Golden Age of Building

By Harj Taggar

Selling software to the government is notoriously hard and not something most builders would even think about doing. Still, the prize is huge if you can figure it out.  Palantir is one of the only startups that managed to figure it out and their market cap today is $125 billion.

Now might be a uniquely good moment to give it a shot. The deficit is running high and there is hope the government will find ways to reduce spending and operate more efficiently. At the same time, AI is now powerful enough to automate many types of administrative work the government spends billions of dollars per year on.

If you combine those ideas, building AI software to automate work done by the government would both reduce spending and make the government more efficient. In particular, we’ve seen that LLMs are especially good at automating rote repetitive admin tasks like filling out forms, reviewing applications or summarizing documents.  As consumers of government services, we’d all benefit from more efficiency, imagine never having to wait in line at the DMV again.

It’s hard to think of something that sounds less appealing as a place to look for startup ideas than the government but if you’re up for digging in that space we’d love to hear from you.

By Garry Tan

We all deserve to be safe in our homes, and while walking around on our streets. This is a basic thing civilization should afford its citizens. Startups are already on the case. License plate cameras built by Flock Safety (YC S17) now help solve 10% of reported crimes in the USA and they’re aiming to get to 25% of all crime by next year. Abel Police (YC S24) cuts the amount of time police officers spend filling out paperwork from hours to minutes, giving them up to 25% more hours per day to do real police work.

Public safety tech can and will make a real difference.  We are especially interested in hearing from you if you’re working on these ideas:

• Advanced Computer Vision: Your startup could develop technology that uses computer vision to enhance safety—identifying suspicious activities or people in distress from video feeds, without compromising individual privacy.
• Emergency Response Enhancements: Technologies that improve the speed and coordination of emergency responses are lifesavers. If you have an idea that can get help to where it’s needed faster, we want to help you make it happen.
• Community Safety Tools: We are looking for tools that improve how communities and law enforcement interact. Your solution could make it easier for neighbors to watch out for each other and stay informed about their safety.
• Efficient Law Enforcement Technologies: We’re interested in technologies that help law enforcement do their jobs more effectively and fairly, from workload management to increasing accuracy in their operations.
  

If your startup is ready to be part of this movement, we want to hear from you.

By Jared Friedman

The UK became the world's richest country in the 19th century by being the workshop of the world. The US did the same in the 20th century. But in recent decades, we've given up this role. The hollowing out of US manufacturing has led to social and political division and left us in a precarious place geopolitically.

Bringing manufacturing back to America is one of the biggest areas of bipartisan agreement. Elon has shown us it is possible to do it by building Tesla Gigafactories in Austin and Nevada. We think other changes in the world have made it easier for new builders to follow his example.

New ML-based robotics systems will make it possible to automate far more, which will reduce the cost-of-labor arbitrage that pushed manufacturing to other countries in the first place.  Companies like SpaceX and Tesla have trained an entire generation of engineers in how to build an American company that makes physical products but operates like a startup.

We know this works because we've had experience working with some of the leading companies in this space. Astranis (W16) is building telecommunications satellites in the heart of San Francisco, in a building that used to build warships for the US Navy during WWII. Gecko Robotics (W16), based in America's old industrial heartland of Pittsburgh, builds robots that do industrial inspections. Solugen (W17) makes industrial chemicals from a large-scale plant in Houston.

By Brad Flora and Harj Taggar

At the start of this year, we posted a request for more stablecoin startups and since then things have only gotten better for stablecoins.   The black cloud over stablecoins has always been regulation, with several efforts to pass regulation in the US failing. The regulatory future for stablecoins in the US now looks more promising and we expect sensible legislation is on its way soon.

Stablecoin payment volumes have surged this year and are now over a fifth of Mastercard’s volume. Almost 30% of global remittances are now facilitated through stablecoins and traditional finance institutions like Visa are offering platforms for banks to issue their own stablecoins.  Also Stripe recently acquired a stablecoin startup, Bridge, for $1 billion, which will only attract more investor interest and capital into the space.

This makes it a better time than ever to start a stablecoin startup. We are especially interested in hearing about ideas that target businesses, helping them to hold and manage stables and also services that make it easy for developers to integrate with them.

If you’re working on stablecoin ideas we’d love to hear from you.

By Garry Tan

Each breakthrough in AI creates demand for more powerful chips to train larger models. No country wants to fall behind in this arms race. Domestic chip design and manufacture is not just about economics anymore, it's about survival in a post AI world.  OpenAI O1 showed us that LLMs with reasoning can power breakthroughs in science and engineering and we’re interested in talking to anyone using LLMs to improve chip design.

We are especially interested in anyone focusing on designing ASICs and FPGAs. Design of customized digital systems whether through FPGAs (field programmable gate arrays) or ASICs (application-specific integrated circuits) has typically been costly because of the amount of custom design, development and testing necessary to bring such a system online. With the advent of large language models, these costs are coming down significantly, such that ever more specialized types of computation could be done.

Our normal computing environment assumes Von Neumann architecture using CPUs that we are familiar with: single shared memory for programs and data, arithmetic unit and a program control unit, operating through fetching and execution cycles serially. Most computers and computation use this because it’s very easy to reprogram such systems.

We know there is a clear engineering trade-off: it is possible to optimize especially specialized algorithms or calculations such as cryptocurrency mining, data compression, or special-purpose encryption tasks such that the same computation would happen faster (5x to 100x), and using less energy (10x to 100x).  This is a diagram (credit: Taner Sadikoglu) showing the differences in how data flows an optimized FPGA system versus a normal CPU.  

Given the order of magnitude improvements possible with specialized FPGAs and ASICs, use of LLMs to optimize this process is likely to produce extremely useful results and great opportunities for startups.

By Dalton Caldwell

The last two years have been a rough time for fintech startups. The collapse of Silicon Valley Bank led to regulators clamping down on new startups and investors fled the space. We're optimistic this is about to change and now is a great time to start a fintech startup.

Historically the hardest thing about starting a new financial startup was the difficulty of getting a deal with a bank or other regulated partner. We are now in a new era where this keeps becoming easier with the advent of providers like Stripe and new core technologies like stablecoins.

AI tools will inevitably cause the financial industry to change rapidly and we believe there could be a structural advantage for a small startup without any legacy systems or processes to quickly build global financial products of the future.

We believe this is the perfect time to start a new generation of financial technology companies built on top of the infrastructure now available. We would like to see ideas around insurance, investment banking, wealth management, international payments and more.

By Jared Friedman and Dalton Caldwell

The cost to reach orbit is falling fast, having fallen over 10x since SpaceX’s first launch in 2006. A startup can now build and launch a satellite on just a seed round.

If you think about how many kilograms of payload get launched into space today, imagine how many will be sent up in one year, in five years, in ten years, and so on.

If we are entering a future with access to space being as routine and inexpensive as commercial air travel, shipping or trucking… what new businesses does that unlock?

Building a space company might scare founders by seeming too ambitious, but surprisingly, it is not necessarily harder than building a software company. YC has funded many space companies — Astranis, Relativity Space, Stoke, and many others  — and their success rate has been no lower, and maybe higher, than our other companies.

By Diana Hu

Engineering tools for the physical world have barely evolved in decades. The CAD/CAM software that mechanical engineers use, the EDA tools for circuit and chip design that electrical engineers use, and the CFD tools for fluid and thermal analysis that aerospace engineers use — All still rely on complex numerical solvers and physics simulations. These are computationally expensive and also require deep training and at times even a PhD to use effectively.

We believe the next generation of AI-powered tools will change that.

With reasoning capabilities built into the new AI models that solve math and physics, we can unleash engineers to design and build physical systems like —planes, buildings, circuits, chips, satellites—faster and better than ever before.

We’re eager to see founders build the AI-aided engineering tools that will drive this transformation as the new generation of Computer-aided Engineering.

By Dalton Caldwell

We would like to fund startups that have a useful need to employ a million workers in a way that uniquely needs humans to do the job and there will be no structural need for the job to be done with AI.

In the past, when massive technological change came about, people have ended up doing different types of work than they did in the past. For example, a lot of people used to be farmers, and then because of machinery a lot fewer people were farmers. The same goes for careers like elevator operators and typists.

Often the types of new careers that are created have better working conditions and are more helpful to humanity.  In this new AI driven world, these might be tools for more people to run their own local businesses, or be able to earn a living by providing services to other humans either online or offline.

Many AI futurists are unsure what the careers of the future are, and we want to fund founders with an answer to that question.