Fred Wilson
Related people
Episodes
Notes
Sourcery · USV’s New $900M Fund: Fred Wilson, Rebecca Kaden & Mike Mignano
-
Does Union Square Ventures Need Trillion-dollar Winners?
Fred Wilson says Union Square Ventures does not need to underwrite to a trillion dollars. A company has to look like a market that can get bigger. Treating today's market size as fixed usually means the investment does not make sense. There was a time when a company worth half a billion dollars was enough for a yes, and then that number moved to a billion. A trillion is not the bar.
If the biggest winners are in the single-digit billions, the fund size and the ownership can still produce healthy returns for the partners. That is where it has been. Some outliers reached $50 billion or $100 billion. A lot of the biggest winners have been in the $10 billion range, and that range produced enorm
-
What Does Obliterate, Don't Automate Mean?
Fred Wilson wrote a blog post in 2014 called Obliterate, Don't Automate. The line is the Union Square Ventures thesis of restructuring a market and pushing value out to the end user, away from the middle. Healthcare, education, and manufacturing are the kinds of markets in reach now, because the tool set is new.
The plain version is: do not hire a lawyer, use an AI lawyer. Do not hire an accountant, use an AI accountant. Do not hire a doctor, use an AI doctor. No human in the loop. It is not about selling AI to those professionals. It is telling an AI to prepare the tax return, giving it access to the bank, and not paying a person to do it. 20 to 30 percent of the legal work at the firm is
-
What Is the Rebel Alliance in AI?
Fred Wilson says Rebel Alliance is the open side of AI, set against established winners: OpenAI, Anthropic, Grok, and Gemini, which is Google. Around them sits a world of open-source models, open-weight models, open-source harnesses, and other infrastructure a developer can stitch into something that looks like Claude or Cursor.
Developers do not want to build on a closed stack. The worry, using Anthropic as the example, is that the company they build on will compete with them and cut them out. That already happened in early social media, when businesses built on Facebook and Facebook later wanted those businesses for itself. The name is for people building a stack they control. That stack
-
Why Did Union Square Ventures Back Radiant Before the AI Power Rush?
Mike Mignano says Union Square Ventures has done well at extremely early bets because it is thesis-driven. Coinbase was very early in crypto. Abridge was very early in AI and in capturing healthcare data. The framework is a strong team and a direction for the idea, even when the first details are wrong.
Radiant is the energy case. Fred Wilson is on the board. The bet was made before it was obvious that AI would need energy to be abundant. The team already had a point of view about energy, and that view is what led to the investment. The firm backed Radiant in 2021 and is five years into the bet. A thesis can also be wrong. The benefit of having one is that the firm can back a space that is
-
Why Can ARR Hide Weak Product Retention?
Rebecca Kaden says annual recurring revenue has become a number a lot of investors watch, because revenue has ramped so fast at many of these companies. A clip from an earlier Sourcery interview has her calling ARR a fake metric, or something close to that wording. The claim under it is to look a layer deeper.
Many of these products are subscriptions. It is easier than ever to get someone to pay without knowing whether they like the product. A lot of plans are annual, and often you have to pay just to try the thing. The check instead is true usage retention: whether people come back every day, every week, or are gone after a month. Holding the revenue for a year, two years, or three will sh
-
Why Do AI Markets Reward Founder Storytelling?
Rebecca Kaden says the founding team matters more than ever, for two reasons. Hiring is harder. Excellent people are paid a lot by large labs, incumbents, and startups that have raised billions, so a founder has to pull them away from those offers.
The other reason is the story. Market narratives are aggregating faster than they used to, and value often shows up before the business does. The founder who drives that story, gathers the attention, and gets treated as the owner of a market early matters for investors, for the team, and for customers. The firm is paying much more attention to that storytelling than it did in other markets. If the product is easier to build, the founders worth ba
-
Why Does Union Square Ventures Keep Its Funds Small?
Rebecca Kaden says a small fund, even one that is only small relative to other funds, is easier to return, and the partners can be pickier. The firm wants the fund built from the bottom up. It starts from strong ideas about what it wants to do. It does not start from a requirement to own some set share of the universe of companies so the math works. The fund should not get bigger than that strategy allows. The number has shifted, and that shift was an ongoing conversation with the limited partners.
A small fund also lines up incentives. Results stay focused on how the portfolio companies perform, and the firm rallies around that. A very large fund can distort those incentives. That alignmen
-
Why Would Shoppers Want Their Own AI Agent?
Fred Wilson says people will want their own agent, not only the agent that belongs to each store. Amazon, Shopify, and Etsy can each have an agent. A shopper will still want one agent that talks to those agents and goes off to do the task, instead of managing every conversation.
Today those agents log into services and pretend to be the customer. Amazon does not like that. Amazon told Facebook, the week of this conversation, that Muse cannot access Amazon, and more blocks like that are coming. The expected next step is a standard way for a person's agent to connect to the agent each service offers. It is early, and the claim is that people want an agent that represents them.
-
Why Are AI Funding Rounds Still Getting Bigger?
Rebecca Kaden says products are dramatically easier to build. The firm recently incubated one quickly and cheaply, and other builders are doing the same. More startups then have to stand out from each other. A lot of the capital goes to that breakout: marketing, messaging, and positioning, especially in consumer.
Compute is a cost that was not there before. Open-source models are a direction the firm likes, and spend on compute sits in a different place than it used to. Nobody knows when models stall or how low compute goes, so that cost is now part of the company. The companies getting funded are also more varied than at any other time in Union Square Ventures' history, from software appli
-
Why Does AI Look Like the Internet in 2003?
Fred Wilson says this moment in AI looks like the internet in 2003. Commercial internet had arrived in 1995, and by 2003 a lot of the underlying infrastructure was built, while the big applications were still ahead. That year already had Google, Amazon, Yahoo, and eBay. It did not yet have Facebook, Twitter, YouTube, or the iPhone. The bet then was that new applications would be built on top of the web.
The same shape is what he sees now. A lot of the AI infrastructure is in place, and the large applications on top of it are still to come. The follow-up in the room is that an application layer is starting to show, in products such as Granola and Suno, and in a newer agent layer such as Inst
-
Why Is It Cheaper to Build a Product but Harder to Break Out?
Mike Mignano says a startup can still be expensive to operate and scale when the product itself is easier to build. Inference is becoming a real line item. Great products get built, and more startups pay for that inference so customers can use them before those customers have to pay.
Standing out is the harder bill. As it gets easier to start a company, more companies show up, and breaking out of the pack gets harder. Startups get more aggressive about go-to-market and marketing so they can get through the noise. Building the product is arguably easier than ever. Running it, scaling it, and breaking out are not.
-
How Did Abridge Go From One Product to Three?
Fred Wilson says Abridge went from one product to three after making all of its money for the last five or six years on a single product. The AI for healthcare company, a star of the 2016 fund, launched two new products in the last 12 months, really two new lines of business.
USV is marked by very long-term investing. Everyone there is still working on companies funded over a decade ago, and the test as companies grow is whether the team can launch new products and new lines of business. At the annual CEO summit last week, a casual conversation covered exactly that record: all the money for five or six years from one product, and the greatest pride in the two launches over the last 12 month
-
Why Does Union Square Ventures Stay Early Stage?
Fred Wilson likes the earlier stage of investing. When Union Square Ventures started, the plan was writing about a $3 million Series A check. The check size has changed a lot over the years, and early stage is still where the firm wants to be.
A typical Series A today starts at 10, and 10 is now on the small side. Some Series A rounds reach 100 or a billion, outside the firm's zone. The Series A checks the firm believes in for the fund will probably be 10 to 30. The checks do not have to be the whole round, so the rounds can be bigger. The firm will also do a lot of seeds and some deals earlier than seed. Exceptions to the usual rule are expected, and a strategic check written in a differen
-
What Is New About Union Square Ventures' Latest Fund?
Fred Wilson says Union Square Ventures closed its most recent set of funds and raised $900 million. It continues what the firm has done for decades: thesis-driven work with founders, and backing businesses it thinks will lead their markets over the next decade. The amount is also bigger than the firm has raised before, somewhat significantly, and it marks a new strategy.
The market has evolved. Outcomes are bigger, the swings are bigger, and the chance to use the new technology and the applications built on top of it is significant. Rounds are bigger, more expensive, and faster, so the strategy is changing to keep up with that market.