I've been in the Crypto industry for six or seven years, and for the past two years, I've also been deeply involved in the AI sector, based in Silicon Valley. Being active in both circles, a very obvious feeling is: in mainstream Silicon Valley circles, the term "Crypto" is mentioned less and less, but the things Crypto does are being used more and more.

I want to bring back some signals from the AI side for Crypto practitioners to consider.

This misalignment is most evident in YC.

YC Winter 2026 has just been announced, with 5 out of 149 companies being Crypto-related. This number isn't high, but if you look at the historical data, you'll find that these 5 companies hide a very clear story behind them.

A Set of Data

YC started investing in Crypto projects in 2014 and has invested in 177 companies so far. I pulled out the numbers for each batch, and the change is直观的:

2018-2019, 3-7 companies per batch, steadily climbing. 2020, 5-7 companies per batch, starting to accelerate. 2021, jumped directly to 13-15 companies per batch. 2022 reached the peak—Winter batch invested in 24 companies, Summer batch invested in 20 companies, a year with 44 Crypto companies invested.

Then came the cliff.

2023 still had 10-13 companies per batch, holding on for a year. 2024 began to collapse—Winter 7 companies, Fall 4 companies, Summer dropped directly to 1 company. For a whole summer, YC only invested in 1 Crypto company.

Winter 2025 briefly rebounded to 10 companies, but then Spring and Summer dropped to only 2 companies per batch.

By Winter 2026, 5 companies.

If you are a Crypto practitioner, seeing "from 1 back to 5" might feel like a warming signal. But if you look at what these 5 companies are actually doing, you'll find they are almost a different species compared to the 24 companies in 2022.

What were the Crypto companies YC invested in in 2022 doing? DeFi protocols, NFT infrastructure, DAO tools, L2 scaling, chain games, social tokens.

What are these 5 companies in 2026 doing? Stablecoin deposit API, cross-border neo-bank, trade execution engine, AI Agent payment gateway, attention exchange.

None are building chains, none are building protocols, none are working on any traditional "Crypto track" you can name.

This is not a warming up, this is a blood change.

Three Definitive Projects

First, a quick overview of three relatively easy-to-understand ones.

Unifold, New York team, building Stripe for Crypto deposits. An API+SDK set, allowing any App to integrate cross-chain, cross-Token on-chain deposits with less than 10 lines of code. Co-founder Timothy Chung previously worked on Streambird (wallet-as-a-service, later acquired by MoonPay and became MoonPay Wallets), and also spent time at Polymarket and Instabase. The other founder, Hau Chu, graduated from Cornell Tech. This is a typical developer tool business—users don't need to know the underlying layer is Crypto.

SpotPay, San Francisco team, cross-border neo-bank based on stablecoins. CTO Thomas was previously at Google, and was the 4th engineer at Brex. CEO Zsika is also from Google, Stanford MBA, grew up in the Caribbean and Latin America, personally experienced how painful cross-border remittances are. The product is straightforward: one account handles overseas收款, local payments, global spending (with a physical card), and interest-bearing savings. Underlying runs stablecoins, but the front end is just a Fintech App, with no visual relation to Crypto.

Sequence Markets, New York, 5-person team, building intelligent trade execution for digital assets. Helps institutional investors with smart routing across exchanges to get better prices and lower slippage. Fully non-custodial, doesn't touch user assets, only does the technology layer—a typical pick-and-shovel model.

The commonality of these three is clear: Crypto is the pipeline, not the selling point.

Two Projects Worth Discussing More

Orthogonal—When AI Agents Spend Money, They Will Use Crypto

This project is one I think Crypto practitioners should take a serious look at.

Founder Christian Pickett previously worked on payments at Coinbase and also spent time at Vercel. Bera Sogut worked on reCAPTCHA and Maps APIs at Google, also at Amazon Robotics, a two-time ACM ICPC (International Collegiate Programming Contest) world finalist.

The problem they are solving is this: There are more and more AI Agents, and these Agents need to call various paid APIs to complete tasks. But Agents don't have credit cards, no bank accounts, can't go through the register-bind-card-pay process like humans. The current practice is for developers to pre-charge the Agent or bind their own API key. This works when there are few Agents, but when thousands of Agents need to autonomously call hundreds of paid services, this system can't hold up.

Orthogonal built a unified gateway: Agents access through MCP or SDK, instantly accessing hundreds of paid APIs, pay-per-request, no need to manage API keys, no need to establish billing relationships. API providers list once, can be discovered and called by all Agents. The underlying layer uses Crypto for settlement, supporting the x402 protocol—an on-chain implementation of HTTP 402 Payment Required.

Why is this relevant to the Crypto industry? Because real-time micro-payments between machines is exactly what the traditional financial system does poorly—credit cards have fee thresholds, bank transfers have settlement delays, these frictions tolerable in human transactions become hard constraints in scenarios where Agents call APIs thousands or tens of thousands of times a day. And Crypto's programmability, instant settlement, and permissionless nature naturally fit this scenario.

Notable timeline: YC's RFS (Request for Startups) in Fall 2025 heavily promoted "Infrastructure for Multi-Agent Systems", and half a year later invested in Orthogonal. Early supporters include Precip (W24), Riveter (F24), Andi (W22), Fiber AI (S23) and a group of YC alumni companies building Agent products, indicating this demand is not theoretical speculation, but real.

There's an interesting intersection here: In the recent viral article by Juzi, it said "Agents are the new owners of software", SaaS needs to go from 2B, 2C to 2A (to Agent). If this judgment holds, then payment between Agents is a fundamental infrastructure problem that must be solved—and Orthogonal is betting on Crypto to solve it.

Forum—Turning "Attention" into a Tradable Asset

This project has the greatest imagination, and also the greatest risk.

Founder Owen Botkin previously did long-short equity trading at Balyasny (one of the world's top hedge funds). Joseph Thomas was an engineer at NASA and DreamwaveAI. The YC Partner assigned to this project is Jared Friedman—a core partner at YC.

Forum aims to build the "first regulated attention exchange". Specifically: Build indices from data from search engines, social media, streaming platforms to quantify the "level of attention" a topic, brand, cultural phenomenon receives, then let users go long or short on the change in this attention.

For example: If you judge that a brand is about to lose public attention due to a PR crisis, you can short its attention index. If you judge that a cultural phenomenon is heating up quickly, you can go long.

Their core argument: Attention is the primary driver of commercial success in the digital age; advertising, traffic, user growth are ultimately monetization of attention. But attention itself has never been directly priced and traded.

This project currently isn't labeled Crypto/Web3, but the form of "regulated exchange" plus "creating a new asset class" will likely involve tokenization. YC's RFS in Spring 2026首次出现 "new financial primitives", and Forum正好踩在这个方向上.

For the Crypto industry, the direction represented by Forum is much farther than stablecoin payments—if the object of tokenization is no longer JPEGs, no longer real estate shares, but something previously unquantifiable like "attention", then this is a completely different story. Of course, whether it can work is still too early to tell.

Changes in RFS

Besides looking at what YC invested in, it's also worth looking at what YC publicly says it wants to invest in.

YC releases an RFS (Request for Startups) every quarter, equivalent to an official选题指南. I梳理了 the Crypto-related content from the last three issues:

Summer 2025: 14 directions, Crypto wasn't mentioned once. Even "AI for Personal Finance", which discusses investment and tax optimization, didn't mention Crypto at all. YC's attention was completely occupied by AI.

Fall 2025: Still no dedicated Crypto section, but two directions埋下伏笔—"AI-Native Hedge Funds" (digital asset markets are 24/7, data is open, naturally suitable for AI量化), and "Infrastructure for Multi-Agent Systems" (this is exactly the scenario Orthogonal later切入).

Spring 2026: The change came. Daivik Goel specifically wrote a section "Stablecoin Financial Services", directly naming the GENIUS Act and CLARITY Act, these two US stablecoin bills, saying stablecoins are in a regulatory middle ground between DeFi and TradFi. The exact words: "The regulatory window is open. The rails are being laid."

The overall introduction of the同期 RFS also首次出现 "new financial primitives",并列 with AI-native workflows and modernized industrial systems.

This is the first time in nearly two years that YC has单独开题 for a Crypto-related direction in the RFS. The wording is also very specific—not saying "blockchain" or "Web3", but precisely "stablecoin financial services", and giving specific directions: yield-bearing accounts, tokenized real-world assets, cross-border payment infrastructure.

How I See It

As someone active in both the Crypto and AI sectors, I think this set of data is actually good news for us Crypto practitioners—just the way the good news comes might be different from what many expected.

YC hasn't abandoned Crypto, but YC has redefined what kind of Crypto companies are worth investing in.

To summarize in one sentence: YC is no longer investing in Crypto, YC is investing in companies that use Crypto.

What's the difference? The former's value proposition is "I am building the Crypto ecosystem", the latter's value proposition is "I am solving a real problem, and Crypto happens to be the most suitable tool".

The former's users need to understand what wallets, Gas fees, on-chain interactions are. The latter's users don't even know they are using Crypto—SpotPay's users think they are using a banking App, Unifold's customers think they are integrating a payment SDK, Orthogonal's Agents don't even have the concept of "thinking".

What does this mean for us?

First, the good news: The stablecoin payment track has moved from insider consensus to mainstream Silicon Valley consensus. YC单独开题 in RFS, advancement of the GENIUS Act and CLARITY Act, Stripe acquiring Bridge—these signals together indicate that the compliance path for stablecoins is opening up. For teams that have been深耕 in this track, the financing environment and market recognition are improving.

Second, new opportunities: Agent payment is a demand growing from within the AI industry, Crypto practitioners have a natural advantage to catch it. Real-time micro-payments between machines, programmable money, permissionless settlement—these things we've been talking about for years suddenly have the most specific application scenario in the Agent economy. This isn't us finding a scenario, the scenario found us.

Of course, there is also a reality to face: The profile of competitors has changed. SpotPay's CTO was the 4th engineer at Brex, Orthogonal's founders are from Coinbase and Google—these people aren't Crypto native, but they are entering with the engineering capabilities and product methodology of traditional tech companies. For us in the Crypto industry to compete with them, understanding the chain alone is not enough, we also need to补上 product experience and engineering lessons.

Also, directions like L1/L2, DeFi protocols, NFT, DAO tools—not to say they have no value, but in the view of mainstream Silicon Valley accelerators and VCs, they are indeed no longer on the priority list. This doesn't mean these directions are finished, but if you are working on these directions, your financing strategy and narrative方式可能需要调整.

Finally, the data line "24→1→5", I think the most accurate interpretation is not "Crypto is recovering", nor "Crypto is declining", but: Crypto is being redefined.

YC spent two years figuring out one thing—the greatest value of Crypto may not be to become an independent industry, but to become the infrastructure for other industries. Whether this judgment is correct还需要时间验证. But as someone in both sectors, I think there are大量 opportunities for Crypto practitioners here—provided we are willing to look at ourselves from a different angle.

Crypto doesn't need to disappear, but the best products of Crypto might be those where users don't feel the presence of Crypto.

This is not a compromise, this might be the greatest victory.

You may disagree with this judgment, but this is the position expressed with real money by the most influential startup accelerator in Silicon Valley today.

Data sources: YC Directory (Crypto/Web3 tag, All batches total 177 companies), YC Winter 2026 Launch List (149 companies), YC Request for Startups (Summer 2025 / Fall 2025 / Spring 2026 three issues). Detailed information on the 5 Crypto-related projects comes from the YC official website and various companies' public materials.

Author: aiwatch, Crypto industry 6+ years, deeply involved in AI sector for the past two years, based in Silicon Valley, focused on GenAI product analysis and Crypto×AI crossover research.