From EVs to Penthouses

How BluSmart cofounders burned millions in public funds.

Hey — It’s Nico.

Welcome to another Failory edition. This issue takes 5 minutes to read.

If you only have one, here are the 5 most important things:

Let’s get into it.

This Week In Startups

🔗 Resources

Inside a high velocity sales-assisted GTM motion.

Swiss Cheese & Vertical Software.

📰 News

Anthropic rolls out an API for AI-powered web search.

Figma releases new AI-powered tools for creating sites and app prototypes.

Netflix is getting into short videos with a new vertical feed for mobile.

OpenAI Adds Shopping to ChatGPT in a Challenge to Google.

💸 Fundraising

German AI agents startup Parloa raises $120m

Finnish startup raises €30 million for world’s first autofocus glasses that work as “nature intended”.

Fastino trains AI models on cheap gaming GPUs and just raised $17.5M.

AI code review startup CodeAnt AI raises $2m seed funding.

Fail(St)ory

Greenwashed and Gone

This week, BluSmart shut down.

The Indian startup promised a cleaner, fairer ride-hailing model. Electric cars. Full-time drivers. No gig exploitation.

Instead, it left behind 10,000 drivers, 800 employees, and a corruption scandal that gutted the company from the inside.

What Was BluSmart:

BluSmart wasn’t trying to be just another ride-hailing app.

It wanted to fix everything people hate about the Ubers and Olas of the world: the surge pricing, the pollution, the endless contractor debates, the driver exploitation.

Instead of following the playbook, BluSmart flipped it.

  • No fossil fuels. Every vehicle in the fleet was electric.

  • No gig workers. Drivers were full-time employees with a fixed salary.

  • No driver-owned cars. BluSmart leased and operated the fleet itself.

  • No surge pricing. The price of a ride depended solely on the distance.

In theory, this gave them full control over quality, pricing, and experience. And it let them address a problem Uber still hasn’t solved: driver trust. In India especially, gig drivers often feel squeezed—low pay, no benefits, no job security. BluSmart saw that as a feature to fix, not ignore.

They built a system where drivers had stable income, riders got quiet and clean EVs, and BluSmart managed the whole supply chain—from the app to the charging hubs.

To do that, they used a hub-to-hub model. Drivers picked up and dropped off cars at centralized charging stations. That solved the range anxiety problem that EVs still face in India, and gave BluSmart control over where and when vehicles were charged.

By 2023, they had more than 8,000 electric cars on the road and thousands of chargers across Delhi NCR. On the surface, they looked like a cleaner, more ethical Uber—at scale.

But underneath, the model was delicate. Every piece depended on the others: the cars, the drivers, the charging stations, the capital to fund it all. If one part broke, the whole system could jam.

And eventually, it did.

The Numbers:

  • 📅 Founded in 2019

  • 🚘 8,700+ EVs and 6,300+ chargers

  • 🧑‍💼 10,000+ drivers and 800 employees affected

  • 💰 Raised $486M from public and private sources

  • 🛑 Services stopped abruptly on March 17

Reasons for Failure: 

  • Public funds, private apartment: Over $31 million meant to fund BluSmart’s electric vehicle fleet was quietly siphoned off by two of its co-founders. Most of that money came from public institutions that were backing India’s push toward green mobility. Instead of going toward EVs and charging stations, it went to personal luxuries, including a $5 million apartment.

  • A cost structure that couldn’t bend: Unlike Uber or Ola, BluSmart paid its drivers a fixed salary. That was great for worker stability—but brutal for cash flow. Salaries didn’t scale down when demand did. And since BluSmart also leased the cars and maintained the charging infra, nearly every cost was fixed. Once growth slowed or funding dried up, there was no wiggle room. The burn didn’t slow down, and they couldn’t sustain it.

  • Too many moving parts, not enough margin: BluSmart wasn’t just building an app. It was running a fleet, managing drivers, leasing vehicles, and building out a network of EV chargers—all at the same time. Each part depended on the others working perfectly. But EV infrastructure in India isn’t exactly plug-and-play, and BluSmart’s hub-to-hub model added even more friction. The complexity made it hard to operate efficiently and nearly impossible to scale cleanly.

  • Breaking the no-surge promise: BluSmart built its brand on the promise of no surge pricing—a key differentiator from competitors like Uber and Ola. However, in January 2024, the company introduced “Rush Hour” pricing, increasing fares by approximately 15% during peak times. This move contradicted their earlier stance and led to customer backlash. This shift not only eroded customer trust but also highlighted the challenges in maintaining a sustainable pricing model without dynamic adjustments

Why It Matters: 

  • If you’re handling public money, the margin for error is zero. There’s no “oops” when you misplace $31M.

  • You can’t fix the gig economy with a broken business model. Treating drivers better is only sustainable if your core economics work.

  • Too much control can become a trap. Owning the fleet and infra sounds good—until you need to scale fast or pivot.

Trend

OpenAI for Countries

Last week, OpenAI introduced OpenAI for Countries — a new initiative to work directly with governments around the world. 

The offer? Help building local data centers, customized versions of ChatGPT, and startup funding — all wrapped in a promise to support “democratic AI.”

Why It Matters:

  • OpenAI wants to be the AI partner for countries outside the U.S.

  • It’s offering infrastructure, tools, and money — but on its terms.

  • It’s also a geopolitical play to keep China out of the AI stack.

What’s the offer?

The idea behind OpenAI for Countries is simple: if other nations want to build serious AI capabilities, OpenAI wants to help — and make sure they don’t turn to China to do it.

Here’s what’s on the table:

  • Help building data centers, so countries can store and run AI models locally

  • Customized versions of ChatGPT, adapted to local languages, services, and laws

  • Co-investment in national AI startup funds.

  • Access to OpenAI tools and infrastructure — as long as you’re aligned with their values

This isn’t just a product expansion. It’s OpenAI trying to shape how AI gets built and governed around the world. The company calls it “democratic AI,” which sounds noble, but it also happens to strengthen U.S. influence in the global AI stack. Countries get infra. OpenAI gets distribution, usage, and a stronger geopolitical moat.

The infrastructure piece builds on Project Stargate — that $500B mega-plan with Oracle and SoftBank to build AI superclusters. Until now, Stargate was focused on the U.S. This new push brings it abroad. Countries will need to co-fund it, but in return, they’re promised more control over how their data is stored, used, and governed. At least on paper.

What’s at stake?

One of the most interesting parts is customization. OpenAI says it’s willing to tailor ChatGPT to the culture, public services, and regulatory needs of each country. This could open the door to local education and healthcare integrations, or more responsive digital services. But it also raises a big question: how flexible is OpenAI, really? Will countries get true customization, or just some translation layers and regional prompts?

Then there’s the national startup fund. OpenAI wants to help seed local ecosystems — putting money into new companies that will be built on top of its own tools. Governments are expected to match the funding. It’s smart: help build the ecosystem and become the default platform powering it.

But none of this is ocharity. It’s a strategic play to make sure OpenAI becomes the global standard before its rivals d. The subtext here is China. Beijing is pushing its own models, exporting them through its own alliances. OpenAI wants to lock in its position early — not just as a provider, but as the foundation for how other countries build AI.

So, while the pitch is full of idealistic language, the reality is more complex. This is about influence, standards, and long-term control. Whether countries are okay with that tradeoff is the real test.

Help Me Improve Failory

How Was Today's Newsletter?

If this issue was a startup, how would you rate it?

Login or Subscribe to participate in polls.

That's all of this edition.

Cheers,

Nico