She Works in VC, and She Says VC Has No Money
“And more importantly,” she said, “VC has no money.”
Manage a million in assets and you collect two hundred thousand in fees over ten years. LPs don’t subtract that fee from the principal. In their eyes, the principal is a million and the expectation is 1.4 million. What the GP has to do is take the eight hundred thousand he actually holds and turn it into 1.4 million.
She isn’t a programmer; she studied finance. VS Code is open on her computer every day with Claude Code running inside it, and her daily workflows are written as md files that run automatically.
“First how you do it, then how you scale it, then how you make money.”
I went to see Amy because I wanted to find out whether there was room for my company and her VC to work together. She does early-stage investing in tech startups at a venture capital (VC) firm in London.
I asked her whether a service like AI transformation applies to VCs.
She said no. A lot of people in VC come from technical backgrounds, and since everything they invest in is a frontline tech startup, they have picked up something of the same character: technically driven, inclined to do everything themselves.
“And more importantly,” she said, “VC has no money.”
I was surprised. An investment firm, with no money?
She explained it to me. There are one or two hundred VCs in the London market, and the top five may take 60% of the entire market. For people working in VC, a large part of their income comes from management fees. At 2% a year, managing a million in assets brings in two hundred thousand in fees over ten years. That two hundred thousand has to cover hiring people and keeping the firm running.
Her boss counts as a general partner, a GP. A GP’s job is to persuade other people to hand their money over for him to manage. The people who put up the money are called limited partners, LPs.
LPs expect a 1.4x return. They don’t consider the principal to be the eight hundred thousand left after two hundred thousand in fees. In their eyes, the principal is a million and the expectation is 1.4 million. What the GP has to do is take the eight hundred thousand he actually holds and turn it into 1.4 million.
“It’s quite a lot of pressure,” she said.
Of course, with some luck, if you hit a company like Uber, the ten-year return may go far beyond 1.4 million. Whatever exceeds it gets split between the two sides at a ratio they agreed on.
We also talked about the enterprise AI transformation my company works on. She said, “This doesn’t need funding. There’s no use for VC money here.”
One more thing I hadn’t expected: the VC industry has a summer holiday, every August.
It started with senior VCs taking time off to be with their kids. With the senior people away, the junior staff couldn’t get much going either. Later, founders discovered that raising money in August was hard, so they simply stopped raising in August, and the junior staff had even less to do. A tacit August holiday loop, closed just like that.
VS Code is open on Amy’s computer every day, with Claude Code running inside it.
She isn’t a programmer. She studied finance, and what she does is early-stage investing in tech startups. She exports part of the data from the management system to her local machine, hooks Claude Code up to APIs for other data sources, and writes her own daily workflows as md files that run automatically every day. This setup has freed up a lot of her time. The firm has now put her in charge of taking it further and optimizing the processes across the whole company: sorting the data into public and private, with the public part serving as context shared across the firm.
I asked her what counts as public and what counts as private.
“A lot of what gets said in conversations with other VCs is private. You talk about work for a while, and then you might talk a bit of gossip. In our industry, you know how it is, information matters.”
Conversations with founders are mostly public, shared across the firm and evaluated together.
Amy doesn’t come from a technical background. When she was an undergraduate in the US, in order to practise her English, she didn’t cluster together with the other students. She moved into a community of foreigners instead, forcing herself to handle daily communication in English.
“It’s flood therapy,” she said, laughing. “You throw yourself into the crowd.”
She said she’s grateful for the age of AI. A lot of the technical things that used to be a slog to learn are finally learnable now.
I asked Amy how she got into this line of work in the first place.
“Chance,” she said with a laugh.
Before joining the VC, she was a founder too. The product was a fitness mirror: users work out following the mirror, and can also use the mirror to correct their own form. Hardware and software combined, sold overseas on a subscription. The software team and the manufacturer were both in China.
She found that the manufacturer they worked with did well on the things that mattered most. The mirror surface was smooth and very flat. But the small details often went wrong: coils kept falling off, and the materials weren’t good. The software team was the same. They built things fast, but always missed small spots. On an overseas user’s screen, for instance, a popup in Chinese would appear.
Amy felt that the users this product was aimed at were precisely the group who care most about details. She thought those details mattered a great deal. Her co-founder thought they mattered less.
In the end the thing didn’t go on. It did become her way into VC.
I asked her what she thinks is one of the hard parts of building AI software right now.
She said user stickiness. With AI agent products, switching costs are too low, and plenty of similar products can do similar things. Especially now that everyone’s user experience is good, one authorization and it connects to a user’s private data. Users don’t mind switching, so willingness to pay is naturally low.
I followed up: so when you invest, how do you judge whether to go in?
She said: first how you do it, then how you scale it, then how you make money.
Finally I asked whether she would still want to go out and start a company in the future.
Amy thought about it, and didn’t say no.