By Uri Poliavich
Capital is abundant again. Investor patience is not
There has probably never been more money looking for the next transformative technology company. Yet for many founders, raising capital feels harder rather than easier.
That apparent contradiction tells us something important about the venture market in 2026. Capital has not disappeared. It has become more concentrated.
According to Carta’s analysis of the US private markets, startups raised $30.4 billion in the first quarter of 2026, with more than 60 per cent of that capital going to artificial intelligence companies. Crunchbase, using a broader global dataset and capturing several exceptionally large AI financings, estimated that AI-related companies accounted for roughly 80 per cent of global venture investment during the same period. The precise percentages vary according to methodology, but the direction is unmistakable: enormous amounts of capital are available, while a disproportionate share is flowing towards a relatively small number of companies.
This changes the way I think about startup investment. When I look at an early-stage company, I am not particularly interested in whether the founders have learned how to present themselves as investable. Most experienced investors can see through a carefully rehearsed pitch relatively quickly. I am interested in something much more difficult to manufacture: evidence that the founders understand a problem unusually well and that reality has begun to confirm some of their assumptions.
This distinction matters because an interesting startup and an investable startup are not necessarily the same thing.
I have seen founders become preoccupied with the mechanics of fundraising: the size of the addressable market, the design of the pitch deck, the vocabulary investors expect to hear and, increasingly, the need to attach artificial intelligence to almost every business proposition. These things may help secure a meeting. They tell me relatively little about whether I would want to invest.
What I want to understand is how the founders think when the presentation ends.
Why this problem? Why now? What have they discovered about their customers that a competitor has not? Which of their assumptions have already proved wrong? What happened when the first version of the product met the market? And, perhaps most importantly, what did they do when reality contradicted the business plan?
At an early stage, investment is necessarily a judgment made with incomplete information. There is no spreadsheet capable of telling an investor whether a company with ten employees will become a billion-dollar business. If certainty existed, venture capital would not produce venture returns. The useful question is therefore not whether uncertainty can be eliminated, but whether the evidence available today makes that uncertainty worth financing.
The Investor Who Said No to Airbnb
Airbnb is useful precisely because its early history demonstrates how unreliable first impressions can be in venture investing.
In 2008, Brian Chesky and Joe Gebbia were trying to finance a company based on the improbable proposition that strangers would pay to stay in other strangers’ homes. Investors repeatedly declined. Chesky later published some of the rejection emails. One investor concluded that the potential market did not appear large enough for the firm’s required investment model.
With hindsight, that rejection looks absurd. I think that is the wrong lesson to take from it.
The investor did not know that Airbnb would become Airbnb. He was evaluating an unfamiliar proposition with limited evidence at a moment when the eventual scale of the market was far from obvious. Investors make decisions under precisely these conditions, and some of those decisions will inevitably be wrong.
What interests me much more is what happened next.
Chesky and Gebbia did not respond to rejection merely by becoming better at explaining the same idea. They gradually made the company itself easier to understand.
During the 2008 US presidential election, when they were struggling for money, they sold novelty cereal boxes called Obama O’s and Cap’n McCain’s and raised around $30,000. Paul Graham of Y Combinator saw something important in this apparently ridiculous episode. Founders prepared to manufacture and sell breakfast cereal to keep a technology company alive were demonstrating a degree of resourcefulness that could not easily be reproduced in a presentation.
After entering Y Combinator in early 2009, the founders went to New York, met hosts personally, photographed apartments and watched how people actually used the service. Graham’s subsequent advice to “do things that don’t scale” became famous, but its underlying logic is more interesting than the phrase itself.
At an early stage, efficiency can be less valuable than learning.
The founders were acquiring information about their market faster than a scalable system could have given them. They were replacing assumptions with observations.
This is much closer to what I look for in a founder.
Persistence by itself does not impress me very much. People can persist with bad ideas for years. What matters is productive persistence: the ability to remain committed to solving a problem while continuously revising assumptions about how it should be solved.
Airbnb’s founders did not simply continue believing. They continued learning. For an investor, that difference is fundamental.
Evidence Changes as the Company Changes
The evidence I expect from a startup depends heavily on its stage.
At pre-seed, there may be little more than the founders, a prototype and an insight into a market. It would be unreasonable to demand mature economics from a company that barely exists. At that point, I am looking primarily at the quality of the problem, the founders’ understanding of it and their ability to move quickly.
At seed, I expect the evidence to begin moving outside the founders themselves. Are people using the product? Do they return? Will anybody pay for it? Are the founders learning something from actual customer behaviour that they could not have learned from market research?
By Series A, the conversation should be different again. I want to understand whether customer acquisition can be repeated, whether retention is meaningful, whether revenue quality is improving and whether additional capital will accelerate something that is already beginning to work.
This is why I am sceptical of the idea that fundraising is principally a presentation exercise. A pitch deck can organise evidence. It cannot create it.
Dropbox is a useful historical example. When Drew Houston applied to Y Combinator in 2007, online storage was not a new idea. Numerous products already allowed users to store or synchronise files. Describing the size of the potential storage market would therefore have proved very little.
Houston instead demonstrated the experience.
He produced a short video showing Dropbox working and directed it towards the technology community most likely to understand the frustrations the product was addressing. The response provided something more valuable than another market-size estimate: evidence that people cared about the particular way Dropbox solved an existing problem.
That distinction remains relevant almost twenty years later. I do not need a founder to prove that the future is certain. I want to see signs that reality is beginning to move in the direction the founder predicted.
What I Look for in a Founder
Venture capital has developed a mythology around founder personality. We are encouraged to look for charisma, extreme confidence, obsession and an almost irrational ability to make an improbable future sound inevitable.
I am cautious about all of these characteristics.
Confidence is useful, but it is relatively easy to perform. Charisma can help a founder recruit talented people and sell a product, but it can also make weak assumptions sound more convincing than they deserve. Even persistence, one of the qualities most celebrated in entrepreneurship, is valuable only when combined with the ability to recognise error.
If I had to identify the quality I value most when evaluating a founder, it would be intellectual adaptability.
I want somebody who can defend an idea vigorously on Monday and abandon part of it on Friday because new evidence has proved it wrong, without interpreting that change as a personal defeat.
This is much rarer than confidence.
Research also complicates the popular image of the successful entrepreneur. Pierre Azoulay and colleagues at MIT, Northwestern, Wharton and the US Census Bureau examined millions of American entrepreneurs and found that the mean age of founders behind the highest-growth new ventures was approximately 45. Relevant industry experience was also strongly associated with entrepreneurial success.
I find this unsurprising. In technology, we sometimes confuse novelty with insight. A founder does not necessarily need to be the first person to encounter a problem. It may be considerably more valuable to have spent enough time with that problem to understand why previous attempts to solve it have failed.
But experience creates its own danger. Someone who knows an industry extremely well can become convinced that its existing assumptions are immutable.
The founders I find most interesting therefore tend to combine two characteristics that should theoretically work against each other: deep knowledge and intellectual discomfort. They understand an industry well enough to recognise its constraints, but remain sufficiently dissatisfied with conventional explanations to question whether those constraints still need to exist.
Slack provides an unusually clear example.
Stewart Butterfield and his team were not originally building workplace communication software. They were developing an online game called Glitch. The game failed commercially and was shut down. During its development, however, the team had created an internal communication system for itself.
They eventually recognised that the tool developed to support the failed company might be more valuable than the product the company had been created to sell.
Slack subsequently became one of the defining enterprise software businesses of its generation, went public in 2019 and was later acquired by Salesforce for approximately $27.7 billion.
I find this story more useful than the conventional celebration of founders who refuse to give up. Butterfield did give up — on the wrong product.
The persistence was directed towards finding value, not protecting the original idea.
AI Has Changed What I Expect from Capital
The current AI investment boom creates another contradiction.
Artificial intelligence is absorbing extraordinary quantities of capital at the top of the market. Frontier models, semiconductors, data centres, energy infrastructure and computing capacity require investments measured in billions.
At the same time, AI is reducing the amount of capital required to build many ordinary technology companies.
A small team can now use AI to write code, generate design, analyse customer behaviour, automate support, produce marketing materials and perform administrative tasks that would previously have required a much larger organisation.
For me, that changes the capital-efficiency question.
If technology allows eight people to accomplish what once required forty, I do not automatically regard a very large seed round as evidence of ambition. I want to know why the company needs the money.
What specific uncertainty will the capital remove?
Will it demonstrate customer demand? Complete a difficult technical prototype? Obtain regulatory approval? Build manufacturing capacity? Establish repeatable distribution? Take a sales model that already works in one market and test whether it works in five others?
“Growth” is not a sufficient answer.
Capital should purchase evidence.
The relationship between money and progress matters particularly in an environment where founders can accomplish considerably more before raising institutional capital than they could even five years ago.
AI Is No Longer a Moat
I apply a similar test when evaluating companies that describe themselves as AI startups.
Five years ago, sophisticated generative AI capabilities could themselves differentiate a product. That advantage is disappearing rapidly. Powerful models can now be accessed from OpenAI,
Anthropic, Google, Meta and other providers without a startup developing a foundation model of its own.
This is enormously beneficial for innovation. It also makes one of my first questions very simple.
If your competitor can access substantially the same intelligence tomorrow morning, what exactly do you own?
“AI-powered” is becoming similar to “internet-enabled” twenty years ago. It describes part of the technology stack. It does not necessarily describe a competitive advantage.
The defensibility may instead lie in proprietary data, distribution, workflow integration, network effects, specialised domain knowledge, regulatory access, switching costs or the accumulation of customer behaviour that makes a product progressively more useful.
This may partly explain why some of the most interesting investment activity is moving beyond generic AI applications towards robotics, defence technology, aerospace, cybersecurity, energy infrastructure and advanced manufacturing.
Hadrian is an interesting American example. The California company is building highly automated factories for aerospace and defence manufacturing. Its proposition is not simply that AI can make manufacturing more efficient. It combines software, automation, physical infrastructure and specialist manufacturing capabilities in a market where barriers to entry are substantial. In 2026, the company announced $1.37 billion in new capital and commitments to expand its manufacturing footprint.
The distinction matters to me. AI can be copied surprisingly quickly. A functioning factory, a specialised supply chain, proprietary operating data and years of accumulated manufacturing knowledge cannot.
I Do Not Invest in Fashion
Founders often ask which sectors are currently most attractive to investors. The obvious answer in 2026 is artificial intelligence. It is also one of the least useful answers.
By the time a sector has become visibly fashionable, thousands of founders and billions of dollars are already moving towards it. Fashion may help a startup secure a meeting. It cannot make the company investable.
I am more interested in understanding why capital is moving towards a sector.
Current investment in AI infrastructure, defence, robotics, cybersecurity, energy and advanced manufacturing reflects deeper economic forces: geopolitical competition, labour shortages, reindustrialisation, increasing demand for computing capacity, cybersecurity threats and the growing connection between digital intelligence and physical infrastructure.
Those forces matter more than the label attached to the company.
The startup opportunities I find most interesting frequently appear where a technological change intersects with an economic constraint.
Uber did not become important simply because smartphones were fashionable. GPS-enabled phones, mobile payments and dense urban demand made a different transportation model possible.
Airbnb was not simply another marketplace. Online payments, digital reputation systems, photography and changing consumer behaviour reduced enough of the trust barrier between strangers to create a market that previously looked implausible.
Stripe did not succeed because internet payments were new. Patrick and John Collison recognised that accepting payments remained unnecessarily difficult for developers building online businesses and reduced a complicated technical process to something dramatically simpler.
In each case, technology mattered because it unlocked behaviour.
That is more interesting to me than technology for its own sake.
Not Every Great Business Should Raise Venture Capital
There is one final distinction I think founders frequently underestimate.
A good company is not necessarily a good venture investment.
A business can be profitable, well managed, useful to its customers and capable of creating substantial wealth for its founders while being completely unsuitable for a venture-capital portfolio.
That is not a criticism of the company. It is a consequence of the mathematics of venture investing.
A venture fund expects many investments to fail or produce modest outcomes. A relatively small number therefore need to generate exceptionally large returns. The potential market, scalability and eventual value of a company matter because the successful investments must compensate for the unsuccessful ones.
This creates a mismatch when a founder builds a company designed for steady profitability and then raises capital from investors whose economic model requires exceptional growth.
I would therefore ask a question before discussing valuation: should this company be financed by venture capital at all?
Taking venture investment does not simply add money to a balance sheet. It changes the expectations surrounding the company. It affects growth targets, governance, future financing and ultimately the expected path towards liquidity.
Capital comes with an idea of what the company should become.
Sometimes the best decision an entrepreneur can make is not to take it.
What Makes the Uncertainty Worth Financing?
The venture market of 2026 can appear contradictory because extraordinary amounts of capital are entering technology while investors are becoming increasingly selective about where ordinary venture dollars go.
I do not see a contradiction. I see a market becoming more demanding about evidence.
When I evaluate an early-stage company, I do not expect certainty. If certainty existed, venture returns would not. Nor do I expect a founder to have correctly predicted every part of the business before building it.
I look instead for a progression.
Has the original hypothesis survived contact with customers? Has the team learned something that was not obvious when the company began? Is the product becoming more difficult to replace as it develops? Can the founders explain why customers stay rather than merely showing that revenue has increased? Does additional capital remove a specific constraint, or does the company simply require money in order to continue discovering what its business might eventually be?
These distinctions matter more to me than the polish of the fundraising process.
Airbnb was initially easy to dismiss because the evidence was weak and the proposition unfamiliar. Dropbox entered a market that already appeared crowded but demonstrated that a substantially better experience could change user behaviour. Slack emerged because its founders were prepared to recognise that the most valuable thing they had built was not the product they originally intended to sell.
None of these stories demonstrates that investors should finance every unusual founder with an improbable idea. That would be a terrible investment strategy.
They demonstrate something more useful: the quality of a startup becomes visible through the way evidence accumulates.
This is why I would never advise a founder to begin by asking how to make a company attractive to investors. The better question is what the company needs to demonstrate so that investment becomes a rational response to what has already been learned.
Good investors are not looking for certainty. They are looking for asymmetry: situations in which the evidence remains incomplete, but what has already been discovered suggests that the potential outcome is disproportionately large.
The founder’s job is not to remove uncertainty from that equation.
It is to show that the uncertainty is worth financing.
Short Biography
Uri Poliavich (born 1981) is an Israeli entrepreneur, philanthropist, and public thinker working at the intersection of technology, digital systems, and their impact on economic and social structures. He currently lives and works in the United States.
Born in Ukraine, he moved to Israel as a teenager, where he completed his education and military service. He later earned a law degree from Bar-Ilan University and began his professional career in commercial law and international business development.
Over time, Poliavich transitioned from legal practice into technology-driven entrepreneurship, building and scaling international digital ventures. His background as a developer and operator of complex digital systems later became the foundation of his work as a commentator on how modern platforms shape economic behavior.
In 2020, together with his wife, he co-founded the Yael Foundation, an international philanthropic initiative dedicated to expanding access to high-quality education. The foundation supports schools, educational programs, and community initiatives across dozens of countries, working to strengthen both academic standards and cultural identity among young people.
Under his leadership, the foundation has developed into a global educational network, providing funding, strategic support, and institutional development to a wide range of educational institutions and initiatives. Its work focuses not only on access, but on long-term sustainability and measurable outcomes in education systems.
Poliavich is also the founder of the Responsible Engineering Lab, an initiative focused on the transparent and ethical use of digital technologies. Through this work, he advocates for a clearer public understanding of how algorithmic systems, platforms, and artificial intelligence shape access to information, economic opportunity, and social outcomes.
He writes and speaks on the transformation of technology, digital systems, and their impact on economic and social structures.
