Hey — It’s Nico.

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

These are the 5 most important things:

• Pandion, a delivery and logistics startup that boomed during the pandemic, announced it is shutting down — learn why below.

• Andreessen Horowitz’s top 50 big ideas for 2025.

• OpenAI now has a complete robotics team.

• Colossal Biosciences, a startup that wants to bring mammoths back to life, raised $200 million.

• GPT now has “Tasks”, which allow it to set reminder and act independently — learn why this a big deal below.

Let’s get into it.

This Week In Startups

🔗 Resources

Five founder tips from someone who’s interviewed 100+ founders.

Andreessen Horowitz’s top 50 big ideas for 2025.

YC believes vertical AI agents can be 10x bigger than SaaS.

The secrets of Stealth Startups and how to build one successfully.

📰 News

Bluesky is getting its own Instagram competitor: Flashes.

Mastodon announces transition to nonprofit structure.

Hundreds of thousands of users are migrating from TikTok to RedNote.

OpenAI now has a complete robotics team.

💸 Fundraising

Stoke Space, a YC-backed resuable rocket startup, raised $260 million.

Synthesia, a startup making AI avatars, raised $180 million.

ElectraLith raises $17 million to produce low-cost lithium.

Colossal Biosciences, a startup that wants to bring mammoths back to life, raised $200 million.

Fail(St)ory

Putting AI into Logistics

Last week, Pandion, a delivery and logistics startup that once promised to challenge the dominance of UPS, FedEx, and Amazon, announced it is shutting down.

Founded in 2020 by a former Amazon and Walmart executive, Pandion entered the market at a time when pandemic-fueled demand was stretching traditional delivery providers to their limits. However, just a few years later, the logistics industry has cooled, leaving Pandion unable to secure a future.

What Was Pandion: Pandion aimed to revolutionize delivery logistics with flexible, AI-driven technology designed to optimize on-time delivery rates at lower costs.

The company operated five sortation centers and partnered with last-mile carriers to offer one- to five-day ground shipping for over 100,000 packages daily. Their pitch to shippers: better delivery performance without the expense of building an in-house logistics network.

At its peak, Pandion employed 63 people and attracted significant attention as an alternative to major delivery giants. It raised approximately $125 million in equity funding over five years and was on track to hit $220 million in sales by 2024.

The Numbers:

• 📅 Founded in 2020.

• 🧑‍💼 Employed 63 people.

• 💰 Raised $125M in funding.

• 📦 Delivered 100,000+ packages daily.

• 📉 Sales projections: $220M in 2024.

Reasons for Failure: 

• Cooling Market Condition: Pandion launched during a period of unprecedented demand for delivery services, driven by the pandemic. Companies like FedEx and UPS struggled to expand capacity, creating a window for new entrants like Pandion. But as the pandemic-era boom subsided, demand leveled off, leaving little room for smaller players to compete. Similar competitors like Point Pickup, Maergo and Convoy also shut down in recent years, signaling a broader contraction in the market.

• Lack of a Sustainable Differentiator: While Pandion’s AI-driven technology promised to improve delivery efficiency, it wasn’t enough to create a lasting edge. Established players like UPS and FedEx had decades of infrastructure and relationships with clients, while Amazon had the scale to undercut competitors on pricing.

• Scaling Ahead of Demand: Pandion quickly expanded its operations, including building five sortation centers and delivering over 100,000 packages daily. While impressive, this rapid scaling was based on a market environment that was already cooling. By committing to such high operational costs without secured long-term demand, the company locked itself into a cost structure it couldn’t sustain.

• Challenges in Securing Acquisition Deals: In its final days, Pandion engaged in discussions with potential acquirers but was unable to secure a deal. This inability to find a buyer or additional investors was a significant factor in the decision to shut down operations.

Why It Matters: 

• Scaling based on temporary demand spikes can backfire as markets normalize.

• Competing with giants requires more than innovative tech—sustainability is key.

• Overreliance on third-party partners weakens control and differentiation.

• Counting on acquisition as an exit strategy is a risky bet.

Trend

GPT Tasks

This week, OpenAI is moving beyond simple chats. With a new beta feature in ChatGPT, called “Tasks,” they’re exploring how AI can help manage our daily to-do lists and remind us of future deadlines.

It’s not just about quick answers anymore—it’s about having an assistant that knows what you need, when you need it, and can help keep you on track.

What Tasks Can Do:

• Schedule Reminders: Set one-time or recurring reminders for various activities. For example:

  • One-time Reminder: Get notified about an upcoming concert ticket sale.

  • Recurring Reminder: Receive daily weather updates each morning.

• Automate Routine Actions: Configure ChatGPT to perform specific tasks at designated times. For instance:

  • Automatically log work hours at the end of each day.

  • Send a weekly summary of industry news every Friday.

• Suggest Tasks Based on Conversations: ChatGPT can even propose tasks aligned with your ongoing discussions, which you can choose to accept or decline. 

Here’s an example: A user set up Tasks to provide daily updates on Apple stock prices. Every morning, ChatGPT automatically delivered a brief summary of the latest trends, helping the user stay informed without any extra effort.

That said, it’s worth clarifying what Tasks can and cannot do. While it allows you to schedule reminders and even automate certain web searches, its capabilities are still limited. Tasks cannot yet integrate with other applications or systems, so you can’t, for instance, ask it to book a restaurant reservation every Friday. Tasks focuses on reminders and simple workflows, leaving more complex interactions for future developments.

And that’s precisely where the true importance of Tasks lies—not in its current capabilities but in the direction it signals for AI. This feature is the first step in a larger vision, one where AI transitions from a responsive tool to a proactive partner.

AI Agents: The First Step Forward

The release of Tasks represents a foundational shift toward what many see as the next evolution of AI: autonomous agents. These are systems that don’t just respond to commands but proactively take action, anticipate needs, and execute tasks on their own.

While Tasks serves as a bridge into this future, OpenAI’s upcoming release of “Operator” will push these boundaries further. Operator will allow AI to independently control computers, moving beyond scheduling to actively managing workflows and complex operations. Together, Tasks and Operator signal the start of a new era in AI—one where the tools we use are not just assistants, but collaborators in how we work and innovate. It's the reason why, for many, 2025 will be the year of the Agentic AI.

Help Me Improve Failory

That's all of this edition.

Cheers,

Nico