nobody else has. That data sharpens their models, which propose better candidates, which generates more proprietary data. It's a loop that gets harder to replicate the longer it runs.
We think the biggest unlock in materials science over the next decade won't come from a lab breakthrough but from AI finally being able to search a space humans never could.
That's the thinking behind our newest investment - Alqem. Here's why:
2
19
3,737
One thing worth underscoring: most AI advantages get competed away quickly. alqem's doesn't.
Every material they synthesize produces data nobody else has. That data sharpens their models, which propose better candidates, which generates more proprietary data. It's a loop that gets harder to replicate the longer it runs.
We think the biggest unlock in materials science over the next decade won't come from a lab breakthrough but from AI finally being able to search a space humans never could.
That's the thinking behind our newest investment - Alqem. Here's why:
3
16
4,344
USV retweeted
Exciting news! The second edition of The World After Capital is now finally available in the US also
3
8
59
6,739
We keep coming back to the idea that agents are becoming less like applications and more like cloud infrastructure. The cloud made compute, storage, databases, and messaging available as services that any developer could build on. We think something similar is happening with intelligence. Rather than every application being built around a single model or provider, they'll increasingly draw on networks of specialized agents coordinated through layers like orchestration, memory, routing, identity, and execution.
.@nickgrossman is betting that in the future, AI agents will look more like cloud infrastructure than digital employees.
"Part of the Rebel Alliance thesis for us is that agents are moving from personified people you hire to cloud infrastructure."
"We're not just talking about, 'Will there be other agents that compete with Claude or ChatGPT?' I think that's one version of what agents look like. Another version is cloud infrastructure where every company that's building any system is going to have agents and the orchestration of hundreds or thousands of agents built into their underlying platform."
"If you think of agents as infrastructure as opposed to agents as employees, I think that's where a lot of these forces come into play."
5
7
61
14,001
.@nickgrossman is betting that in the future, AI agents will look more like cloud infrastructure than digital employees.
"Part of the Rebel Alliance thesis for us is that agents are moving from personified people you hire to cloud infrastructure."
"We're not just talking about, 'Will there be other agents that compete with Claude or ChatGPT?' I think that's one version of what agents look like. Another version is cloud infrastructure where every company that's building any system is going to have agents and the orchestration of hundreds or thousands of agents built into their underlying platform."
"If you think of agents as infrastructure as opposed to agents as employees, I think that's where a lot of these forces come into play."
4
12
54
29,040
USV GP @nickgrossman explains the firm's "Rebel Alliance" thesis: AI won't be won by a handful of model providers. It will create an entire ecosystem of new companies.
"The Rebel Alliance is the idea that the opportunity is so much bigger than any one company and that there is a massive ecosystem exploding around these incredible magic core models."
"We believe there will be generational companies built up and down the stack... The opportunity is so big and so massive and so expansive, and there are so many forces pushing for this motion that there are going to be companies building around the stack."
"It feels like the beginning of the whole economy, but maybe the whole Internet, getting rewritten on the agentic stack and with an agentic approach."
2
5
37
17,801
Check out @nickgrossman on @tbpn at 4 ET today to talk about The Rebel Alliance and the Emergent Agent Stack.
Happy Wednesday. On today's show:
- @markpinc (Zynga)
- @luke10ferrari (Bending Spoons)
- @wayneting (Lime)
- @nickgrossman (USV)
- @SHUBH17_ (Integral)
- @_jeff_liu & @_JonWang (Assort Health)
- @EliotPence (Dominion Dynamics)
- @julianhoenig (Amble)
See you on the stream.
3
16
3,489
USV retweeted
Hiring 2–3 designers for ongoing contract work.
The job: turn complex ideas into clean, simple diagrams. Often.
You: enjoy translating into pictures, ship fast, want your work to be seen. NYC-based and up for collaborating in person now and then.
Me: Shouldn't be designing diagrams anymore. Can provide clear scope, real work, and pay on time every time.
Reply or DM of stuff you're proud of.
37
2
101
10,832