That Startup

In 2021, Climeworks charged $1,000/ton to pull CO₂ from the air. Most called it too expensive. They signed 2,000+ customers anyway. People will pay premium prices for proof something works.

In 2019, Vernacular.ai hired linguists before engineers. They needed people who understood how Indians actually speak. Not textbook Hindi. Sanjay Goyal built a voice AI that handles 30M calls/month. Language-first hiring was the whole bet.

Most founders want distribution after they build. Lei Jun built the distribution first. Find your 100 obsessive early users. Let them shape the product publicly. Make them feel like co-creators. By launch day, they won't just buy. They'll sell for you. #startups

In 2010, Xiaomi had zero phones and zero factories. But Lei Jun spent 18 months just talking to engineers before building anything. Most founders would've launched in month 2. He didn't ship until he had 13 co-founders who were all ex-Google, ex-Motorola, ex-Microsoft.

By 2014, Xiaomi was valued at $45 billion. Faster than any consumer company in history at that point. The hard truth: Lei Jun didn't win on the phone. He won because he spent 18 months building a fanbase that trusted him before he asked them to buy anything.

In 2011, Xiaomi sold 300,000 phones in 34 hours. They didn't buy a single ad. The entire launch ran through MIUI forums where fans had been beta testing software for a year. Those beta testers became the sales force. Community was the product before the product was.

Everyone said Xiaomi was a hardware company. Lei Jun said it was a software company that happened to sell hardware. Big difference. He priced phones at cost and planned to make money on apps and services. No carrier subsidies. No retail stores. Pure internet model.

In 2020, Patreon hired 4 creators as full-time employees. Not to build product. To veto product decisions. Creator Advisory Board with actual power. Retention jumped. Churn dropped. The people using the product became the people protecting it.

In 2023, Lelapa AI built Africa's first Zulu AI model. No VC backing. Just a small Cape Town team training models on languages Big Tech ignored. Now enterprises pay to reach 12M+ Zulu speakers. The next AI gold rush is languages no one else bothered with.

The next 10 years in SEA fintech will be won by boring specialists, not blockchain missionaries. Pick one corridor. One use case. One country. Omise built something real before they got distracted. Most founders never even get that far. #fintech

In 2016, Omise raised $17.5M and was processing payments across Thailand, Japan, and Singapore. By 2021, OmiseGO — their blockchain spinoff — had collapsed from a $1.5B valuation to near zero. The parent company survived. The hype vehicle didn't.

Southeast Asia's fintech wave is real. 200M unbanked adults. Mobile-first populations. Fragmented banking infrastructure. But the founders who win here aren't selling the vision of disruption. They're solving a specific payment problem in one country first.

OmiseGO promised to 'bank the unbanked' across Southeast Asia using blockchain. But Omise's core product already did that — quietly, profitably, without a whitepaper. They had the solution. They just dressed it up in a costume no one needed.

Everyone blamed crypto winter for OmiseGO's collapse. The real problem: Omise had a working fintech business and abandoned focus to chase a token sale. The $25M ICO in 2017 felt like a shortcut. It was actually a trap.

In 2023, Charm Industrial charged $600/ton to bury CO2. Every other carbon removal startup charged less. They kept the price high on purpose. It signaled quality. Microsoft signed a 10-year deal. #climatetech

In 2023, Writer AI grew from $0 to $50M ARR. No ads. They embedded their AI writing tool directly into enterprise workflows — IT never had to approve it. Employees just started using it. Then IT had no choice but to buy it. Bottom-up wins again.

If your product creates real value, charging less upfront isn't desperation. It's strategy. Get users dependent on the workflow first. Monetize the expansion later. Ramp didn't discount their way to growth. They priced for dominance.

In 2024, Ramp cut its software pricing to $0 for small teams. Within months, it was processing $55B+ in annualized spend. Their biggest growth lever wasn't sales. It wasn't ads. It was giving the product away.

This is the lesson most B2B founders miss. Price high to look premium and you slow adoption. Price low to drive adoption and you build a moat out of habit and integration. Ramp chose the moat. Now it's valued at $13B. #fintech

Ramp's bet was simple: make expense management so cheap to start that the switching cost from competitors becomes the only real barrier. Once your finance team lives in Ramp, you don't leave. They optimized for lock-in, not early revenue.

Most people think free tiers are charity. Ramp saw it differently. Every free user was a future enterprise buyer, a referral engine, or a Trojan horse inside a company that would eventually upgrade. Free was the funnel.