Author: Y Combinator
Compiled by: TechFlow
The Golden Age of Construction
This is the most opportune time in history for builders. We have just witnessed an astonishing scene: a giant robot "chopstick" precisely grabbing a falling skyscraper from the air. This is not only a technological marvel, but also a symbol of a huge leap in construction capabilities. Artificial Intelligence (AI) is changing the way we work at an unprecedented pace, especially for builders, AI is having a profound impact. It can be said that we are entering a golden age of construction, and this also provides us with a rare opportunity to create things that can truly make the country better. Here are some areas of innovation that we believe are particularly worth focusing on and exploring in this golden age.
Government Software
Author: Harj Taggar
Selling software to the government has always been known for its high barriers, and most entrepreneurs don't even consider entering this field. But if you can crack this problem, the rewards will be huge. For example, Palantir is one of the few startups that have successfully entered this market, and its current market value has reached $125 billion.
Now may be a particularly good time to try. Due to persistently high fiscal deficits, the government is urgently in need of relief through reduced spending and increased efficiency. At the same time, the rapid development of AI technology has made it possible to automate many of the government's administrative tasks that cost billions of dollars each year.
Combining these two points, developing AI-based software to help the government automate its work can not only reduce expenditures, but also improve efficiency. In particular, large language models (LLMs) have performed exceptionally well in handling repetitive administrative tasks such as filling out forms, reviewing applications, or summarizing documents. As users of government services, we will all benefit from more efficient services, such as no longer having to wait in long lines at the DMV.
Although the government may sound like an unattractive startup field, if you are willing to dig deeper, we look forward to hearing your ideas.
Public Safety Technology
Author: Garry Tan
Everyone should feel safe at home and when walking on the streets. This is a basic guarantee that a civilized society should provide for its citizens. Startups have already made efforts in this area. For example, Flock Safety (YC S17), which developed license plate cameras, has already helped solve 10% of reported crimes in the US, and their goal is to increase this to 25% by next year. Meanwhile, Abel Police (YC S24) has reduced the time police officers spend on paperwork from hours to just minutes, saving them up to 25% of their time each day for actual police work.
Public safety technology is bringing and will continue to bring real change. If you are innovating in the following areas, we are particularly interested in hearing from you:
Advanced computer vision: Develop computer vision-based technologies that can identify suspicious activities or people in need of assistance from video streams, while protecting personal privacy.
Emergency response technology: Improving the speed and coordination of emergency response is crucial. If you have ideas to get help to the needed location faster, we hope to help you make it a reality.
Community safety tools: Develop tools that can improve the interaction between communities and law enforcement, such as solutions that help neighbors take care of each other and stay informed about safety conditions.
Efficient law enforcement technology: Technologies that can help law enforcement work more efficiently and fairly, such as workload management systems or tools to improve operational precision.
If your startup is ready to join the wave of innovation in this field, we look forward to discussing with you.
American Manufacturing
Author: Jared Friedman
In the 19th century, Britain became the wealthiest country in the world by becoming the "world's factory". The United States also replicated this success in the 20th century. However, over the past few decades, the US has gradually abandoned this role. The hollowing out of manufacturing not only exacerbated social and political divisions, but also left the US in a geopolitically unstable position.
Bringing manufacturing back to the US is one of the areas of high bipartisan consensus at the moment. Elon Musk has already shown the feasibility of this goal by establishing Tesla super-factories in Austin and Nevada. We believe that the current technological advancements provide more opportunities for a new generation of builders to emulate his success.
New robotic systems based on machine learning (ML) make more production processes automatable, reducing the labor cost gap that has led to the offshoring of manufacturing to other countries. In addition, companies like SpaceX and Tesla have cultivated a whole generation of engineers who have mastered how to create an American enterprise that produces physical products but operates like a startup.
We have already seen successful cases of this model. For example, Astranis (W16) is building telecommunications satellites in the heart of San Francisco, which was once a shipbuilding site for the US Navy during World War II. Gecko Robotics (W16) is headquartered in the old industrial center of Pittsburgh, focusing on manufacturing robots for industrial inspection. Solugen (W17) produces industrial chemicals in large-scale factories in Houston.
Stablecoins 2.0
Authors: Brad Flora and Harj Taggar
Earlier this year, we released a call for more stablecoin startups. Since then, the stablecoin space has only gotten better. The main challenge for stablecoins has long been regulatory issues, and the US has failed in its attempts to regulate stablecoins. But now, the prospects for stablecoin regulation in the US are more optimistic, and we expect reasonable legislation to be enacted soon.
This year, the transaction volume of stablecoins has surged, now accounting for more than a fifth of Mastercard's payment volume. Nearly 30% of global remittances are now done through stablecoins, and traditional financial institutions like Visa are also providing platforms for banks to issue their own stablecoins. Furthermore, Stripe recently acquired a stablecoin startup Bridge for $100 million, which will undoubtedly attract more investors and capital to this field.
Therefore, now is one of the best times to start a stablecoin startup. We are particularly interested in ideas in the following directions:
Enterprise-focused services to help them more easily hold and manage stablecoins.
Developer tools that make it simple and easy to integrate stablecoin functionality.
If you are exploring innovations related to stablecoins, we look forward to discussing with you.
Large Language Models (LLMs) for Chip Design
Author: Garry Tan
Each breakthrough in AI technology will drive the demand for more powerful chips to support the training of larger-scale models. In this technological race, no country wants to fall behind. Chip design and manufacturing is now not only an economic issue, but also a key to survival in the post-AI era. OpenAI's O1 model has shown us that large language models (LLMs) with reasoning capabilities can drive major breakthroughs in science and engineering. We are very interested in any team that utilizes LLMs to improve chip design.
We are particularly focused on teams that specialize in designing ASICs (Application-Specific Integrated Circuits) and FPGAs (Field-Programmable Gate Arrays). Traditionally, the design of customized digital systems has required a large amount of development, design, and testing costs, making FPGA and ASIC R&D a high-cost, high-threshold area. However, with the emergence of large language models, these costs are significantly decreasing, making a wider range of specialized computing possible.
Currently, most computers use the von Neumann architecture, which processes programs and data through a single shared memory and operates through a serial fetch-and-execute cycle. The advantage of this architecture is its flexibility and ease of reprogramming. However, for specific tasks (such as cryptocurrency mining, data compression, or specialized encryption tasks), optimized algorithms and hardware design can achieve a 5 to 100-fold increase in computational speed, while reducing power consumption by 10 to 100 times.
The following chart provided by Taner Sadikoglu shows the difference in data flow between an optimized FPGA system and a traditional CPU.
Given the order-of-magnitude performance improvements that FPGAs and ASICs can bring, leveraging LLMs to optimize this process could yield highly valuable results and create tremendous business opportunities for startups.
Fintech 2.0
Author: Dalton Caldwell
The past two years have been challenging for fintech startups. The collapse of Silicon Valley Bank has led regulators to tighten restrictions on new startups, and investors have also been withdrawing from this field. However, we believe this situation is about to change, and now is the best time to start a fintech startup.
In the past, the most difficult part of starting a financial startup was reaching agreements with banks or other regulated partners. Now, with the emergence of service providers like Stripe and the proliferation of stablecoins, this process has become increasingly simple.
The rapid development of AI tools will inevitably drive transformation in the financial industry. For small startups without the baggage of traditional systems, this transformation presents a structural advantage, allowing them to quickly build the next generation of global financial products.
We believe now is the ideal time to create new-generation fintech companies based on existing infrastructure. We hope to see innovative ideas around insurance, investment banking, wealth management, and international payments.
New space companies
Authors: Jared Friedman and Dalton Caldwell
The cost of entering space is rapidly decreasing, with a more than 10-fold reduction since SpaceX's first launch in 2006. Today, a startup can build and launch a satellite with just a seed round of funding.
As entering space becomes as routine and low-cost as commercial aviation, shipping, or freight, it will unlock many entirely new business opportunities. Imagine how many kilograms of payload are being launched into space today, and how that number will grow in one year, five years, and ten years.
While starting a space company may seem ambitious, it is not necessarily more difficult than starting a software company. YC has already funded multiple space companies, including Astranis, Relativity Space, and Stoke among others, and their success rate may even exceed that of companies in other sectors.
AI-aided engineering tools
Author: Diana Hu
Physical engineering tools have seen little substantive progress for decades. The CAD/CAM software used by mechanical engineers, the EDA tools used by electrical engineers, and the CFD tools used by aerospace engineers - these tools still rely on complex numerical solvers and physical simulations. Not only are these tools computationally expensive, but they also require deep expertise, sometimes even requiring a Ph.D. to use effectively.
We believe that a new generation of AI-driven tools will fundamentally change this landscape.
By embedding the reasoning capabilities to solve mathematical and physical problems into new AI models, we can help engineers design and build physical systems like aircraft, buildings, circuits, chips, and satellites faster and with higher quality.
We look forward to founders developing AI-assisted engineering tools to drive this transformation and become the driving force behind the next generation of computer-aided engineering (CAE).
One million jobs 2.0
Author: Dalton Caldwell
We hope to fund startups that can create one million jobs that require human labor and are not dependent on AI to replace them.
Historically, whenever there have been major technological shifts, the types of jobs people do have changed. For example, many people used to be farmers, but with the advent of mechanization, agricultural labor has declined significantly. Similarly, jobs like elevator operators and typists have gradually disappeared.
However, technological changes often create better conditions for new jobs and bring greater value to humans. In this AI-driven new world, these jobs may include providing tools for more people to run their own local businesses, or earning a living by providing services to others, either online or offline.
Many AI futurists are uncertain about the future of occupations, and we hope to fund founders who can answer this question.
