Fruits of the Walled Garden
Companies are climbing further up the stack
When Infrastructure Climbs the Stack
When generative AI first broke into the mainstream, companies like OpenAI and Anthropic were understood primarily as infrastructure providers. Developers were encouraged to build on top of them, with the promise that AI models would be the foundation layer of a vast new ecosystem of applications. But today, these same companies are climbing further up the stack.
Take OpenAI’s recent release of Sora2, a consumer-facing app for video generation. What was once just a raw capability (text-to-video) is now packaged into an end-user experience, competing head-on with startups that thought they’d have room to build applications. Similarly, Anthropic has launched Claude Teams, not just offering API access to Claude models but delivering a ready-made productivity suite for enterprises.
Think of the model companies as farms. They used to sell ingredients to restaurants — the startups — who turned them into meals. Now the farms are running restaurants too. To stand out, you can either cook better with the same ingredients or source ones no one else can.
More from the a16z fintech team
The Greenfield Strategy: AI-native startup Bingo
By James da Costa & Alex Rampell
With every new technology cycle comes a new set of business software companies. AI will be no different. In past cycles, much of enterprise software consisted of digitizing an offline process: turning paper records into databases with a front-end interface. In the first era of software, companies like PeopleSoft digitized HR records and made them available on-prem. In the cloud era, Workday made those digital HR records available anywhere, displacing PeopleSoft. Now, AI can do the work of humans who once used the software: software is becoming labor.
a16z Finance Forum
Ivan Makarov and Seema Amble hosted our second a16z Finance Forum Demo Day at the NYSE. The event featured over 200 CFOs and presentations from AI startups disrupting the finance stack, including: Concourse, Rillet, Vantage, Sequence, Runway, Sphere.
“Never let a good crisis go to waste.”
By David Haber
The SVB / First Republic banking crisis of 2023 made managing uninsured deposit balances existential for all regional and community banks. This crisis accelerated ModernFi’s consolidation of the market with the development of the leading bank-led deposit network. With more than 500+ banks in the network (with hundreds more to come) the flywheel is spinning.
a16z Partner Interview, Seema Amble: 3 things she looks for in early-stage startups, how to get a meeting with her, & more
Recent M&A Deals and Market Intel
IPO activity continued over the past month, with Ethos filing with the SEC on September 26 to raise up to $100 million and Wealthfront following on September 29 with an estimated $250 million offering. Meanwhile, PicPay is reportedly exploring a U.S. IPO that could raise around $500 million as soon as this year, though the size and timing remain under discussion.
Morgan Stanley announced its acquisition of EquityZen on October 29. The deal supports Morgan Stanley’s strategy to expand its private markets platform, deepening relationships with private companies and offering MS at Work participants additional liquidity options, while broadening access to private shares for Wealth Management clients.
Barclays’ U.S. Consumer Bank (USCB) announced its acquisition of Best Egg on October 28 for $800 million. The deal complements USCB’s partnership-driven credit card business and adds digital and risk capabilities that strengthen its consumer finance franchise. The transaction value represents a high single-digit P/E multiple.
OpenAI completed an acquihire of the Roi team, the company behind an AI-powered personal finance app. Roi CEO and co-founder Sujith Vishwajith announced the move on October 3, noting he is the only member of the four-person team joining OpenAI.
S&P Global announced its acquisition of With Intelligence on September 30 for $1.8 billion. The deal expands S&P’s data coverage, insights, and analytics capabilities across private markets, strengthening its position in alternative asset intelligence.
FIS announced its acquisition of Amount on September 24. The deal strengthens FIS’s digital account origination capabilities with a unified platform that integrates AI-driven features across deposits, lending, and card solutions for SMBs.
This newsletter is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. Furthermore, this content is not investment advice, nor is it intended for use by any investors or prospective investors in any a16z funds. This newsletter may link to other websites or contain other information obtained from third-party sources - a16z has not independently verified nor makes any representations about the current or enduring accuracy of such information. If this content includes third-party advertisements, a16z has not reviewed such advertisements and does not endorse any advertising content or related companies contained therein. Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z; visit https://a16z.com/investment-list/ for a full list of investments. Other important information can be found at a16z.com/disclosures. You’re receiving this newsletter since you opted in earlier; if you would like to opt out of future newsletters you may unsubscribe immediately.
The walled garden dynamic in fintech is fascinating. Platforms that control the data layer end up controlling the relationship layer too. From what we see tracking VC deal flow at VC Brain, the funds that back fintech infrastructure plays are betting on exactly this: whoever owns the rails sets the rules.
This really clicks for me because it names the dynamic a lot of builders are quietly living through right now: the “platform” you depend on eventually wants to become the “product” your customers use. That is not malicious... it's just gravity. If you sit on the capability, you will eventually package the capability.
The walled-garden data framing is a strong antidote to “my moat is my prompt.” But I think there’s a second layer that matters just as much: the workflow moat that manufactures the data moat. The winning vertical tools will not just license a corpus; they will earn the right to continuously generate cleaner, more structured, higher-trust data because they are embedded in the daily work. Said differently, defensibility comes less from owning a static pile of information and more from owning the feedback loop that keeps it current, compliant, and "decision-grade".
One tension I can’t stop thinking about: “walled gardens” are great for startup defensibility, but they can be terrible for the broader ecosystem if they become pure toll booths. The most interesting companies might be the ones that build a garden with gates, meaning they capture value while still making the surrounding market more legible through APIs, standards, or selective openness.
Curious where you land on this: in 3–5 years, is the durable edge “exclusive data,” or “exclusive distribution of the workflow that creates the data”? My gut tells me the latter.