Martin Tobias has spent most of his career getting to major technology shifts early.
He joined Microsoft when it was still small enough for employee equity to create founder-level wealth. Then he founded Loudeye Technologies, built it from three people to 450, and took it public. He later started a renewable-energy company, became a venture partner at Ignition Partners, made hundreds of angel investments, and launched Incisive Ventures.
He has now backed eight investments that reached unicorn status.
Most of what we cover on this show is built around avoiding permanent loss. Real estate, private equity, debt. You underwrite the downside, protect the cash flow, and try to produce a relatively tight range of good outcomes.
Venture works almost exactly backwards.
Most of the investments can fail. A result that looks extraordinary in another asset class may barely move the fund. And one extreme outlier can make nearly everything else irrelevant.
Martin has lived both sides of that equation. He created a $500 million paper fortune through one concentrated position. Then he built a portfolio designed to survive being wrong most of the time.
Here are the parts worth your time.
The $500 million was accurate, but it wasn't cash
When Loudeye went public, The Wall Street Journal calculated Martin's net worth at more than $500 million.
The number was accurate. It was also mostly theoretical.
Martin was locked up for six months and couldn't sell his shares. Even after the lockup expired, a public-company CEO faces trading windows, limits on how much can be sold, and the signaling problem that comes with the CEO unloading stock.
Within 12 months, the value of his position fell roughly 90%.
He didn't buy a new house. He didn't buy a new car. He didn't rebuild his lifestyle around a number he couldn't access. He already liked the life he had, so he left it alone.
A friend made the opposite decision. The friend had about $150 million in newly public stock, borrowed $20 million against it, bought a large house in Miami, and started spending as though the valuation had already become permanent wealth.
The stock fell 95%. The loan was called. He lost the house.
That is the distinction founders miss. The stock is volatile. The loan payment and lifestyle costs are not.
An IPO can be a route to liquidity without being an actual exit for the person holding the shares. Until the position can be sold or diversified, spending against it creates a very certain liability against a very uncertain asset.
The wife joke was the setup. The 80% was the answer
I asked Martin which purchases or investments actually improved his life after he became wealthy.
His first answer was a new wife. In his words, probably the best upgrade you can make.
The useful financial answer came next.
Martin watched people take one big win and put everything back on the table. They made money once, decided that proved they were exceptional capital allocators, and immediately started taking the same level of risk again.
He went the other direction.
Martin put roughly 80% of his winnings into low-leverage, cash-flowing assets, including commercial real estate. The remaining 20% became his higher-risk capital for venture and similar investments.
The important part isn't copying those percentages. It is the order of operations.
If you spend $50,000 a month, the first problem after a large exit isn't finding another 100x opportunity. It is creating a conservative, post-tax income stream that reliably covers the $50,000 with room to spare.
Once that is solved, the rest really can be upside.
Martin's current portfolio is unusual, and he says directly that he wouldn't recommend it to everyone. About 75% is in wholly owned, low-leverage commercial real estate. Approximately 5% is in public markets. The remaining 20% is in venture and private equity.
The real estate covers his burn rate. He believes he could lose the entire risk portfolio without changing his monthly life.
That is a better definition of being rich than a large number on a brokerage statement.
Venture math is supposed to look terrible
Within Martin's B2B software thesis, he says six of 75 investments became unicorns. That is an 8% unicorn rate.
It also means 92% did not become unicorns.
His working model for a venture portfolio is even more extreme. Approximately 80% of the investments may go to zero. Around 15% may produce good outcomes in the 10x to 50x range. The remaining 5% to 8% need to become genuine outliers capable of returning more than 100x.
In a 35-company fund, one 100x investment made at a standard check size can theoretically return approximately three times the entire fund even if everything else goes to zero.
That is completely different from real estate.
If we buy a good real estate deal and execute reasonably well, the expected outcome might fall somewhere around 1.5x to 2.5x over five years. We aren't underwriting one property to return 100 times our money while assuming most of the other buildings disappear.
In venture, a 10x result may not be enough to drive the fund. If it is one position in a 30-company portfolio, it only returns about one-third of the capital.
You need the outlier.
That doesn't make venture better than real estate, private equity, or debt. It gives venture a different job in the portfolio.
The more reliable assets protect the floor and produce cash flow. Early-stage venture provides exposure to an outcome the more reliable assets almost never produce.
Martin's conclusion is not to put everything into startups. It is that someone with a sufficiently large balance sheet may want a single-digit allocation to early-stage venture as an asymmetric option.
I think that framing is right. Fund the life first. Then buy the upside.
Building software got cheaper. Getting attention didn't.
Martin focuses on very early B2B software companies. His description is “two guys and a dog.”
The attraction is capital efficiency.
Ten years ago, building a serious software company required servers, hosting infrastructure, and a team of engineers. Martin estimates it could cost $5 million before the company had much of anything.
Today, a founder can begin with a $20 AI subscription and inexpensive cloud infrastructure. In Martin's first fund, five of roughly 30 companies reached profitability after raising less than $2 million.
A three-person software company may become profitable at $350,000 to $400,000 of revenue.
That still doesn't make it a venture-scale company.
The venture question is whether the company can grow revenue at something closer to 10x a year for several years. Small-scale profitability is useful, but extreme growth is what creates the outlier.
The scarce resource has also changed.
Ten years ago, finding enough technical talent to build the product was often the hardest part. AI has made building easier. That means more people can launch products, which makes getting customer attention harder.
Martin now looks for two things in the founding team: someone who deeply understands the customer and someone capable of building the product. Then he looks for a distribution wedge; a credible way to reach the first 100 customers without spending a fortune.
The product became cheaper to build.
Trust, distribution, and attention did not.
AI may remove the apprenticeship before it removes the profession
Martin has a 15-year-old daughter and says he doesn't know what to tell her to study.
When he went to college, computer science was an obvious answer. Learn to program and you could enter the workforce with a valuable, well-paid skill.
He thinks entry-level programming is now one of the first job categories being materially disrupted by AI. The same thing is beginning to happen to the junior work historically given to lawyers, architects, engineers, and consultants.
The interesting problem isn't only whether AI eliminates jobs.
It is what happens when AI eliminates the work people used to learn the job.
Martin's first tasks at Accenture included building PowerPoint presentations for senior people. It wasn't sophisticated work. It was apprenticeship. He sat inside the process, watched how decisions were made, and learned how to become useful.
If AI completes all the basic work, companies still need some way to develop experienced professionals.
My view is that roles built around trust, judgment, and relationships have more protection. Strategic sales is an obvious example. Medicine has a long tail of human judgment. The trades have physical execution that software cannot easily replace.
In our paving business, simply answering the phone, returning a quote quickly, and running a professional process can separate an operator from competitors who take weeks to call someone back.
Martin's formulation is sharper: don't focus on finding a career AI cannot touch. Become the person in your field who is best at using AI.
The advantage belongs to the AI-native operator, not the person relying on a credential that used to guarantee a job.
A venture fund buys selection, not just diversification
Before Martin invested professionally, he made 250 angel investments.
The process was familiar. Meet a smart founder. Review a polished deck. See that a friend is participating. Write a $25,000 or $50,000 check.
He eventually concluded that it wasn't a very good strategy.
As an angel, he reviewed approximately 20 to 30 opportunities a month. As a venture manager, he sees roughly 300. That larger funnel gives him a better view of what is being funded, what comparable businesses look like, and what a reasonable price should be.
The structure of the round also matters.
Angel-backed companies can end up raising money hand to mouth. If the business burns $50,000 a month and raises roughly the same amount each month, it never receives enough runway to complete a meaningful set of objectives.
A professional venture round may fund 18 to 24 months. That gives the company time to answer the questions that actually matter: Can the team build the product, and will customers buy it?
Despite making hundreds of direct investments, Martin says his personal angel portfolio performed worse than his venture portfolio.
The difference wasn't merely diversification. The fund selected from a larger opportunity set, used a more consistent process, and provided enough capital for the companies to reach meaningful milestones.
For anyone trying to become a venture manager, Martin's advice is equally direct: invest your own money first. Build a record. Then raise individual deals through a syndicate. Only after people have seen your judgment should you ask them to commit capital to a blind pool.
We see the same trust ladder in real estate. Raising money for a specific property is easier because investors can inspect the asset. A blind pool requires them to underwrite the manager before they know what the manager will buy.
You earn that authority through evidence, not a pitch deck.
Give the full episode a listen. Martin is unusually candid about what a $500 million paper fortune actually felt like, how he protected the wealth that survived, and why successful venture investing requires being comfortable with a portfolio that looks wrong most of the time.
— Sam Silverman
Silverman Capital
