That Startup

If you're building in a crowded space right now: Don't compete. Reframe. Srinivas didn't fight Google on Google's terms. He changed the question entirely. Find the thing everyone accepts as fixed. That's your opening. #startups

In 2024, Perplexity AI was worth $520M. By early 2025, it crossed $9 billion. Arvind Srinivas built it in NYC with 3 people and almost no marketing budget. The product grew because of one weird decision most founders would never make.

The lesson isn't about fundraising. It's about framing. Srinivas never said he was competing with Google. He said he was replacing the search bar. Same market. Totally different story. Investors funded the story before they funded the numbers.

In 2022, Srinivas cold emailed Jeff Bezos, Nvidia's Jensen Huang, and Elad Gil. All three invested. He didn't pitch a deck first. He sent them a working product and one sentence: 'This is what search should feel like.'

Most people think Perplexity won because of better AI. That's not it. Google had better models. OpenAI had more users. Perplexity won because Srinivas refused to build a chatbot. He built an answer engine. Small difference. Massive outcome.

In 2009, Maicih chips sold zero in stores. They only sold through 20 "agents" on Twitter. No retail. No ads. Followers had to find the agents to buy. FOMO did the rest. $1M revenue in year one.

In 2023, Ninja Van lost $170M. Everyone called it dead. It still delivered 2.5B parcels across SEA since launch. Unprofitable ≠ dying. Sometimes scale IS the strategy. #logistics

If you're building hardware in a price-sensitive market: Don't chase the premium customer first. Own the middle tier nobody respects. Boat hit ₹2,500 crore revenue by 2022 by being the best option in a bracket everyone else ignored. #startups

In 2015, Boat's co-founder Aman Gupta had ₹0 in revenue, a warehouse full of unsold cables, and a bank account nearly empty. Most people think Boat was an overnight success. It wasn't.

Hardware startups in India kept failing because they copied Western pricing for Indian wallets. Boat did the opposite. They built decent quality at a price point where the next option was a fake Chinese product. That gap was worth billions.

Gupta's fix was brutally simple. Stop trying to compete with Sony and JBL in retail stores. Go where those brands weren't paying attention: Amazon India. They priced aggressively, collected reviews obsessively, and owned the ₹500-₹2000 segment.

The common story: Boat blew up because of good branding. The real story: Boat almost died before anyone knew its name. In year one, they couldn't move inventory. Retailers wouldn't stock an unknown Indian audio brand. Returns piled up.

In 2000, Shaadi.com had 1 million users but zero revenue. So they charged ₹500/month. Half the users left overnight. The other half became paying customers. That's when they knew they had a real business.

In 2008, Spotify couldn't get labels to license its music. So Daniel Ek flew to every major label and demoed it in person. No pitch deck. Just the product. That hand-to-hand distribution got them their first licenses. The rest is history.

Before your next raise, answer this honestly: Can you hit your yield targets today, at current size? If not, fix that first. Northvolt had the money, the partners, and the mission. It didn't have the operational proof. That was enough to end it.

Northvolt raised $15 billion. Had backing from Goldman Sachs, Volkswagen, and the EU. Filed for bankruptcy in 2024. The hard truth European climate founders don't want to hear is buried in how it happened.

The AI era is pushing European climate founders to raise bigger and move faster. Northvolt is the warning. If your unit economics and production yields aren't proven at small scale, more funding is just a longer fall. Validate the process. Then scale.

In 2023, Northvolt delivered only 8% of its contracted battery volume to BMW. BMW cancelled a $2B order. That one number triggered a collapse that $15B couldn't stop. One operational failure became a financial death sentence. #ClimaTech

Everyone assumed Northvolt failed because of bad tech or wrong market. The actual problem: they scaled manufacturing before they mastered it. Yield rates at their Swedish gigafactory were catastrophically low. Capital can't fix an unproven process. It just burns faster.

In 2024, Hippocratic AI ran out of money mid-Series A. Munjal Shah cold-emailed every investor he knew. They closed $141M in 6 weeks. Now they're building AI nurses for 100M+ underserved patients. Rejection nearly killed it. Urgency saved it.