investment strategy

A few years ago, I started paying more attention to how people evaluated technology companies.

Most conversations sounded familiar.

How fast is the company growing?

How large is the market?

How much funding has it raised?

Who are the founders?

Those questions matter, but after spending more time around software products, engineering teams, AI systems, blockchain infrastructure, and early-stage companies, I started looking at investments differently.

I stopped asking only, “How big could this become?”

I started asking, “What can break this business?”

That change in thinking has probably been more useful than any complicated investment model.

Growth Can Hide Technical Debt

A product can look successful from the outside while becoming increasingly fragile underneath.

Imagine a startup growing from 10,000 users to 500,000 users in a year.

From an investor's perspective, the chart looks excellent.

But inside the engineering team, something else may be happening.

The database was designed for the original product. Services depend on each other in ways nobody fully understands. Deployments require manual steps. Monitoring catches problems only after customers complain.

Every new feature makes the system slightly harder to maintain.

Eventually, growth stops being an advantage and starts becoming pressure on the architecture.

This is why I pay attention to technical debt when evaluating technology businesses.

Technical debt is not automatically bad. Every company takes shortcuts.

The real question is whether management knows where those shortcuts are and whether the engineering organization has a realistic plan for fixing them.

AI Has a Reliability Problem

AI has created one of the biggest technology investment opportunities of this decade, but it has also created a strange engineering challenge.

Traditional software usually behaves predictably.

Give the system the same inputs and you generally expect the same result.

AI systems don't always work that way.

Models hallucinate. Outputs change. Prompts behave differently after model upgrades. Costs can increase quickly when usage grows.

A company saying, “We use AI,” tells me almost nothing.

I want to understand what happens when the AI is wrong.

Does a human review important decisions?

Can the company measure model accuracy?

Does it know how much inference costs per customer?

Can it switch models if its current provider changes pricing?

Those are not small engineering questions. They can become major business risks.

A strong AI product needs more than a good model.

It needs good data, evaluation systems, monitoring, fallback mechanisms, and clear limits around where automation should be trusted.

Security Is Usually Invisible Until It Isn't

Security is another area investors often underestimate because good security is difficult to see.

When everything works, security looks like nothing happened.

Then one incident changes everything.

A leaked API key, compromised dependency, exposed database, or badly designed authentication system can turn years of growth into a crisis.

Modern software also depends heavily on third-party packages.

A development team may write 50,000 lines of its own code while indirectly depending on millions of lines maintained by other developers.

That creates supply-chain risk.

I don't expect every startup to operate like a bank, but I do want to know whether security is part of engineering culture or something the company plans to think about “later.”

Later can become very expensive.

Blockchain Has a Different Kind of Risk

Blockchain businesses introduce another interesting technical problem.

Code can control real assets.

That changes the consequences of software bugs.

In traditional software, a bad deployment might cause downtime.

In decentralized finance, a smart contract error can move or permanently lock millions of dollars.

This is why audits, permissions, upgrade mechanisms, key management, and contract architecture matter so much.

I am also cautious when a blockchain project's entire value depends on token price instead of product usage.

The stronger projects usually have a reason to exist even when market speculation slows down.

They solve infrastructure, identity, payments, settlement, ownership, verification, or coordination problems that remain useful regardless of the market cycle.

The Best Investment Question I Know

Today, when I look at technology opportunities, I try to understand four things.

Can the system scale?

Can it be secured?

Can the economics survive growth?

Can the product continue working when one important dependency fails?

That last question is especially useful.

What happens if AWS goes down?

What happens if OpenAI changes its pricing?

What happens if a blockchain becomes congested?

What happens if the company's biggest data provider disappears?

Good engineering teams think about these scenarios before they happen.

Great companies build around them.

That doesn't mean avoiding risk.

Technology investing will always involve risk.

The goal is understanding which risks are temporary problems and which ones are fundamental weaknesses.

That distinction matters.

A startup may have ugly code today and still become an excellent company if it has strong engineers and understands what needs to change.

Another company may have a beautiful demo but no defensible architecture, no security discipline, and economics that collapse as usage increases.

The second company often looks more impressive at first.

The first one may be the better investment.

Smart Strategy Is Mostly About Asking Better Questions

Investing in technology is not only about predicting the next trend.

AI will change industries.

Blockchain will continue evolving.

Automation will remove entire categories of manual work.

New platforms will appear.

But the companies that survive usually solve much less glamorous problems: reliability, security, infrastructure, data quality, distribution, and cost.

Those are the problems I increasingly pay attention to.

Because behind almost every impressive technology company is an engineering problem somebody solved extremely well.

And behind many failed technology companies is an engineering problem everyone noticed too late.