Investment Thesis

How I think about angel investing, in my own words.

The Infinite Hold

I invest in startups I wish I had built, hold them permanently, and borrow against appreciating positions to fund an ever-expanding portfolio. I exit only when a business becomes irrelevant, the culture changes, founders lose integrity, or the math overwhelmingly favors reallocation. The goal is an infinite compounding machine that grows to as many investments as a lifetime allows.

Why I invest

I do not invest in startups as a financial exercise. I invest because these are companies that, in a parallel universe, I wish I had built myself. Every investment I make represents a founder whose problem I care about, a product I admire, and a mission I would dedicate my own life to. I am not buying equity. I am joining a team. That is why I never want to leave.

When I look at my portfolio, I do not see ticker symbols or entry prices. I see the future being built by people I believe in. That emotional conviction is not a weakness in my investment process. It is the foundation. It is what makes holding permanently feel natural rather than forced.

The case for holding permanently

The greatest technology companies do not stop growing when they go public. They accelerate. Look at the companies everyone can name: the bulk of the value they created came after the listing, not before, and the pattern repeats endlessly.

If the best-known venture firms had simply held every position and never sold, they would be far larger businesses today. The entire venture capital model, invest, exit, return capital, raise a new fund, is structurally designed to force selling at precisely the moment when compounding is most powerful. I reject that model.

My approach is different. I hold. Not because I am passive, but because I understand that the best outcomes in technology come from the longest holding periods. Every year I do not sell is a year my winners continue compounding. Every exit I avoid is a taxable event I defer. Every position I maintain is collateral I can borrow against.

The compounding machine

Here is the mechanism I am building.

Deploy. I invest in startups with personal capital, spread across sectors and geographies so that no single outcome decides the result.

Compound. The best companies in my portfolio grow from tiny to enormous. I do not sell at any of those milestones. I hold through every later financing, through a public listing, and beyond.

Collateralize. As positions become liquid, they become borrowable assets. I borrow against my holdings rather than selling them, which unlocks capital without triggering a sale or a tax event.

Reinvest. The borrowed capital funds the next generation of angel investments. Those investments grow, become collateral, and fund the generation after that.

Repeat. Each cycle adds more collateral, more borrowing capacity, and more investments. The portfolio becomes a self-reinforcing engine.

Can this actually work? I believe yes. The logic is straightforward: as long as the portfolio appreciates faster than the cost of borrowing against it, the spread funds continuous deployment into new startups. Every public listing expands the collateral base. Every year of appreciation increases borrowing capacity. The cycle compounds.

My exit conditions

I am not dogmatic about holding forever in every case. I will exit a position under a small set of specific conditions.

The business becomes irrelevant. Technology shifts, markets evolve, and some companies fail to adapt. If a company's core product or market is structurally declining with no credible path to reinvention, holding becomes value destruction rather than value creation.

The culture fundamentally changes. I invest in missions and people. If a company's leadership abandons the values that made it worth backing, if the culture shifts from innovation to extraction, from ambition to complacency, then the company I invested in no longer exists, even if the legal entity persists.

The founders lose integrity. If I discover that the leadership is dishonest, misleading investors, manipulating metrics, or acting unethically, trust is broken permanently. I cannot compound alongside people I do not trust, regardless of how promising the business looks on paper.

A better use of capital demands it. If selling a position at fair value allows me to deploy that capital into something with dramatically higher expected upside, and I cannot access that capital through borrowing, I will make the trade. This is rare, but intellectual honesty requires acknowledging that no position is sacred if the math overwhelmingly favors reallocation.

Outside of these conditions, I hold. Through drawdowns, through bear markets, through temporary setbacks. My conviction is in the long-term trajectory of technology, not in quarterly earnings.

The long-term bet

The underlying thesis is simple: technology companies, public or private, are the greatest wealth-creation engines in human history. Software scales without limit. AI compounds intelligence. Deep tech solves problems that unlock entirely new markets. The secular trend is undeniable and accelerating.

As long as these conditions hold, permanent holding outperforms any exit-driven strategy. I am not building a portfolio to sell. I am not optimizing for distributions or fund metrics. I am building a portfolio to keep: a growing, compounding, self-funding machine that lets me invest in as many startups as a lifetime allows.

Decades from now, I do not want to look back at a list of companies I sold too early. I want to look at a portfolio of companies I believed in, held through everything, and watched become defining institutions of their era. That is the game I am playing.

Proof in public markets

6 companies in this portfolio have reached public markets, the clearest validation of the Infinite Hold, because a public listing is the moment the conventional model sells and this one keeps holding. Alongside them, as of October 2026, 10 companies have been acquired, 5 have shut down, and 213 are operating privately, out of 234 published here.

This site shows each company's name, category and status. It does not show what is held in any of them or what any position is worth; the rules on that are set out on the FAQ.

How I evaluate

Four things run through every decision.

Founder-first

Team quality and execution ability matter more than market-size slides. I look for founders with domain expertise and an unfair advantage: a reason this person, of all people, is the one to solve this problem. I am joining a team, so the team is the decision.

Speed and conviction

I make decisions fast. If the founder, market, and timing align, I commit quickly rather than putting a founder through a drawn-out diligence process. Conviction, once formed, is worth more than a longer list of questions.

AI-native

AI is the most transformative technology shift since the internet. I am heavily weighted toward companies building the infrastructure and applications of AI: the models, the chips and compute beneath them, the agents on top, and the businesses that AI makes possible for the first time.

Active support

I am not a passive check-writer. I help with introductions, strategic advice, hiring, and the next fundraise. Because I hold rather than sell, my interest in a company does not end at the next milestone, and neither does the help.

Where I focus

The portfolio is organised into 14 categories and 69 sub-categories, and every company page carries both. The themes below are the ones I care about most.

AI. Foundation models, AI infrastructure, applied AI, and machine learning platforms. The core thesis.

Fintech. Neobanks, corporate cards, payments infrastructure, and financial automation tools.

Robotics. Humanoid robots, autonomous vehicles, industrial automation, and manufacturing.

Software. Enterprise software, developer tools, collaboration platforms, and business-to-business applications.

Aerospace. Space technology, satellite communications, launch vehicles, and orbital manufacturing.

Defence. Autonomous defence systems, cybersecurity, intelligence platforms, and military technology.

Beyond these, the portfolio reaches into climate and energy, biotech and health, semiconductors and hardware: anywhere technology creates a step-function improvement rather than an incremental one.

The labels and the themes are not the same thing. A category is a label a company carries on its page, and a company can carry several: a humanoid-robot company sits in Robotics and in AI, a battery company in Climate and in Hardware. The themes above are what I am looking for underneath the labels, which is why an AI company can appear in almost any category and why the categories overlap as much as they do. If you are unsure which label your company would carry, do not worry about it; say what you are building and the label will follow.

What I look for

Strong signals

• Technical founders with deep domain expertise
• Clear founder-market fit: why this person for this problem
• Products with early traction or waitlists showing demand
• Markets undergoing structural shifts (AI, regulation, demographics)
• Capital-efficient business models with a path to profitability
• Strong co-founder dynamics and complementary skill sets
• Alumni of accelerators such as Y Combinator and Stanford StartX
• Repeat founders, or operators from top-tier companies

Red flags

• Solo non-technical founders in deep-tech spaces
• Overcrowded markets without a defensible moat
• Founders who cannot articulate their unfair advantage
• Businesses that require massive capital before any validation
• Teams that have not talked to customers before building
• Pricing expectations that run ahead of anything the company has proven
• Copy-paste business models from other geographies
• Founders who optimize for fundraising over product

Track record

In December 2024 Lenny Rachitsky published his ranking of the most active angel investors in the U.S., built from Crunchbase data. He wrote that I was "the #1 most active angel investor in 2024," and that I was the most surprising person on the list to him: he had not heard of me, and my list of investments struck him as incredible. His ranking comes with its own caveat, that it rests on the data available to it, and I would rather repeat that caveat than drop it.

Why founders choose me

I have been a founder more than once. I co-founded MovieLaLa, which Gfycat acquired, and founded RemoteTeam, which Gusto acquired; MojiLaLa and Leo AR are also mine. I have been through every part of the founder journey, including the part where you sell the company, so I know what it takes to get to the finish line.

I went through Stanford StartX, 500 Startups, FI, Betaworks Vision, and Quake Accelerator as a founder, which left me with a deep network across the startup ecosystem. That network is most of what I have to offer after the money.

I also publish the whole portfolio, including the companies that failed. Other investors and founders learn more from the full record than from a highlight reel, and a portfolio that hides its losses is not one anyone should trust. The pages here show what each company does, its category and its status; they do not show terms, and they are not meant to.

What this means for founders

If you are raising and thinking about writing to Sahin, the thesis above translates into a few practical things.

You are pitching a holder, not a flipper. An angel who intends to hold through a public listing is not going to push you toward an early sale and is not watching a fund clock. The flip side is that he is choosing a company he would want to belong to for decades, so the mission and the people carry more weight than the spreadsheet. Write the email that a future teammate would want to read, not the one a banker would.

Start with why you. Founder-first means the first question is whether you and your team are the people to solve this problem. Domain expertise, an unfair advantage and evidence that you have already talked to customers do more work than a market-size slide. If the email could have been sent by any competent team, it is missing its most important sentence.

Expect a quick answer, either way. Speed and conviction cut both ways: a fast yes and a fast no are both more likely than a long process. That means the first email has to carry the argument on its own. Do not save the interesting part for a call.

Be honest about where you are. Two of the red flags above, pricing expectations that run ahead of proof and teams that build before talking to customers, are about candour more than about progress. Early is fine. Vague is not.

AI is the centre of gravity, not the boundary. The core thesis is AI, but the strong signals are about structural shifts of any kind, and the portfolio's 14 categories reach well beyond AI. If technology is creating a step-function improvement in your market, you are in scope.

Ask for what you need after the money. Introductions, hiring help and support for the next fundraise are stated parts of the offer, and because he holds rather than sells, they do not stop at the next milestone. Founders who use that get more from the relationship than founders who do not.

Read the portfolio before you write. The category and sub-category pages show which companies here are adjacent to yours, and each company page lists the investors that public sources report alongside him. That is useful in two ways: it tells you whether your market is one he already knows, and it tells you who else has backed companies like yours. It does not tell you what any company here raised or on what terms, because the site does not publish that.

Nothing on this page is investment advice, and nothing here describes terms; those are discussed directly, if at all. The practical next step is the pitch page.

Frequently asked questions

Does the Infinite Hold mean you never sell?

Almost never. He exits only when a business becomes irrelevant, its culture fundamentally changes, its founders lose integrity, or a far better use of capital demands it. Outside those conditions he holds, through drawdowns and bear markets.

Do you only invest in AI?

No. AI is the core thesis and the heaviest theme, but the portfolio spans 14 categories, including robotics, aerospace, climate, semiconductors, biotech, fintech and software.

What do you do after investing?

Introductions, strategic advice, help with hiring and support for the next fundraise. He describes himself as not a passive check-writer, and because he holds rather than sells, the involvement does not end at the next milestone.

How do I get my company in front of you?

Through the pitch page: one email covering what you are building, why now, why you, where the company is today, what you are raising, and a link to the deck. There is no guarantee of a reply.