Angel Investing

Disclaimer: This is not personalized investment, legal or tax advice. Startup investing is very high risk, and you can lose everything you put in. In fact, you probably will.

Picking the best founders

Okay, so you've got deals coming in. That's sourcing.

Next is picking, where you build the judgment to back great startups at a high rate.

For picking, I'll cover founders first, then markets, then ideas, then how to confirm your hunches in diligence.

Founders come first because VCs rank them as the most important criterion for picking a winning startup. In a survey of 885 VCs, 95% named the team as an important factor and 47% named it the most important, more than any other factor.

Of course great market timing matters too, but it's rarely enough, because another team with the same idea at the same time and better founders will beat yours.

Reading a market means predicting a chaotic future, whereas founders—people—are easier to assess. You'll still be wrong some of the time, but it's far less of a guess, because a founder's past performance, unlike a stock's, says a lot about their future. Founders who had already taken a company public had a 30% chance of doing it again, against 18% for first-timers.

Another reason why great founders are so important is because the idea a founder starts with often isn't the one that it evolves into before winning. Great founders "pivot" well, so even when they start with a mediocre idea, they find their way to a better one in a better market, and bad founders don't have the skill to do that. In fact, plenty of companies you use started as something else:

• Slack was the internal chat tool of a game studio whose online game, Glitch, had failed.
• YouTube launched as a video dating site with the slogan "Tune in, hook up."
• Twitter started as a side project at Odeo, a podcasting startup that Apple's iTunes had made irrelevant.
• Shopify was the software behind Snowdevil, an online snowboard shop, until other merchants wanted the software more than the snowboards.

I've personally backed enough unicorns (startups worth $1B+) at seed to be confident in what great founder traits are. So here's what I look for, starting with identifying exceptionalism:

• Spiky, with a growth curve. Spiky means extremely strong in at least one thing, like sales, technical or recruiting ability. We're talking top 0.1% in the world. The most successful founders start out weak in many areas but outlier strong in at least one, ideally the one the business most depends on, and every top VC I've spoken to echoes this. The founder needs a steep growth curve at the same time, helping them learn fast and fix their weaknesses early, because at seed what counts is whether they can figure it out as they go.
• A history of outlier execution. If they're spiky, their history shows it: they've already accomplished big things, even so-called first-time founders. Patrick Collison won Ireland's Young Scientist award at 16 for writing a programming language, and at 19 he sold his first startup with his brother John, who was 17. Stripe came two years later and counts as their first big company, so a "first-time founder" has often been tinkering for years. In Ali Tamaseb's study of US unicorns started from 2005 to 2018, about 60% of founders had started a company before. Specifically, ask founders what they've become world class at, and what they built before this that they're most proud of.

Then look at how deep the founder has gone in their historical thinking:

• Mastering their idea maze. The idea maze (a term from Balaji Srinivasan, made famous by VC Chris Dixon of a16z) is every route to building an idea, dead ends included. The founder should have explored the history of theirs and the market realities of attempting it today: who tried this, why it failed, and why it only works now. Tamaseb found that fewer than half of unicorn founders had deep experience in their industry, so a founder can learn the maze from scratch.
  • What does this look like in practice? We press founders on hypotheticals about their market to see how far their thinking stretches. If they can't answer, they may be YOLOing, and a founder who YOLOs gets beaten by one who does their homework. Every $1B founder we've backed has depth—not shallow strategic thinking. I'll give you questions to ask founders in a moment.
  • Gokul Rajaram, who's invested in 700 companies, tests it by trying to throw founders off course. He names five or six other ways to solve their problem and asks why theirs is better. A problem has many possible solutions, so a founder's route only means something if they can defend it against the others.
• Founders need obsession. It's what drives them through the idea maze, which takes a lot of work. Their startup has to be what they think about in the shower, and it has to feel like life or death to them. Look for a chip on their shoulder, something they want to prove, because it often powers the obsession. Obsession also keeps a founder going past rejection: Canva cofounder Melanie Perkins took more than 100 rejections before Canva got funded, and Eric Yuan applied for a US visa nine times before he got one. Years later, as CEO of Zoom, Yuan personally emailed every customer who canceled to ask what Zoom could have done better.
  • Obsession shows up outside work too. At Chris Sacca's house, Uber founder Travis Kalanick won a few close games of Wii Tennis against Sacca's father, then switched hands, and Sacca's father didn't touch the ball for the next three games. Kalanick then pulled up the global leaderboard, where he was ranked second in the world. He was building Uber at the time. The lesson is that people who can be obsessed in one thing can become obsessed in another. Those who never obsess about anything may not have that necessary maniac in them.
  • Honestly, if you remember one thing from this guide, remember Kalanick switching hands and still sweeping the score board. Go find the person in your life who'd do that, and ask what they're working on.

More founder criteria you need to know about are how fast they move and who they hire to keep that pace:

• Speed. Matt Loszak of the unicorn Aalo showed me this better than anyone: the best founders are viscerally uncomfortable with slow progress, and feel it as a physical irritant. Less than a year after breaking ground, Aalo reached criticality (a self-sustaining nuclear chain reaction) with a test reactor on July 4, 2026, hitting the deadline set by Trump's executive order.
  • DoorDash's founders started very humbly. They built a marketing page for restaurant delivery in Palo Alto in an afternoon, and soon a stranger called to order Thai food. They delivered it themselves between Stanford classes, tracking orders in a Google Doc and their drivers with Apple's Find My Friends. Dispatching by hand taught them what their driver-assignment software should do. The point is how fast they collected these learnings, even when they had nothing built. They found faster paths to everything.
• A magnet for people. Startups don't run themselves, so the founder has to be persuasive enough to recruit great people to their cause. A founder is like Obi-Wan, who has to convince someone on dusty Tatooine that joining will transform their life. Stand in for their future hires and notice what the founder's pitch does to you. If a founder can get you excited about their future, they can probably do the same to a great engineer with a safe job.
  • Unicorn founders treat each early hire as foundational. They run a deliberate process to find the best person in the world for each role, not whoever is closest at hand ("hiring within arm's reach").
  • The first ten hires set the ceiling for everything that follows, and done right, the team becomes an advantage rivals can't easily copy.
  • A bad hire in an important early role does more damage every month they stay, so if the founder has hired anyone, ask who their first bad hire was and how long it took them to act. Charlie Songhurst, who's invested in 483 companies, finds that how quickly founders remove the first employee who doesn't fit often tracks closely with success. Acting fast shows they'll accept conflict to shape the company, and founders who avoid it often end up with a toxic culture.

Finally, when assessing outlier founders, ask what will keep them in it for ten years. Are they mission aligned? Building a unicorn takes a decade. Without a mission beyond money, most founders lose steam and get passed by hungrier competitors. WhatsApp founder Jan Koum grew up in Ukraine, where phone lines were often tapped. That drove him to create something new. The mission doesn't have to be world-saving, but it has to be genuine.

We've covered a lot in this section, so here it is as four questions:

What to ask on a diligence call

Diligence is the homework you do on a startup before you invest, and it starts on your first call or meeting. Let's cover how to turn those founder traits above into questions you can ask.

To start, a simple way to open is to ask how the business started. VC Trae Stephens of Founders Fund opens nearly every pitch meeting with the origin story, because it shows whether a founder arrived at the idea through real mission alignment or is just YOLOing on an opportunity. He's screening out what he calls whiteboard founders (I call them tourist founders), who want to be entrepreneurs first, so they list every idea on a board and narrow it down to the least bad one. He wants someone who's staked their time and reputation on an unpopular belief.

Stephens says the first 10 minutes should make it obvious that nobody is better placed to pursue this idea. Yasir Arafat, nuclear startup Aalo's cofounder, shows what that looks like. He grew up with regular power outages in Bangladesh and has wanted to fix energy for good ever since.

From there, don't spend much diligence time on what's obviously good. Brian Singerman, a former partner at Founders Fund, takes a company's strengths as given and spends the meeting digging into its weak spots, because that forces unrehearsed thinking. A great founder owns the weakness and has covered it with a strong teammate or a thought-through answer, or can reason it out on the spot.

Past the origin story, these are the questions I pick from, with what I count as strong and weak answers. Yours may differ over time.

I start with the person:

• What are you world class at?
  Strong: one thing, with proof and a growth curve. Weak: a list of competencies.
• What did you spend your spare time on as a kid or teenager?
  Strong: built or obsessed over something unprompted. Weak: only did what was assigned.
• What did you build in the last 12 months that nobody assigned you? Can you show proof, so we know this isn't an AI response?
  Strong: shows they're proactive and curious about building. Weak: only what was assigned.
• What's the most money you've made outside a paycheck, and how?
  Strong: shows their entrepreneurial hustle so far. Weak: only ever earned a salary.
• Who is someone senior in your industry you've cold-contacted? What did you send, and what happened?
  Strong: shows their networking hustle. Weak: can't name anyone.

Then test how well they know the problem and who'll pay to solve it:

• Who tried this before, and why did they fail?
  Strong: knows each attempt and what changed since. Weak: assumes earlier founders were less talented.
• What do you understand that others don't? (compiled YC interview questions)
  Strong: a specific, non-obvious belief backed by what they've seen firsthand. Weak: repeats what everyone in the space already says.
• Who signs the purchase order?
  Strong: knows who the buyer is and why they'd buy. Weak: can only discuss the science.
• What's most likely to kill this company?
  Strong: names the core risk and the plan for it. Weak: claims there's no real risk.
• How does this business get harder to compete with over time?
  Strong: names how it stays ahead of competitors and keeps buyers loyal as it grows. Weak: only cites patents.
• How will you get your first $10M, $100M and $1B in revenue?
  Strong: names the customers and channels at each step. Weak: quotes a share of a big market.

Finish with momentum and the plan:

• What changed in the last 30 days?
  Strong: real progress that shows how fast they're moving. Weak: plans and meetings.
• What have you made happen so far? (Sheel Mohnot of BTV, via Signature Block)
  Strong: concrete results that happened because of them. Weak: effort, plans and intentions.
• Who have you convinced to join?
  Strong: strong hires who took a risk on them. Weak: hires within arm's reach.
• Who are your next 10 hires? (David Tisch, when he ran Techstars New York)
  Strong: names the roles in order and why, because your money pays for what comes next. Weak: says they'll hire as needed.
• How much do you want to raise, and what milestone do you want to hit? (Arjun Sethi of Tribe Capital, via Signature Block)
  Strong: a specific milestone that sets up the next round. Weak: a round size with no milestone attached.
• How did you meet your cofounder, and what convinced you to "get married"?
  Strong: a deliberate search for the best possible person, rather than a haphazard YOLO. Weak: met recently, and it's untested.

These are the questions we've kept asking after years of pruning, so feel free to save them for your own calls.

Those lists describe a strong answer, but you being able to actually recognize one in real-time on a call is harder. (This is why you should always ask to record your calls, by the way.) You learn this mostly by doing 100+ calls. After that, you start to recognize what exceptional sounds like.

Until then, here's a thought exercise. In this made-up example, a founder building grid batteries answers "Who tried this before, and why did they fail?" Pick the best answer before you read on:

• A. "Nobody has really tried this. The big battery companies are slow, and they don't understand software."
• B. "Three companies tried. Two ran out of money and one got bought. They were too early, and we have a stronger team."
• C. "Three companies tried. Two died because cells cost too much for any utility to sign. The third had the economics but couldn't survive a three-year buying process. Cell prices have since fallen by more than half, and small power co-ops can now buy in six months, so we're selling to them first."

It's C. A doesn't know its own history, and B is the weak answer from the list: it sounds informed but blames the earlier teams. C knows what killed each attempt and what's changed since, and it answers "Who signs the purchase order?" before you ask. Most pitches sound like B, so after enough calls, a C stands out.

Picking markets and technology

Let's move on from founder criteria. Next is the market.

The first thing to know about markets is that the best investments typically look weird to start. Plenty of big winners looked odd at first:

• Uber began as UberCab, a way to summon a black luxury car from your phone for about 1.5 times a taxi fare. It looked like a limo service for rich people in San Francisco. No one was ride sharing with strangers, and that sounded unsafe.
• Twitch started as Justin.tv, a 24-hour live feed of one founder's life from a webcam on his head. Its gaming streams became Twitch, which Amazon bought for $970M in 2014.

Weird alone is no reason to invest, though, because it usually just means bad, and most founders who seem crazy turn out to be crazy. Portfolio math can still favor a pile of crazy-looking deals though over a pile of decent ones, because a decent deal rarely returns enough to matter. Twenty decent deals that each return 2x leave you at 2x. Nineteen zeros and one 100x leave you at 5x.

The point is that the biggest future companies, like Uber, usually break a long-held paradigm, and that's what winds up feeling weird in the early days.

Anduril is a recent example. In 2017, Palmer Luckey, who had left Facebook three years after selling it Oculus, started a defense company with Trae Stephens of Founders Fund and named it after Aragorn's sword in The Lord of the Rings. Defense was the industry Silicon Valley kept away from: a year later, Google dropped a Pentagon drone-imaging contract after thousands of its employees protested. Anduril built its own drones and surveillance towers with its own money instead of waiting for government contracts, and in May 2026 it was valued at $61B.

Your job as an angel investor is to tell which weirdness comes from a flaw and which comes from a massive insight.

One way to tell them apart is to ask what's about to change in the world that makes this possible—that people haven't clued into yet.

• You want ideas that couldn't have worked 24–48 months ago. This is called the "why now." If it was possible before now, then why hasn't someone already built a huge company doing it? Because it's not a great idea. But if it's only newly possible, then its future is unwritten.
• The change that makes an idea newly possible can be in technology, consumer behavior, regulation or distribution. Better batteries made AirPods and Tesla's electric cars practical. The SEC's 2016 crowdfunding rules let startups raise money from anyone, which led to Republic, a platform later worth $1B+. COVID moved work and workouts home, and Zoom and Peloton grew with it. Facebook's app platform let games like FarmVille reach millions of players cheaply. My startup guide has more examples.
• That tailwind should arrive within 2 years. A tailwind is an outside change that pushes demand toward the company. It's what makes something eventually have a "why now." If it hasn't arrived yet, that's okay, since it could mean you're early to something inevitable. Beyond 2 years, I'd probably not invest on the assumption it'll arrive, because predictions that far out are unreliable and the world changes in unexpected ways.
• Meanwhile, competitive threats can't neutralize this startup idea's approach. This is what I call orthogonal risk: what doesn't exist yet that would make this irrelevant if it did? Imagine you were investing in taxi dispatch software before Uber was born. Uber was your unseen orthogonal risk. No one talks about calling a cab anymore. The tailwind warning signs were there if you looked closely: GPS smartphones were spreading fast after Apple opened its App Store in 2008, a year before UberCab started. That means real time location and mobile apps were about to change everything. Spotting this tailwind takes real mental effort, and getting good at it over the years is how you better avoid bad investments.

When assessing how a startup idea fits into a market, you also need to ask how big the prize can be and who'll pay to get it there:

• Can the market reach $3B+? That test forces big thinking, and a company that falls short can still reach $1B. Remember to judge a market by what it can become, because today's number can fool you. Your job as an angel is to be good at predicting the future. Tobi Lütke says well-known VCs passed on Shopify, which lets anyone run an online store, partly because only 40,000 online stores existed in 2008. He calls that a category mistake, because a great company can grow the market itself instead of fighting over a static slice of it. Uber shows how badly a market estimate can miss. In 2014, finance professor Aswath Damodaran valued Uber at about $5.9B by treating its market as the existing $100B taxi business. Bill Gurley, an Uber board member, answered that cheaper, faster rides would bring in people who rarely took taxis. Uber went public in 2019 at about $82B. So ask who would start buying if the product were far cheaper or easier than what exists today. Those new buyers are the market a great company grows into.
• There's a first buyer. If nobody can say who'll pay for this, it's probably a dud.
• To assess buyer demand, ask whether buyers are pulling it out of a founder's hands. Interest is cheap, so look for customers truly chasing the product with cash in-hand: asking when they can have it, paying before it's done, or dropping a tool they already use, which costs them time and risk. If buyers only call it "interesting," the pain isn't big enough to beat the cost of switching.

Picking ideas

Mike Maples told Tim Ferriss that 93% of his firm's profits from exits came from companies that pivoted. Like I said, great founders often move on to better ideas, so I'm not going to make many points about picking ideas. Frankly, judging an idea also takes case-by-case knowledge of its competitors, technology, customers and market, and Claude will do a better job of that than I can.

The one point I do want to make on ideas, though, is about moats—what gives a company protection from competitors—because my experience refutes a massive misunderstanding that most of my VC peers still hold:

Most moats are illusions.

Few companies wind up having true intellectual property (patents), regulatory (government red tape protection), or network effect (Facebook, X, LinkedIn) moats. If they do have them, heck yeah—lucky you for investing in them.

But most companies simply have to keep moving fast enough to stay about 1.5 years ahead of the competition while building goodwill with buyers, which is why speed is on the founder quality list above.

Speed matters even more than being first. In a study of about 500 brands across 50 product categories, nearly half of the companies that pioneered their category failed.

Goodwill, by the way, means buyers stick with you after competitors catch up, because you were there first, they've worked with you for years and they trust you not to let them down. It's a soft moat, but it's what protects most companies.

A 1.5-year head start is why I said it's okay for markets and ideas to look weird early on. Investor Jerry Neumann says that for the first few years, a startup's early protection is that nobody else wants in, because the idea looks stupid and the market looks absent. Again, be okay with being early. Invest in something where tailwinds haven't even come yet, but you know they're coming, and you're timing things brilliantly. That's non-consensus investing. Not to be confused with moronic investing, which is investing in things that no one wants now and no tailwinds will ever materialize for either.

How much diligence to do

Here's a hypothetical for you: a startup makes free compliance software for contractors, and you as the investor know nothing about contracting. Pick your first move:

• A. Read market reports on construction software.
• B. Ask the founder to introduce you to a few happy customers.
• C. Call 10 contractors and pitch the product as if it were yours.

Kent Bennett of the VC firm Bessemer did C. Six of the 10 asked where they could get it, and he invested. He says only one or two of every 20 companies he digs into give a signal that clear. Reports are secondhand, and customers the founder picks were picked to be polite. A cold pitch to real buyers needs no industry knowledge, and strangers have no reason to flatter anyone, so his method works in any industry.

Consider this: eight publishers had turned down the first Harry Potter book when Bloomsbury's chairman, Nigel Newton, took the sample chapters home. Instead of reading them, he gave them to his eight-year-old daughter Alice, who came downstairs an hour later glowing and then nagged him for months to see what came next. He wrote J.K. Rowling a check for £2,500.

Like Bennett, he tested the product on its buyer. Skip the idiots in the middle.

Digging on demand like this pays. In a 2007 Kauffman study of angels in investing groups, investors who spent more than the 20-hour median on a deal earned 5.9x their money, against 1.1x for the rest, and multiples were twice as high on deals inside the angel's expertise. That's another reason to invest in what you know and who you know.

Here's how I'd spend those 20 diligence hours, in order:

• Ask Claude for a map of the space first.
• Pitch 10 buyers cold, as Bennett did.
• Ask for proof of demand. Revenue beats a deposit, and a deposit beats a customer's letter of intent (LOI), which is usually non-binding, so read it. Many great early deals have none of this yet.
• Talk to the team. Check that the founder has hired killers and that everyone tells the same story about the market and the technology.
• Use the competitors' products, then ask how theirs differs. The answer shows whether they explored the idea maze.
• Call references, and count how many go out of their way to say something amazing. One great reference can outweigh the pitch: Don Valentine funded Nvidia on a tip from Jensen Huang's boss, despite a pitch Huang calls horrible, then said, "If you lose my money, I'll kill you."
• If the field is new to you, bring in a technical expert to check for fraud and risks you can't see. Ask what specifically rules the company out, then ask the founders for evidence. Experts get new tech wrong too: Neil Armstrong doubted the commercial space newcomers in 2010, ten years before SpaceX flew NASA astronauts.
• Write a memo before you wire and when you pass. Cover the problem, why now, the competition, the market, the team, who might buy the company, why you're in and the strongest case against. Never edit it, so it keeps you honest when the results arrive years later.

Diligence pacing

How fast you run diligence depends on the fundraising round. Heavy investor interest means you should speed up your diligence in case the round fills. It isn't a reason to invest, and neither is a deadline, since urgency is how FOMO (fear of missing out) gets you into a bad deal. Pricier seed rounds do turn into unicorns more often: in Carta's data on 5,700 seed startups, the most expensive quarter hit 5.6% against 0.8% for the cheapest. That's still about 1 in 18, and you pay up to get in.

Also, don't say yes on the call. Calacanis takes a few days to answer, and our fund takes 7 to 14 days, with three of us looking at the deal in parallel.

Before a final call, decide what has to be true for you to move forward, so you don't drift into a yes or waste the founder's time. Then ask yourself whether you'd decide the same way if the round stayed open 30 days and you couldn't see the other investors.

Here's a rule of thumb to keep you from running in circles: if you don't know within 14 days, you probably don't care enough, or it's not enough of a hell yeah. Wait for the hell yeahs.

Passing costs less than it feels like. When in doubt, say no.