How did Funes secure investment from YZi Labs?
TechFlow Selected TechFlow Selected
How did Funes secure investment from YZi Labs?
Founder's personal account.
Navigating Web3 tides with focused insightsText by: Hanyang MasterPa
Woke up to news that a nuclear power plant cooling tower in Tennessee had been blown up. The cover image looked familiar, and when I clicked in, I realized I’d photographed it last year. It was a rainy morning; Chongqing and I drove over an hour from Nashville to get there. Aside from deer and wild chickens, the road was empty—until we parked at the entrance and were suddenly surrounded by a group of burly men. Turns out the site next door is a private prison, its front gate facing the nuclear plant. They'd been watching us since we arrived, our visit adding a splash of excitement to their dull morning.
The cooling tower, now gone forever
Like many North American stereotypes, the guys relaxed immediately upon seeing two Asians. After warning us not to fly drones—though it didn’t matter much—they rushed off toward another alarm. So this abandoned nuclear plant and its cooling tower became one of the few structures we’ve only photographed, not modeled. Not that modeling would’ve been easy—the scale was simply too vast for small drones’ remote range.
But now, no model will ever exist. Its massive, even sublime volume couldn’t prevent it from collapsing like any ordinary house near my hometown. The inmates and guards inside the prison must have enjoyed quite a show.
Humans can't achieve eternity, but we always try to resist time. That’s why Funes exists—and why we pour our energy into it.
What is Funes?
It feels like I’ve never properly introduced Funes here. So before telling our fundraising story, let me explain what Funes is.
In 2012, extremists captured Timbuktu and destroyed historical sites. In 2014, Shangri-La suffered a devastating fire. That same year, the ancient city of Aleppo became a battlefield in Syria’s war, suffering severe damage. Nearby, the ruins of Palmyra were flattened by ISIS in 2015. In 2019, Notre-Dame de Paris burned. As global instability and climate disasters intensify, world cultural heritage grows increasingly fragile and endangered. Every corner of the planet could be facing its final day. Our physical world urgently needs digital preservation and documentation.
Yet despite this, the physical forms of a nation’s landmarks, ancestral homes passed down generations, or buildings you pass daily—are themselves information, carriers of knowledge. But our understanding of them is far more limited than you might imagine. From grand architecture, monuments, and ruins, to smaller artifacts, toys, and crafts—these creations are traces and witnesses of humanity on Earth. They’ve almost never been digitized at scale for universal purposes like collection, protection, and display, then uploaded to the internet for public access and use.
That’s why we founded Funes.
Some Funes models already online and maps based on them
Funes is like a GitHub for the physical world. Together with users worldwide, we model and archive all human-made structures. Funes’ database now contains over 1,000 three-dimensional models of human-built structures (not all yet published). Spanning continents and over 4,000 years of history, the collection continues to grow rapidly.
We won’t claim certainty, but within known scope, Funes is likely the world’s largest open-access 3D architectural archive.
Funes now adds 5 to 10 new models daily—some collected directly by our team, others submitted by contributors from places as distant as Singapore and Moldova. These models form an unprecedented digital repository, offering rich research material for computer vision and graphics researchers, filmmakers, and game creators. They also open entirely new research directions in heritage conservation, architectural history, archaeology, and related fields previously unattainable.
When enough models accumulate, unimaginable possibilities naturally emerge—like cross-regional, cross-cultural visual comparisons. With these models, historians can trace the spread of cultural ideas, studying how different civilizations responded to natural conditions and social challenges through built environments. A researcher of ancient trade can now “stand” simultaneously at a frontier pass in northwest China and a caravanserai ruin in Anatolia, meticulously comparing stonework techniques and structural designs. Scholars can use our large-scale map function to “connect dots into lines, lines into surfaces,” revealing trans-regional, trans-historical development trends.
Each 3D model in Funes comes with coordinates and modern addresses, linked to 3D terrain data. Thus, the database becomes a powerful tool for GIS analysis and spatial humanities research, helping urban planners analyze settlement patterns and architectural distribution across varied landscapes. For large archaeological site models, researchers can correlate locations with geographical features, ancient transport routes, and early urban planning, examining how climate and topography influenced architectural evolution—even mapping the regional spread of architectural styles. These high-precision geospatial datasets also support heritage tourism planning and development.
For professional users, Funes offers a suite of research-enhancing features:
The “wireframe mode” allows engineers to see through surface materials and examine structural details of the model’s triangular mesh. Heritage conservators can use it to deeply analyze internal architectural structures.
Wireframe and survey line-drawing mode, Sopoćani Monastery
Unlike typical online model libraries, Funes also provides a professional “orthographic view” feature. Originating in descriptive geometry, orthographic view eliminates perspective distortion—where objects appear smaller the farther they are. For example, in standard photos, the top of a tall tower or cathedral often looks narrower than its base. Orthographic view avoids such deformation entirely, enabling accurate comparison of proportions in large-scale architecture—whether Cologne Cathedral or a Tang Dynasty pagoda—rendered with the precision of Liang Sicheng-style architectural drawings, eliminating both perspective and lens distortions, greatly aiding archaeologists in understanding architectural proportions.
Top and bottom show standard vs. orthographic view from the same angle, Sheqi Shan-Shan Guild Hall in Henan
Built upon “orthographic view,” Funes has pioneered a globally unique “survey line-drawing mode,” trained on extensive architectural drawings and optimized 3D model structures. It automatically generates archaeological-grade architectural plans—including floor plans, elevations, sections, site plans, axonometric diagrams, and orthophotos—meeting the standards of the U.S. Historic American Buildings Survey (HABS), suitable for academic publication and comparative studies. This mode also adopts UNESCO’s standard archival format for heritage documentation, effectively supporting international conservation efforts and World Heritage nomination collaborations.
Overlay of survey line-drawing and 3D model of Balhae Temple's Palsangjeon Hall, Korea’s National Treasure No. 55
Thanks to optimized photogrammetry algorithms, every dimensional structure and geometric relationship in Funes models can be measured and calculated. Mathematical structures and modular designs in Bauhaus, Metabolism, and Brutalist architecture can be precisely extracted using measurement tools, helping researchers explore spatial philosophies—such as the eave curves in East Asian traditional architecture or complex proportional systems in Greek temple columns—with extreme accuracy in 3D space.
These professional modes provide researchers tools previously accessible only through costly field surveys.
Soon, Funes will support 360-degree panoramic viewing—not just exteriors, but interiors too—offering complete spatial immersion. Researchers can “walk into” a 13th-century Gothic cathedral, gaze upward at towering vaults, observe light from stained glass casting patterns on stone pillars; or step into a Chinese courtyard house, experiencing the shifting views created by skywells, corridors, and rock gardens. Complete interior panoramas help restoration engineers pinpoint beam positions and spatial relationships among frescoes—like a digital “Noah’s Ark” for cultural heritage.
Panoramic photo from St. Stephan’s Church of Mainz, where Marc Chagall designed the entire church’s blue stained glass
Documentation accompanying each building model is equally vital. We’re currently using our proprietary AI pipeline to enable machines to understand expert descriptions, academic sources, and visual features of 3D models, generating accurate, accessible encyclopedia-style entries. We’re gradually rolling out real-time AI narration and interactive browsing. This experimental educational tool lets users freely browse, rotate, and zoom into buildings while AI narrates the structural and historical context of visible details—deeply integrating visual and textual information. A demo is already live; check it out on the Selimiye Mosque page.
Funes aims to ensure that even if physical structures are damaged, the spatial memory of civilization endures intact.
Architecture embodies the tangible crystallization of human wisdom across civilizations, eras, and individuals—it should be shared by all humanity, not vanish silently into oblivion. Often, buildings are deemed worthless or obstacles to urban progress and casually demolished. Decades later, someone discovers their extraordinary significance—but by then, only archives or photos remain, and reversing the loss is impossible. Modern people’s ability to appreciate their surroundings is questionable; every space around us holds unimaginable meaning, if only we truly study and attend to the messages they convey.
By lowering modeling barriers and opening more data, Funes drives genuine democratization of cultural heritage: a child from a remote mountain village can now study the details of world-famous buildings just as easily as a student on the U.S. East Coast. Open access to models coexists with diverse model contributions—whether high-precision scans by archaeologists or smartphone photos by local residents—all archived together. This means heritage documentation is no longer the privilege of a few experts, but a shared human endeavor.
The value of a particular place is inherently immeasurable. Its importance can never be assessed purely from a societal standpoint. The site of our first date, the grass where we walked our pet, the room we once called home now gone, the restaurant filled with memories of a loved one no longer with us, the office where we began our career, the classroom we last sat in—these places may not interest historians, but they hold immense personal value, defining our human existence through experiences, emotions, and memories. Therefore, protecting these seemingly insignificant spaces is equally crucial.
Qitaihe’s筒子楼 (tongzilou), residential buildings that are childhood memories for me—and countless others
Funes’ ultimate ideal is to connect the digital realm with the physical world—not just meticulously studied, curated “significant spaces,” but everyone’s personal spaces, every ordinary yet intimate moment of life.
About Funes’ Fundraising
Since starting Funes, three questions come up in nearly every conversation:
-
How do you model? Feel free to click here to learn more.
-
How do you make money?
-
Who invested?
These questions follow a strict sequence. I’ll write a separate article about the second question later. This time, I’ll focus on the third: who invested in this round, and how did we talk? Fundraising itself is a long, tense process that gives founders stomachaches—so writing about it isn’t particularly exciting. For readability, I’ll focus on key moments.
In today’s market environment, I feel there are mainly two types of early-stage startups: those locked in by investors before even launching, and those unable to raise at all. Companies that successfully fundraise after a 4–6 month effort, like a few years ago, are now statistical outliers. When we started preparing for fundraising last June, we braced ourselves—we’d be lucky to close by December. So we aimed to build as many models and gather as much data and results as possible while spending minimally.
Some think fundraising is a single moment—you find an investor, take the money, and it’s over. But it’s actually a process: initial contact, formal review, term negotiation, disbursement… it takes time. This means founders can’t just “save up something big,” work intensely for months, then present spectacular results. It doesn’t work that way. Because fundraising is a process, founders must keep the company progressing throughout, so investors confirm their judgment remains sound.
So my life split into two parts: going out to model and sharing with users, and meeting investors. These couldn’t be done together—investors are often in drone-restricted zones.
After deciding to build Funes, I visited Professor Liu Feng in Xiamen. His only advice: meet people quickly, because the market could turn anytime. He discouraged me from preparing first before approaching investors, urging me instead to start conversations early to understand investor perspectives sooner. At the time, I was traveling between Shanghai and Hong Kong; HG’s KK and Walter, with strong networks, introduced me to my first batch of investor meetings and helped refine my pitch. That’s how fundraising began.
Liu’s prediction proved spot-on. By September, I’d accumulated nearly four months of rejection. Not a big deal—rejection usually reflects differing perspectives on your company, and sometimes helps founders grow. Side note: sadly, most investors don’t ask questions that make founders go “wow.” When one does, I’m deeply grateful. The first major breakthrough came in Silicon Valley. Due to the pandemic, I hadn’t been to SV in three years. Chaoge connected me with several SV friends. One U.S. investor rejected me but spent an hour analyzing current SV fundraising challenges. Without him, I’d have wasted a lot of time.
By September, with many events, I came to Singapore. It was unbearably hot—felt like an endless summer. One afternoon, I was modeling CHIJMES Hall. Since MegaETH was hosting an event there, Shuyao asked if I could showcase Funes. I said I’d model the venue first. Right after finishing, Siyuan from ABCDE messaged: “Want to meet our founder, Mr. Du?”
I doubted ABCDE would invest—too far, too hot. I was wearing my summer modeling outfit: quick-dry shirt and shorts, drenched in sweat—not ideal for meeting someone. I considered declining. But Siyuan gave an irresistible reason: ABCDE’s rooftop offers a view of CHIJMES Hall’s roof—a no-fly zone I couldn’t photograph.
He took me straight to meet Mr. Du. When I said we’re like a Wikipedia for the 3D world, he shared how he loved Wikipedia as a kid—clicking from one entry to another, imagining future travels. He’d spend hours reading, though back then Chinese content was scarce, requiring self-translation and half-understanding. Then he said: Funes’ philosophy resonates with Wikipedia—he completely understood why we wanted to build this.
Then he said: “Let’s not talk investment yet” (my heart tightened), “but personally, I’ll donate $100,000 right now” (my heart jumped).
We then discussed his experience co-founding Discuz!—I’ll record that episode with him later. Finally, he said: Funes deserves investment, but the process takes time. So he’d donate first, then discuss investment. The funds would help us move faster. As I left, he added: he genuinely loves Wikipedia, but for future investor meetings, I should avoid comparing Funes to Wikipedia—everyone claims to love Wikipedia, but no investor wants to fund it.
I thought “donate first, invest later” was just polite talk, but I was still deeply grateful for the $100,000—Funes’ first funding. The next day, investment discussions began smoothly. However, ABCDE has since ceased operations under that brand, rebranded under a new name. Funes may have been ABCDE’s final investment.
On leaving, Siyuan kept his promise, taking me to the window overlooking CHIJMES Hall—yes, visible, but too small to photograph 😂.
Special thanks to Shuyao for tremendous help—much appreciated! MegaETH is also an outstanding project; I like everyone involved (even considered recruiting some). From the start, Siyuan introduced us to MegaETH, and we’ve met many supportive friends there. Finding entrepreneurial peers who help each other is crucial—often overlooked. Fundraising matters, but so does who walks that path with you.
At a MegaETH event—back when I still had yellow hair (actually silver fading)
The day after meeting Mr. Du, Dragonfly’s GM messaged: Could I stay another day in Singapore to meet Brother Bo (Mr. Feng Bo)? They’d cover the change fee. I was chatting with Mable, who said: “If you can meet Brother Bo, go now!” The revised ticket was cheaper, and Ctrip refunded over a hundred RMB.
I met him at his home. The moment I entered, I was drawn to an Ernst Haas photograph.
Since this newsletter has many subscribers, some may not know my background. To clarify: my day job is essentially photography. Ernst Haas was an early pioneer of color photography. I’d seen that photo in teaching slides before—assumed it was small, but the original was surprisingly large.
Unlike typical fundraising, my first question to Brother Bo was: “That’s an Ernst Haas piece, right?” He probably didn’t expect anyone to recognize it, and we ended up discussing photography. That day’s conversation mixed Funes, large-format cameras, Polymarket, Mamiya 7, Protra 400, interests in the future, and the meaning of life—topics wildly unrelated. Despite years of fundraising, I’d never had such a wide-ranging discussion.
GM has overseen Funes closely and sincerely. But we posed a major challenge to Dragonfly: we were asking for too little. To explain: for large funds, investing in small projects still demands significant effort and post-investment management. They calculate how many deals their team can handle per fund cycle, setting minimum investment thresholds.
For founders, raising more money isn’t always better. More capital either dilutes ownership or inflates valuation. Frankly, I believe pre-IPO valuations are just paper—often feeding founder ego. Being a founder of a $100M company holds zero appeal for me. We needed to fundraise at fair terms without excessive dilution.
But Brother Bo didn’t pressure me. He offered: whatever amount remained unfunded in this round, he’d personally cover.
So I secured both my first and final investor. Mable, who urged me to meet Brother Bo, is also a podcast host—Chongqing and I first publicly introduced Funes on her show.
I’d long wanted to meet York from Generative Ventures offline, but timing never worked—this consumed 80% of fundraising. Even with remote work, investors prefer face-to-face meetings. Eventually, York suggested: “Why not meet with my partner Will first?” But again, schedules clashed. Finally, he came to Beijing, I went to Shanghai—we met at a KFC in Hongqiao Airport. I brought a 3D-printed plastic model of Liaozhongjing Daming Pagoda and used it to explain. He sensed it was my gift—but no, we hadn’t printed extras; I needed it to show everyone. Yet, the investment was sealed right there in KFC.
The white pagoda I brought—I’ve since gifted all three in the photo to friends; left two are plastic 3D prints, right is stainless steel 3D print + polished
Later, he introduced me to BAI’s founder Anna. Before meeting, I asked Xiaowan, founder of Wan Dian, if she knew Anna. She said Anna’s great fun and talkative—exactly as it turned out. So listeners, if you see this, please join or recommend your portfolio companies to appear on our podcast “Mayfly World”!
Back to that plastic pagoda. After the airport, I went to HashKey’s office. I’d met Jeffrey from HashKey at a dinner hosted by Li Yang in Chiang Mai. I didn’t know he was an investor, but after returning home, he reached out to chat. We didn’t discuss the project much—instead, talked extensively about Liao pagodas. So this time in Shanghai, I brought the plastic version. Thus, the HashKey deal unfolded amid discussions about Liao pagodas.
Of course, the fastest close was during a dinner at our investor Michael Jin’s home, where Owen committed funding in one meal. Well, maybe fundraising really does require more meals. Our collaboration with D11 was also negotiated at a Dubai restaurant that seemed geared toward tourists.
Also thanks to a friend Jarseed—for helping me and several partners grasp industry knowledge, whether at Jing-A or online.
No more rambling—some investors aren’t mentioned here, my apologies. Many of you likely read this because you saw news of YZi investing in us. So finally, let’s talk about that.
I actually spoke with YZi very early. Just days after meeting Mr. Du, I met Dana from YZi Labs. Back then, it wasn’t renamed yet—still Binance Lab. I wasn’t optimistic beforehand—Funes didn’t seem like a typical Lab investment. But after a two-hour chat with Dana, I felt hopeful. Dana was more curious about our founding team. We shared many stories from modeling trips. Of course, we also discussed why I once wanted to write about Binance.
If you’ve read this far, you might realize: investors who might back you rarely spend the entire first meeting solely on your project. Those inclined to invest tend to be broadly interested in you as a person. As for the project, every detail gets clarified over the long fundraising journey. So from Dana’s questions, I sensed momentum.
Then two things happened: Binance Lab rebranded to YZi, and Siyuan left ABCDE to join YZi. Internal changes took time, and I hesitated to push—Siyuan had led ABCDE’s investment in us, creating a conflict. Throughout the rest of fundraising, except as friends, Siyuan deliberately stayed out of the process.
But Dana later reached out, suggesting a meetup in Hong Kong. Coincidentally, KeDa, Chongqing, and I were all heading there—we brought our newly printed booklet. It contained photos taken during our Funes journey. So the six of us—us three, plus Dana, Ella, Siyuan, and Nicola—sat across the largest café table, beginning with that booklet.
In the distance: Liaozhongjing Daming Pagoda, Funes’ first photographed and published model
As a founder, I enjoy conversations with exceptional investors. Good investors never rush to prove themselves right. Meeting enough investors, you realize many just want to demonstrate their intelligence in front of you. So fundraising with good investors isn’t Q&A—it’s dialogue. Founders ask questions too. Our main one was:
What does YZi envision Funes to be?
Many investors suggested: since Funes relies on photo-based modeling, could it become a 3D version of Instagram? There’s a huge logical gap: should a startup become the founder’s true vision, or agilely morph into “any big company”? Many founders self-deceive here, saying our company will become XXX—where XXX can be any big name. For example, claiming Funes could become the next-generation Instagram. When I say that, do I mean Instagram’s model is worth learning, or merely that Instagram is a successful giant we can associate with? That balance lies within the founder’s conscience. For us, Funes has a clear identity in our minds—becoming “any big company” isn’t the goal.
What mattered most was Dana agreeing that when founders discuss visions with investors, calmly rejecting what they’re not is essential. Dana casually gave an example: GitHub was like that—its founders built the world they believed should exist, not chasing markets. GitHub’s creators believed code should be organized and shared in specific ways—they pursued that vision, not victory over rivals for monopoly. Founders and investors alike must know what not to do.
Dana paused, then said: Funes is like a GitHub for the Physical World. Hearing that, I realized it was the best metaphor for Funes I’d ever heard—though we didn’t start Funes to become GitHub, it’s an excellent, understandable analogy. I now use it widely, but credit isn’t mine (wish it were haha).
First photo in the booklet; this booklet exists only in print, never fully published online
Later, Ella messaged: though she had lots to carry leaving Hong Kong, she took our booklet home. That made me happy—these images are the path we’ve walked.
Next came the chat with CZ. KeDa and I video-called him from a hotel in Huludao. I worried about poor internet, tested repeatedly. Joked this might be the most important video call since the hotel was built.
During the formal chat, CZ said he’d known me for years, knew I was idealistic and serious—no scammer looking to cash out. So he wanted to meet the rest of the team. Mostly, KeDa spoke with him—about KeDa’s past projects, what architectural archaeology is, which countries’ heritage most needs protection... If you don’t know KeDa, I strongly recommend this podcast episode to experience his charm. Honestly, if history remembers me a century from now, it’ll be for having driven KeDa around. Anyway, CZ and KeDa hit it off. Then we met “Sister One.” Before the meeting, I realized my mistake as a founder: she was one of the earliest Funes followers—top ten. Even before many in our own company followed, she already did. Sadly, I forgot to ask how she first discovered Funes 😂. Next time, I’ll ask.
One extra note: founders enjoy talking to fellow founders, meaning fundraising flows easier when the lead investor is a former founder. Many contexts need no explanation.
Reading this, it might seem opposite to my earlier tone—wasn’t fundraising supposed to be hard? Seems like casual chats. Yes—because otherwise it’d be boring. A strictly chronological account would be a long, dull log.
This article differs from my usual ones, so I’m unsure who’ll read this far. But whoever does, it’s fate. Maybe you’re a founder like me. Beyond dramatic moments, I want to share what else we did. These snapshots represent just 1% of our fundraising. The rest was mundane, tedious, boring work. But founders live this reality—we don’t survive on that 1% of highlights; the remaining 99% is the essence of our existence.
Our own actions during this fundraising mattered greatly.
-
Funes’ MVP evolved from a concept webpage into reality. We built a full cloud-based modeling and front-end rendering pipeline from scratch. Given buildings’ unique properties, we developed a custom data management system. Each step was visible to investors. We ensured every subsequent meeting showed product progress.
-
We photographed many models. The team never stopped, modeling everywhere possible. At its core, Funes is manual labor. Why would someone stand under the sun, sweating, photographing an unnoticed building? Investors funded us by understanding exactly that.
-
We have users. We hold weekly model-sharing sessions, each larger and better attended than the last. Many attendees go on to photograph their own buildings afterward. Ultimately, a project doesn’t survive on funding, but on users. I’m deeply grateful to all friends contributing to Funes.
-
Luck played a role. Many pivotal moments weren’t due to effort—luck was involved. How to view effort and luck? We partners love this quote:
A warrior, before his duel, prays to God—not for victory, but for courage and resolve.
Beyond fundraising, there are other costs: I likely won’t write new Binance series articles soon. That series is for media, posing conflicts of interest. But if possible, I hope to interview CZ or Sister One on my personal channel “Mayfly World.” Also, long ago, I interviewed Xiao Feng from HashKey. That interview, for various reasons, wasn’t published. Now, with our partnership, it’s unsuitable for media. But it’s a great interview—I’ll see if I can publish it on my newsletter.
Words fall short, meanings run deep. Funes’ fundraising received help from many friends. I can’t name everyone, but thank you all.
Hanyang walking toward Nan Ta of Liao Shangjing
Finally, Why Funes?
Modeling the entire world—why such obsession?
Because we may be the last generation living fully in the physical world. As technology advances, the cyber world will grow ever more central—more important than the physical. Humanity’s grandest, most magnificent construction era in the physical world is history. We may be the last generation dedicating time, money, ambition, and lives to building the physical world. For the next generation, the cyber world will be the ideal to strive for.
So it’s time to leave a final portrait of this material, spatial world. We’re making a death mask in plaster for a world destined to stiffen and decay. Before it dissolves into points, triangles, UVs, projections, pipelines, and parameters in cyberspace, we preserve one final definitive record.
Humanity’s ten-thousand-year effort reshaping Earth’s surface with earth, wood, stone, brick, glass curtain walls, and reinforced concrete—was but a dream. Now the dream ends. We take notes on the dream, sketching side profiles of yesterday’s illusions.
You could say this project is the final museum—one that can never be completed. Eventually, every physical object that ever existed, every museum housing such objects, every museum of museums—will become part of this ultimate museum’s collection and subset.
That’s why it’s called Funes. It comes from Borges’ short story “Funes the Memorious”—a tale encompassing all tales, just as our plan is a museum collecting all museums.
In building this endless, eternal, perhaps futile museum, everything will emerge.
Join TechFlow official community to stay tuned
Telegram:https://t.me/TechFlowDaily
X (Twitter):https://x.com/TechFlowPost
X (Twitter) EN:https://x.com/BlockFlow_News
