14 Fragments, 14 Years of Investment Thinking | Linear View

Recently, Harry Wang, founder of Linear Capital, gave a speech titled "The Power of Long-Termism: My 14 Years with Manycore Tech" at the Hangzhou Unicorn Conference, part of the city's annual "All Things Grow" summit. Beyond the speech, he also recorded the debut episode of Linear's video podcast Between the Lines with Manycore Tech's Huang Xiaohuang.

Today, we've pulled 14 excerpts from Harry's reflections — about how an investor makes judgment calls, how he stands by a company through its darkest hours, and how he thinks about long-termism, Hangzhou, and entrepreneurship itself.

If you're walking this path too, we hope these offer something useful.

A little over a month ago, Manycore Tech rang the bell at the Hong Kong Stock Exchange, becoming the first among Hangzhou's "Six Little Dragons" to go public.

Fourteen years ago, "spatial intelligence" wasn't even a concept. Yet Harry wrote Huang Xiaohuang his first angel check. Fourteen years later, that investment has returned roughly a thousand-fold.

The story's beginning may have involved some luck, but it rested more on belief in a person — and unwavering support through the decade that followed.

A young man in flip-flops and a T-shirt, an NVIDIA CUDA scholarship recipient, who sold every share of NVIDIA stock the day he quit to start a company back home, fighting alongside his brothers to realize their dream from another angle.

This, perhaps, is the magic of entrepreneurship and early-stage investing. Some things can't be measured by ROI. But as we believe — if it's worth doing, do it a little longer.

01 Cherish the Connection

In 2009, I was volunteering at a Zhejiang University Silicon Valley alumni event when I met Xiaohuang, who was also volunteering.

During my years studying and working in the US, I never liked holding long-term titles, but I was always happy to ask, "Is there anything I can help with?" I genuinely had no idea at the time that this chance encounter would lead to knowing Xiaohuang — and later, investing in him.

After we met, I referred him for an interview at Facebook. But during the interview, he said he still wanted to work on GPUs.

This was 2009. NVIDIA's market cap was under $10 billion. The term "spatial intelligence" didn't exist.

02 The Most Expensive Whim of My Investing Career

Let me tell you a joke on myself.

Fourteen years ago, I wrote that first angel check to Huang Xiaohuang. Manycore's valuation then was roughly one-thousandth of its market cap at IPO. Around that time, many of my earlier angel investments had cratered, which had shaken my confidence in angel investing altogether.

But I admired Huang Xiaohuang, so I gritted my teeth and cut the check I'd planned to write in half — $500,000 for the angel round instead of the $1 million Chen Hang had hoped for.

So my feelings now are complicated. Half pride for Xiaohuang, half a lesson in early-stage investing — I only earned half of what Manycore had wanted me to earn. This may be one of the most expensive whims of my investing career.

03 Betting on Obsession, Not a Sector

Fourteen years ago, what I saw was a guy in flip-flops and a T-shirt, an NVIDIA scholarship winner, who'd given up NVIDIA stock to return home and start a company.

Incidentally, he sold every share of NVIDIA the day he quit — poverty limited our imagination, I suppose. NVIDIA later climbed to trillions in market cap. Whether that trade was worth it is impossible to calculate now.

I wasn't investing in a sector. I was investing in the obsession this person carried.

04 The Resume Stain Nobody Understood

Xiaohuang once told me a joke.

When he first returned and was fundraising, he'd introduce himself as being "from NVIDIA." VCs didn't respond well. In his own words: "Basically, I felt like in VCs' eyes, it was about as fancy as saying I worked at Foxconn. Definitely not a glamorous resume."

Several VCs directly told him his team's work history wasn't great. He didn't ask for specifics at first. When he did, they said: "First, that company's not so great. Second, you didn't work there very long. Your cofounder doesn't even have work experience."

Xiaohuang quickly revised his resume — just said "worked in Silicon Valley," no company names.

05 Go Heavy on the Right People

Over fourteen years, Manycore pivoted directions, pulled its US IPO at the last minute due to changing conditions, got doors slammed by overseas clients, was sued by competitors... yet Xiaohuang, Chen Hang, and Zhu Hao never gave up once. The three of them stood together to ring the bell at the Hong Kong Stock Exchange on April 17.

If you want to go fast, you can speed alone. If you want to go far, you need brothers and sisters who truly get you.

Founders finding investors, investors finding founders — these aren't one-time transactions. When you find the right person, bet heavy. This was the most expensive lesson I learned from cutting that check in half.

06 Stay on the Right Path

Were there faster paths for Manycore? Yes. Major tech companies came calling multiple times with acquisition offers. They refused. Good call.

Were there cheaper paths? Yes, too. Back then, no SaaS company built its own infrastructure. They insisted on building their own GPU cluster. Looking back, this "dumb decision" was precisely the solid foundation for their spatial intelligence story today.

Over 500 million structured 3D scenes — not bought, but rendered frame by frame over fourteen years. Brutally hard. Genuinely impressive. I want to say to young entrepreneurs: Shortcuts tire you out, even kill you. The right path, walked long enough, steadies you and wins you allies.

07 When Starting Up, You Do Everything

How rough were Manycore's early days?

For the first two years, their GPUs would burn out the moment they hit the servers — temperatures spiking to 110°C, destroying a card every few minutes.

It finally forced Huang Xiaohuang to write his own algorithms to monitor temperatures. A CUDA scholarship recipient, pushed into becoming a thermal engineer. When you're starting a company, you do everything.

08 The 81st Trial

Xiaohuang once told me an investor pointed at his nose and demanded: "You've got a perfectly good SaaS business — why are you doing spatial intelligence?"

Later, when he faced resistance from legacy shareholders around the IPO, I helped mediate. After countless calls, we still landed on a hard-to-swallow solution. I told him, "Brother, I really tried my best. But in situations like this, this might just be your 81st trial."

In the end, Xiaohuang got through it with that survival instinct, that willingness to stake everything — plus the support of the vast majority of reliable investors.

09 The Most Anxious Times Are Often the Most Interesting

In the months after this trial, DeepSeek exploded, and spatial intelligence became capital's darling overnight. In Xiaohuang's own words, it happened over one Spring Festival.

He once said something I've never forgotten: "The most anxious period was actually the most interesting."

So to my founder friends: If entrepreneurship exhausts you and makes you want to lie flat, you may be headed the wrong direction. If it feels brutally hard but you want to charge forward every day, that's the right place.

10 Why Rush to Finish a Fun Game?

One thing Linear has always believed — long-termism.

Put simply: if it's worth doing, do it longer. Anything truly meaningful compounds with every additional year you stay with it.

And the dopamine on this path comes in installments, not a lump sum at the finish line. The game is this fun — why rush to finish it?

11 Some Things Have No ROI

Let me share a candid number. In my first two years as a personal angel, I invested in 20 early-stage projects. Nineteen lost money. Manycore was the only winner — but it returned a thousand-fold.

Looking back, my capabilities were limited, but my luck was decent. Around the same time, I invested in a privatization deal that returned 10x in three years. Fast, clean, genuinely thrilling. But something felt off.

So when I founded Linear, I ultimately chose the lower-success-rate early-stage game — and specifically the Hard Mode within early-stage tech investing. I don't regret it at all, because it means more.

Watching Xiaohuang and his team grow from naive startup kids into "worldly-wise yet unworldly" tech entrepreneurs — that has no ROI. And because over these years I've watched our country, amid the great transformations of this century, increasingly treat technological innovation as a pillar of national destiny — that fulfillment is priceless.

12 Two Lines for Xiaohuang

Two things I'll leave with Xiaohuang: First, as an adult, you want both — make serious money, and hold onto that original desire to change the world. The two reinforce each other. I hope many years post-IPO, the Xiaohuang in flip-flops and a T-shirt is still in there.

Second, a selfish wish: I'm not selling after your IPO, but please take the Manycore team and earn me returns higher than what I made from Facebook stock.

13 Why Hangzhou?

I suppose I'm something of a "Hangzhou product" myself — a Zhejiang native, Zhejiang University graduate, just left too early. My achievements don't match the founders who stayed. I genuinely admire Zhejiang University entrepreneurs like Huang Xiaohuang and Wenfeng Liang.

But when I need to think through something, I do come back, take a lap around West Lake, and things seem to clarify.

These past couple years everyone's talking about Hangzhou's "Six Little Dragons." I've been asking: what does Hangzhou have going for it?

After I invested in Manycore's earliest angel round, the Hangzhou Shangcheng District government also gave them nearly 1.5 million RMB in startup subsidies. The difference: I took equity, they didn't. Of course, they had reasonable other expectations.

Many years ago, I didn't fully grasp what "socialism with Chinese characteristics" meant. Now I understand a little. Respect.

14 The Mountain Slope Greets Us with Slanting Light

Hangzhou's understanding of innovators didn't start today. A thousand years ago, my favorite and China's most resilient literary figure, Su Shi, didn't lie flat when he was down on his luck in Hangzhou. He oversaw the dredging of West Lake, using the excavated silt to build a causeway.

The West Lake beauty we see today was created by a politically frustrated scholar working through his confusion.

So if any founder here ever hits a wall, can't see a way forward — come to Hangzhou, walk Su Causeway, circle West Lake. If once around isn't enough, do two, do three. Eventually you'll find, as the mountain slope greets you with slanting light.

May every founder find their own Su Causeway.

These stories are just the tip of the iceberg. For the full version, search for Between the Lines on WeChat Channels, Bilibili, Xiaoyuzhou, Xiaohongshu, Tencent Video, and Weibo.