An Agent Infra Founder with $25M ARR Says: We're at 6:05 AM on Day One of the Agent Era | A Conversation with Xiaosu Technology's William Du
July 20, 2025
What opportunities has the AI Agent explosion created for upstream and downstream industries?
In this episode, we invited William Du, co-founder and CEO of Xiaosu Technology, to share how his company — an AI infrastructure provider behind Agents — has seized the moment as Agents take off. He tells the story of how Xiaosu Intelligent Search and SkyRouter.ai, their one-stop model aggregation platform purpose-built for Agents, helped push the company's total ARR past $25 million.
As one of the "shovel sellers" in the Agent space, William described the current moment to us as "6:05 a.m. on day one of the Agent era" — the sun has just risen, with a long road ahead and enormous room for growth. Beyond Agents, William also shared his career choices, his transition from secondary market investor to founder, and key lessons learned along the way. We hope this helps listeners better understand Agents and the AI infrastructure field behind them.
PS: Xiaosu Technology has prepared a gift for "Crossroads" listeners. Register using this link wj.qq.com to receive 2,000 free intelligent search API calls or an equivalent professional services package.
Part One: The Infrastructure Provider Behind AI Agents — How to Reach $25 Million ARR
02:02 — Crossroads vs. Xiaosu Technology William: rapid-fire Q&A
03:08 — From aerospace engineering to finance to entrepreneurship: each career move was fairly "cross-disciplinary"
05:21 — What exactly does Xiaosu Technology do? AI intelligent search + SkyRouter.ai, a one-stop large model aggregation platform for Agents
06:39 — As more people search through Agents, search behavior becomes machine-to-machine
08:48 — Serving Agent clients revealed their need for one-stop infrastructure services
10:19 — Building intelligent search: three phases. Phase one: full-web full-text reading paired with search APIs
11:46 — Phase two: anticipating industry developments early, building a self-developed alternative to the Microsoft Bing Search API
13:20 — Phase three: after the Agent explosion, search capabilities plus multilingual and customization capabilities became essential
15:57 — Seizing timing through self-development plus M&A: search talent is limited
19:10 — For B2B companies, improving through customer feedback earlier is itself a competitive moat
20:25 — Doing homework in advance to anticipate Microsoft's Bing Search API adjustments
24:42 — The old search competitive landscape: 2-3% of global daily search queries happened in the B2B segment, with Microsoft Bing dominant
26:06 — Consumer-facing search giants also offer search API services, but this isn't their core business
29:52 — Optimizing products and services for Agent needs: what search requirements do Agents have that people wouldn't expect?
30:23 — Most Agents are going global from day one today; providing multilingual support and data compliance is critical
33:13 — "Today is 6:05 a.m. on the first day of the Agent era"
34:20 — Many Agent clients spend their time on go-to-market; they outsource infrastructure to suppliers who can provide one-stop services
37:13 — Two convictions about building AI Agents: first, Chinese and U.S. teams will become the main players in the AI industry, and Chinese founders going global will have hard requirements
38:01 — Second: starting from 2025, inference demand for compute will significantly exceed training demand — hence SkyRouter.ai
39:48 — SkyRouter.ai's three competitive advantages: sufficient global nodes and distributed compute resources; platform stability; resource operations capability
41:18 — Migration to SkyRouter.ai reduces calling costs by 10-15%: how exactly is this money saved?
Part Two: AI Infrastructure Industry Analysis
43:49 — From the post-ChatGPT AI infrastructure startup boom to the infrastructure services needed after the Agent explosion: what changed these two years? From scrambling for GPUs to scrambling for data, from IaaS compute resource competition to Agent vendors needing full-stack services from IaaS to PaaS
46:53 — The biggest challenge right now isn't insufficient infrastructure capability, but that the Agent industry is still searching for product-market fit
47:39 — Who has found PMF? Coding, office software — reference scenarios are AI assisting humans to search and complete work simultaneously
49:17 — Beyond office scenarios, some vertical use cases: AI for ad placement, AI for travel itinerary planning, AI for information acquisition, etc.
50:27 — Near-term AI infrastructure: what matters most is whether AI-native applications can achieve qualitative leaps in user scale and retention
52:09 — Long-term AI infrastructure: the previous logic will continue — compute will reach a baseline, and data will become more important
54:01 — "Leading model companies will definitely build their own search, but they won't build all of it"
55:06 — "The AI Agent ecosystem will become personalized and richly varied; leading model companies won't capture a large share of Agents"
Part Three: Personal Career Choices and Entrepreneurship Lessons
56:08 — From secondary market investor to diving into the deep end of entrepreneurship: investing is allocating resources; entrepreneurship is creating new productive forces and new resources
57:21 — From starting a business in senior year of college to investing at Hillhouse
01:00:13 — Most investment decisions are short-term; the uncertainty of entrepreneurship is not knowing what results you'll get after doing the work
01:02:39 — "After starting a company, I realize how shallow I was as an investor. I am now filled with respect for the business itself"
01:04:44 — In secondary market investing, the consequence of being wrong is stopping losses; strategic misjudgments in entrepreneurship carry massive consequences
01:06:42 — "Think thrice, think more, try first, then go all in"
01:09:51 — Anxiety is useless. The beauty of B2B is: if you don't win a client this time, go back and improve enough, and you'll have another shot next time
01:12:16 — What I wish I'd known when starting out: 1) strategic choices matter enormously; 2) business is hard, comprising many factors; 3) maintain an open mindset
Subscribe to the "Crossroads" podcast 🚦 We track the industry transformations and new entrepreneurial opportunities brought by the new wave of AI technology.
🚦 Crossroads is Steve Jobs's metaphor for Apple — standing at the intersection of technology and liberal arts, where great products are born. AI is transforming every industry. We seek out, interview, and bring together a new generation of AI founders and active participants in the AI era. Together with them, we explore and embrace the changes and new possibilities.
👦🏻 Host Koji: I co-founded Jiepang / The Fair / Tangdao, and started AI Hacker House, a community space for a new generation of AI entrepreneurs. I believe technology, especially AI, represents the greatest value-creation opportunity of our generation. Feel free to reach out to chat, bounce ideas, and connect with what's next. Koji on Jike | Koji's website
👧🏻 Host Ronghui: I've worked at a dollar-denominated VC and spent five years as a Silicon Valley correspondent, following tech development and business stories. Feel free to reach out and exchange ideas. Ronghui on Jike
🎄 This podcast is supported by The Fair's Sound Forest Podcast Initiative.