Breaking Into Silicon Valley, David Wei Blazes a New Trail for Dollar Funds | Getting in the Game

"Entering the Game" is a recurring column from Dark Tide Waves (暗涌Waves). It grew out of our observation that once-reliable operating models are facing new challenges, and the industry rules long inherited from West to East have been shattered. People are hungry for new maps and new orders of innovation and capital. "Entering the game" is the most precious posture of all. "Entering the Game" was born amid transformation. To sum up this column's subject in one sentence: we hope to find the new players and new playbooks better adapted to a changing environment. Over the past six months, "Entering the Game" has identified the market's most low-profile healthcare dark horse fund — Longpan Capital, Chaoxi Capital, "heterodox" yet "perfectly timed", and the unconventional and aggressive FA firm V Creation. Against the backdrop of exits becoming the dominant theme in the primary market, we also explored the new RBF investment model through our coverage of new players like Drip Capital and Yongchuan Capital. In 2024, "Entering the Game" will continue — we look forward to more "players entering the game" amid transformation. The following is this column's fifth article.

By Zhiyan Chen

Edited by Jing Liu

Is the world growing more connected or more fragmented? Faced with this question, most would struggle to give an optimistic answer. People have grown accustomed to two coexisting narratives in the venture capital market: one side talks about AI large models, cross-border e-commerce, and Middle Eastern capital; the other's keywords are early-stage and small-ticket investing, advanced manufacturing, and government-guided fund reciprocity. At this moment, former Alibaba CEO David Wei's JiaYu Capital has chosen to back the launch of a dollar fund rooted in Silicon Valley: ENVISIONX Capital. The fund has completed its first close, with major LPs including multiple entrepreneurs and family offices. Beyond locking in AI — Silicon Valley's absolute hottest track — a core keyword for ENVISIONX is "globalization." "On one hand, we provide capital to support innovations we believe in; on the other, through our partners' networks and industry experience, we help non-US companies expand in the US and connect non-US founders with mainstream American universities and investment institutions," Hong Zheng, ENVISIONX's founding managing partner, told Dark Tide Waves. ENVISIONX carries echoes of JiaYu Capital. For instance, it emphasizes the team's corporate management backgrounds and hands-on investment experience. Zheng previously served as CEO of BlueFocus International, leading operations across 12 countries and directly executing two IPOs on Nasdaq and the NYSE, seven international M&A deals and post-merger integrations. He worked at a Fortune 500 tech company in management roles and has extensive connections in Silicon Valley's tech and VC circles. Among the three partners, Hui Liang is a founding partner of Oriza Ventures, a Silicon Valley fund, with over a decade of experience in institutional and corporate compliance in the US; George Maloney is a native American with 15 years in tech investment and management, including 10 years in China, and speaks fluent Mandarin; partner He Shen was previously CFO of BlueFocus International. Another parallel: the playbook of providing value-added services first, then making investments. In its official introduction, ENVISIONX explicitly lists "helping growth-stage companies mature through the relay of operations and capital" as part of its mission. There are also clear differences. ENVISIONX's all-in posture on AI, for one. More crucially, in structure, ENVISIONX is a 100% American fund. According to public information, Dark Tide Waves found that JiaYu Capital manages nearly $1 billion in dollar funds and has experience raising and managing USD funds. But in recent years, amid shifting regional dynamics and changes in the primary market landscape, raising a new dollar fund focused on China has become genuinely difficult. Wei believes that, on one hand, AI-driven changes will create opportunities large enough and worth capturing, with Silicon Valley as their epicenter; on the other, "under current conditions, it's no longer possible to build an investment institution spanning China and the US — it must be a legally completely independent entity." ENVISIONX is a fresh answer. In JiaYu Capital's office in Shanghai's Kerry Parkside, Wei and Zheng told Dark Tide Waves about ENVISIONX's founding purpose and expectations. To some extent, what this emerging dollar fund aims to do — and its very birth — represent an optimistic answer to where "globalization" is headed. The following is our conversation:

The Second Half of Globalization

Dark Tide: Why establish a dollar fund based in Silicon Valley at this moment?

Wei: In the summer of 2022, I visited Silicon Valley for the first time in three years and formed two convictions: first, geopolitical risks are unavoidable; second, AI will bring massive change. Why set up a fund in the US, and why focus that fund on AI, both stem from these two convictions. Under current conditions, it's no longer possible to build an investment institution spanning China and the US — it must be a legally completely independent entity, and it needs a professional team permanently based in the US.

Dark Tide: Some believe the world is undergoing "deglobalization," and many RMB funds that once considered dual-currency structures have abandoned the idea.

Wei: I often say we can't change the macro environment, but we can create micro environments. The old globalization had one benefit: one global pocket, investing on both sides — that was a hard connection. Now, capital may no longer have legal-based connections, but the value of soft connections between capital will only grow. JiaYu and ENVISIONX will be a soft connection in the future; the exchange of ideas between people cannot be stopped.

Zheng: Some say "deglobalization" or "anti-globalization," others say "slowbalization." We tend to view this more positively — soft connections may be the second half of globalization, and ENVISIONX hopes to serve as one medium within it.

Dark Tide: What's the current state of Chinese entrepreneurs and companies in North American capital markets?

Zheng: I've noticed a phenomenon in Silicon Valley: it's not hard for Chinese entrepreneurs to get their first check, nor is it hard to grow and reach Series D or E and secure major PE money. The difficult hurdle is the Series A and B gap — from connecting with mainstream local markets, securing investment from mainstream North American funds, to local hiring, local customer acquisition, and board governance, they need more help. We want to be that help. I encountered a real case: a Chinese entrepreneur had a competitor doing almost exactly the same thing, with half the revenue and growth rate, but because the competitor was American, the Chinese founder's valuation was only one-sixth. That's the reality. You could say that while ENVISIONX is legally a 100% American fund, our team composition feels familiar enough to Chinese entrepreneurs. Our partners have multiple cross-border IPO experiences, local American backgrounds, and experience engaging with law firms at various levels. These are concrete ways we can help entrepreneurs who want to build global companies.

Dark Tide: But looking at your invested projects, not all are Chinese entrepreneurs.

Zheng: Chinese founders account for about half. We're willing to help promising non-US companies and entrepreneurs develop in the US — they include Chinese, but also Europeans. To add, while ENVISIONX is rooted in Silicon Valley, almost none of our current investments are native Silicon Valley projects. One reason is that Silicon Valley project valuations are generally quite high, and with so many investment institutions there, the chance of finding overlooked deals is slim. Additionally, we hope that after investing, our portfolio companies can enter the Silicon Valley ecosystem with our help — playing on a spatial differential allows us to maximize our value-add.

Dark Tide: Since fundraising began, have prospective LPs bought into ENVISIONX's strategy?

Wei: We've basically not raised externally, because ENVISIONX is a first-time fund, so the vast majority of LPs are relationships JiaYu accumulated globally over the years and who are compliant under US law. Of course, by Fund II, once the strategy and track record are proven in the market, we'll certainly need to raise institutional capital from US sources. What I've heard most from American investors these past two years is: no one can afford to not invest Chinese. If in the mobile internet era, a founding team without Indian members was unviable, then in this AI wave, you won't see a core AI R&D team without Chinese members. Guaranteed. China today has two advantages: one is Chinese experience, especially in commercialization; the other is the core R&D teams composed with Chinese elite participation.

Dark Tide: Some have called 2023 the "first year of Chinese enterprise globalization." How much did last year's wave of Chinese companies going global influence ENVISIONX's founding?

Wei: JiaYu divides going global into two parts: hardware going global and software going global. Honestly, hardware going global is what JiaYu does best and has long focused on. Starting with Anker Innovations, we've produced 11 world champions, including Ziel Home, Merece, Cupshe, EcoFlow, Hicon, and Ziel Technology. Hardware going global is fundamentally about selling products — it doesn't necessarily require an offshore dollar fund to help. But software going global involves core data, and may need a more compliant, offshore-structured investor. That's a major backdrop for creating ENVISIONX.

Dark Tide: Can the JiaYu-ENVISIONX model be replicated by other institutions?

Wei: We very much hope it can be replicated, because having only one player in the world isn't a good thing. And we believe it can be replicated. When I communicated with domestic investment institutions this year, many GPs were quite interested in our approach. I often joke that new district development is easier than old city renovation. The JiaYu-ENVISIONX model is new district development — we have no historical baggage.

The All-In AI Methodology

Dark Tide: From an industry perspective, AI isn't in the core investment direction of mainstream RMB funds. Why did JiaYu support an AI-focused institution?

Wei: RMB funds should of course invest in AI too. I compare AI infrastructure and AI large models plus various vertical models to roads and the cars running on them. When cars change, roads must change too. We judge that both AI-related hardware and software will undergo major changes. So Chinese teams investing RMB in hardware-leaning AI new infrastructure, while US opportunities lie in application-layer software.

Dark Tide: Specifically, how will ENVISIONX invest in AI?

Zheng: A major piece is AI new infrastructure, plus the application layer. We'll focus on three AI-native application industries: healthcare, financial services, and gaming.

Dark Tide: Why these three industries?

Zheng: These three industries share one characteristic: they continuously produce proprietary, private-domain data.

LLM large models are based on public data. So any vertical application built on public data — we'd worry about being crushed by large model iterations. Conversely, what large models can't crush are these three industries I just mentioned.

From another angle, looking at globalization's current state, these three industries are also ones that can become large enough within a single market. ENVISIONX's invested healthcare AI project achieved a strategic exit through M&A in less than a year.

Dark Tide: How do you choose stages?

Zheng: Must have product and revenue — pre-product and pre-revenue, we don't touch. If the product isn't built, we won't invest even at Series C; if product and revenue are there, we'll invest even at angel. We don't fight for the first round, and we don't pepper our shots.

Like JiaYu, ENVISIONX also practices services first, investment second. By introducing customer prospects to companies we're evaluating, we help ourselves test and validate our judgment.

Dark Tide: AI investing is extremely competitive — doesn't the services-first approach risk missing the best timing?

Zheng: In communicating with many corporate executives, I've found that companies today don't actually use just one supplier. So investors need to be careful — some startups will present a dazzling array of logos as their customers, but we need to carefully examine whether these customers have real stickiness and what their actual share is. These things require a certain amount of time and stage to verify.

So is it possible to miss out? Absolutely. But we're not positioned at seed stage, so we'd rather miss than invest wrong.

Wei: And it's not too much to worry about. Generally, after startups sign customers through our introductions, they'll often create a special add-on investment opportunity for us, and also understand our value, with some valuation balance.

Dark Tide: So specifically in AI, JiaYu and ENVISIONX have a complementary relationship?

Wei: I understand AI infrastructure as three forces: transmission force, storage force, and computing force.

China leads in transmission force — for instance, Innolight's soaring stock price last year, and our portfolio company Bosen Optoelectronics, both made rapid advances in large model infrastructure. In transmission force, JiaYu focuses on optical chips, electrical chips, optical modules, including next-generation long-distance transmission between data centers. ENVISIONX focuses more on data flow and distribution software related to transmission force. Once transmission force rises, storage force bottlenecks emerge. So in storage force, JiaYu focuses on hardware chips, while ENVISIONX focuses on compression technology and storage software. In computing force, we don't look at GPUs because large chips mean large investments. JiaYu invested in VirtAI Tech, empowering hardware through software to achieve breakthroughs in computing force.

With the three forces, we also have three uses — usable, well-used, and affordable. China will certainly find AI large models usable; the question is how to use them well and affordably. Because China faces restrictions on high-spec computing chips, this may force out different AI application solutions — not everything needs to be done with the most expensive GPUs. There will be numerous application scenarios and investment opportunities in between.