Elizabeth Yin πŸ’› Profile picture
Aug 24, 2019 β€’ 12 tweets β€’ 4 min read β€’ Read on X
Some Saturday morning thoughts on the next wave of unicorns:

1) We invest in the US, CAN, & SEA, and ppl ask us why we lump all our investments into 1 fund as opposed to regional funds. This is because the next wave of startups are global from day 1.

What do I mean by that?
2) 5 years ago, you'd be crazy to try to go global first. At that time, it was unfocused and challenging to manage. Today, you're at a disadvantage if you are not global first.
3) Let's dive into the data. 5 years ago, most unicorns that were started then were in the US and China. news.crunchbase.com/news/unicorn-c…

I believe that we will continue to see this trend. But you have to look under the hood at what that means.
4) If you look at the up-and-coming companies that will be unicorns in the next couple of yrs, many of them will be SF HQed companies. But their roots were elsewhere. E.g. companies like Front, Talkdesk, & I'd put Intercom in that category too.
5) And it's important to look at beginning stories, because that's how startups get their foot on the first rung of the ladder. E.g. If you can keep your costs down on engr talent AND manage well, you can go far.
6) I recently caught up w/ 2 past portfolio cos. I was floored that both were doing $5m ARR w/ fast growth and had raised very little seed funding. I asked them how they were able to build so much and sell so much w/ so little.
7) All of their devs were in Vietnam and Argentina (respectively). Both teams had a tech co-founder leading product outside the US. Both teams sell to companies in the US and have all customer acq (sales / mktg / BD) in the US.
8) When I look at my @HustleFundVC portfolio which is newer, I also see the same trend. Even if not hiring abroad, they are hiring outside of SF. One of my current portfolio cos who actually has raised a lot of cash based in SF has more employees in Dallas than SF.
@HustleFundVC 9) Then I have a couple of current portfolio cos doing some crazy global arb things. E.g. Canada gives startups a TON of grants reducing the cost of tech talent: a 3-5x difference to SF tech talent costs. 1 of my SF based cos set up a Canadian entity to take advantage of this.
@HustleFundVC 10) On the business side, we have portfolio cos selling to Asia and SEA cos selling globally / to the US. 5 years ago, this seemed weird / impossible. Today, this makes a lot of sense for the right biz.
@HustleFundVC 11) When I looked at our portfolio, I could not quite "bucket" so many of our companies. They are SF based & US corp but have most ops and devs in Y country / city. Or a SG incorporated co that sells to the US.
@HustleFundVC 12) There's no clear cut geography anymore and founders who can navigate this new way of hiring / managing remotely (or w/ hub and spoke) / and selling remotely are at a serious advantage in this new economy.

β€’ β€’ β€’

Missing some Tweet in this thread? You can try to force a refresh
γ€€

Keep Current with Elizabeth Yin πŸ’›

Elizabeth Yin πŸ’› Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @dunkhippo33

Apr 4
Anthropic's policy changes just forced me to optimize my OpenClaw setup, and honestly... it was overdue. While I'm not super happy about this, I get it - they have a company to run, and they are losing so much money. Here's what I decided to do. More >>
1) I had looked into other models including Gemma but ultimately decided to stay with Anthropic. The biggest waste was coming from heartbeats when there was no work to do. So I just slowed mine from frequent pings to once every 60 minutes.
2) My OpenClaw doesn't have nearly enough to do, so it doesn't need to check for tasks constantly if tasks are rare. Also, my OpenClaw now goes completely inactive at night. No use burning through tokens on heartbeats when there's no work.
Read 6 tweets
Apr 2
I tried Paperclip - an open source project that lets you set up an "autonomous company" with multiple AI agents. The concept is wild: you act as a board member setting vision while agents coordinate to build and run the company. More >>
1) What works: The agent orchestration is surprisingly smooth out of the box and was easy to get going. AI content writers, SEO agents, and engineers all coordinate without much manual intervention. It's the closest I've seen to truly autonomous business operations.
2) However, I found that even "autonomous" companies need nudges. My workers would get stuck, or the SEO specialist needed better direction, and coordination would break down more than expected. Still way less work than building a company manually, but not hands-off.
Read 7 tweets
Mar 27
I'm seeing a lot of founders, incl myself, build AI apps with tons of integrations.

It's tempting to integrate to everything - Gmail, Notion, Slack, Telegram, SendGrid... but there's a cliff your app will fall off of if users connect with all these integrations. More >>
1) Latency stacks fast: Every integration adds round trips. What starts as snappy responses becomes sluggish as integrations compound.
2) Context window bloat: More tools = more tokens loaded per request, even for unused integrations. Your token budget gets eaten up before you even start.
Read 7 tweets
Mar 24
The startup landscape has changed dramatically in the last 2 years with AI's rise. But there are 5 specific changes that have been absolute head fakes for many people. Here's what I'm seeing: 🧡
1) Product-market fit is really easy to lose now. I used to believe that if you got past $10M ARR, you were pretty set. These days, that's not true. I've seen companies reach high levels then go to zero because of fast followers and AI-powered competition.
2) Software moats are much harder to come by because anyone can vibe code something in a day. This has made older, less tech-savvy industries MORE attractive because once you get in with your workflow, they're less likely to rip you out for new software.
Read 7 tweets
Nov 22, 2025
Most AI startups will go bankrupt within 18-24 months.

After reviewing hundreds of AI startups' financials at Hustle Fund, I'm seeing a sustainability crisis that nobody's talking about.

The numbers are brutal. 🧡 Image
1- The typical pre-A or series A AI startup is burning $500K/month. A lot of this is on GPUs.

At that pace, a $10M funding round only buys you 20 months of runway and no margin on the revenue.
2- The problem is that companies are growing users.

But expensive GPUs eat into any and all revenue that a company may be making.

There’s just no margin.
Read 10 tweets
Apr 30, 2025
Yesterday, I spoke with some students at UCLA about entrepreneurship, and people asked me what makes for a good startup idea.

Here are some "layered" thoughts...

More >>
1) The first layer you have to cross with a good startup idea is that you need to at scale be bringing in more money than is going out the door.

This is obvious.
2) But, many entrepreneurs don't understand how this breaks down. Your COGs + marketing spend must be less than the lifetime value.
Read 13 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us!

:(