Y Combinator: 10 innovative directions worth exploring in the field of AI
Original source: Y Combinator
Original translation: TechFlow
The Golden Age of Construction
This is the best time in history for builders. We just witnessed an amazing scene: a giant robot "chopsticks" accurately grabbed a falling skyscraper from the air. This is not only a miracle of technology, but also a symbol of a huge leap in construction capabilities. Artificial intelligence (AI) is changing the way we work at an unprecedented speed, especially for builders, AI is having a profound impact. It can be said that we are entering a golden age of construction, which also provides us with a rare opportunity to create things that can truly make the country better. Here are some innovative directions that we think are particularly worth paying attention to and exploring in this golden age.
Government Software

(Video link)
By Harj Taggar
Selling software to government is notoriously difficult, and most entrepreneurs won’t even consider entering the field. But if you can crack it, the rewards can be huge. For example, Palantir is one of the few startups that has successfully entered this market and is now valued at $125 billion.
Now may be a particularly good time to try. With fiscal deficits high, governments are eager to ease pressure by reducing spending and improving efficiency. At the same time, the rapid development of artificial intelligence technology has made it possible to automate many administrative tasks that cost governments billions of dollars each year.
Combining these two points, developing AI-based software to help automate government work can not only reduce expenses but also improve efficiency. In particular, large language models (LLMs) excel at handling repetitive administrative tasks such as filling out forms, reviewing applications, or summarizing documents. As users of government services, we all benefit from more efficient services, such as no more long lines at the DMV.
While government may sound like a less sexy area to start a business, if you are willing to dig deeper, we would love to hear your ideas.
Public Safety Technology

(Video link)
By Garry Tan
Everyone should feel safe in their homes and on the streets. This is a basic security that a civilized society should provide to its citizens. Startups are already working in this area. For example, license plate cameras developed by Flock Safety (YC S17) have helped solve 10% of reported crimes in the United States, and their goal is to increase this to 25% by next year. Meanwhile, Abel Police (YC S24) reduces the time officers spend filling out paperwork from hours to minutes, saving them up to 25% of their time each day for actual policing.
Public safety technology is and will continue to make a real difference. We’d especially like to hear from you if you’re innovating in the following areas:
· Advanced Computer Vision:Developing computer vision-based technology that can identify suspicious activity or people in need of help from video streams while protecting individual privacy.
· Emergency Response Technology:Technology that improves the speed and coordination of emergency response is critical. If you have an idea that can get help to where it’s needed faster, we want to help you make it a reality.
· Community Safety Tools:Developing tools that improve community interactions with law enforcement, such as solutions that help neighbors look out for each other and stay informed about safety conditions.
· Efficient Law Enforcement Technology:Technology that helps law enforcement work more efficiently and fairly, such as workload management systems or tools that improve operational precision.
If your startup is ready to join the wave of innovation in this area, we’d love to talk to you.
Made in the USA

(Video link)
By Jared Friedman
In the 19th century, Britain became the world’s wealthiest country by becoming the “world’s factory.” The United States replicated this success in the 20th century. However, over the past few decades, the United States has gradually abandoned this role. The hollowing out of manufacturing has not only exacerbated social and political divisions, but also put the United States in a geopolitically unstable position.
Bringing manufacturing back to the United States is one of the areas where there is currently strong bipartisan consensus. Elon Musk has shown us how feasible this goal is by building Tesla Gigafactories in Austin and Nevada. We believe that current technological advances provide more opportunities for a new generation of builders to emulate his success.
New robotic systems based on machine learning (ML) allow more production processes to be automated, thereby reducing the labor cost gap of outsourcing manufacturing to other countries. In addition, companies like SpaceX and Tesla have trained a whole generation of engineers who have mastered how to create an American company that produces physical products but operates as a startup.
We’ve already seen this model work. For example, Astranis (W16) builds telecommunications satellites in the heart of San Francisco, where warships were built for the US Navy during World War II. Gecko Robotics (W16), based in Pittsburgh, the old industrial heartland of the US, builds robots for industrial inspection. Solugen (W17) produces industrial chemicals in a large factory in Houston.
Stablecoin 2.0

(Video link)
By Brad Flora and Harj Taggar
Earlier this year, we published a request for more stablecoin startups. From there, things have only gotten better for the stablecoin space. For a long time, the main challenge with stablecoins has been regulation, with several attempts to pass stablecoin regulation in the US failing. But now, the outlook for stablecoin regulation in the US is more positive, and we expect sensible legislation to be on the horizon.
This year, stablecoin payment transactions have surged, now accounting for more than a fifth of Mastercard’s payment volume. Nearly 30% of global remittances are done in stablecoins, and traditional financial institutions like Visa are also providing platforms for banks to issue their own stablecoins. Additionally, Stripe recently acquired a stablecoin startup, Bridge, for $1 billion, which will undoubtedly attract more investors and capital to the space.
Therefore, now is one of the best times to launch a stablecoin startup. We are particularly interested in ideas in the following directions:
· Services for enterprises to help them hold and manage stablecoins more easily.
· Provide developers with easy-to-use tools to quickly integrate stablecoin functionality.
If you are exploring innovations related to stablecoins, we are very much looking forward to communicating with you.
Large Language Models for Chip Design (LLMs for chip design)

(Video link)
Author: Garry Tan
Every breakthrough in AI drives the need for more powerful chips to support the training of larger models. In this technological race, no country wants to fall behind. The design and manufacture of chips is now not only an economic issue, but also the key to survival in the post-AI era. OpenAI's O1 model shows 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 uses LLMs to improve chip design.
We are particularly interested in teams that focus on designing ASICs (application-specific integrated circuits) and FPGAs (field-programmable gate arrays). Traditionally, designing customized digital systems requires a lot of development, design, and testing costs, so the research and development of FPGAs and ASICs has always been a high-cost, high-threshold field. With the emergence of large language models, these costs are falling significantly, making more types 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 serial instruction fetch and execution cycles. The advantage of this architecture is high flexibility and easy reprogramming of the system. However, for specific tasks (such as cryptocurrency mining, data compression, or specialized encryption tasks), by optimizing algorithms and hardware design, it is possible to achieve a 5 to 100 times increase in computing speed while reducing energy consumption by 10 to 100 times.
Below is a diagram provided by Taner Sadikoglu showing the difference in data flow between an optimized FPGA system and a traditional CPU.

Given the orders of magnitude performance gains that FPGAs and ASICs can deliver, optimizing this process with LLMs could be extremely valuable and create a huge business opportunity for startups.
Fintech 2.0

(Video link)
By Dalton Caldwell
The past two years have been challenging for Fintech startups. The collapse of Silicon Valley Bank led to regulators tightening restrictions on new startups and investors retreating from the space. However, we believe this is about to change and that now is the best time to start a Fintech startup.
In the past, the hardest part of starting a financial startup was getting an agreement with a bank or other regulated partner. Now, with the emergence of service providers such as Stripe and the popularity of new technologies such as stablecoins, this process is becoming easier.
The rapid development of AI tools will inevitably drive changes in the financial industry. For small startups without the burden of legacy systems, this change brings structural advantages, allowing them to quickly build future global financial products.
We believe that now is the ideal time to create a new generation of fintech companies based on existing infrastructure. We hope to see innovative ideas around insurance, investment banking, wealth management, international payments and other fields.
New space companies

(Video link)
By Jared Friedman and Dalton Caldwell
The cost of accessing space is falling rapidly, having dropped more than 10-fold since SpaceX’s first launch in 2006. Today, a startup can build and launch a satellite with just a seed round of funding.
As access to space becomes as routine and low-cost as commercial aviation, shipping, or cargo transportation, this will unlock many new business opportunities. Can you imagine how many kilograms of payload are launched into space today? How will that number grow in one year, five years, and ten years?
While starting a space company may seem ambitious, it’s not necessarily more difficult than starting a software company. YC has funded several space companies, including Astranis, Relativity Space, and Stoke, and their success rate may even exceed that of companies in other fields.
AI-aided engineering tools

(Video link)
Author: Diana Hu
Engineering tools for the physical world have made little real progress for decades. CAD/CAM software used by mechanical engineers, EDA tools used by electrical engineers, and CFD tools used by aerospace engineers—these tools still rely on complex numerical solvers and physics simulations. These tools are not only computationally expensive, but also require deep expertise, sometimes even a Ph.D. to use effectively.
We believe that a new generation of AI-driven tools will change all that.
By infusing new AI models with the reasoning power of solving math and physics problems, we can help engineers design and build physical systems—planes, buildings, circuits, chips, satellites—faster and with higher quality.
We look to founders to develop AI-assisted engineering tools to drive this change and become the driving force of the next generation of computer-aided engineering (CAE).
One million jobs 2.0

(Video link)
By Dalton Caldwell
We want to fund startups that can create one million jobs that require humans to do the work and do not rely on AI to replace them.
Historically, whenever there has been a major change in technology, the types of jobs people do have changed. For example, in the past, many people were farmers, but with the spread of mechanization, the agricultural labor force has been greatly reduced. Likewise, professions such as elevator operators and typists are disappearing.
However, technological change often creates new professions that are better equipped and provide greater value to humanity. In this new AI-driven world, these professions may include giving more people the tools to run their own local businesses, or to make a living by providing services to others online or offline.
Many AI futurists are uncertain about what the future of professions will look like, and we want to fund founders who can answer that question.
Original link
You may also like

Dragonfly Partners: Most agents will not engage in autonomous trading, how can crypto payments prevail?

US AI Startup Goes All In on Chinese Mega-Model | Rewire News Morning Brief

Trump Lies Again: A "Five-Day Pause" Psyop, How Wall Street, Bitcoin, and Polymarket Insiders Synced Uposciogen

When a Token Becomes Labor, People Become the Interface

Ceasefire News Leaked Ahead of Time? Large Polymarket Bets on Outcome Before Trump's Tweet

BlackRock CEO's Annual Shareholder Letter: How is Wall Street Using AI to Keep Profiting from National Pension Funds?

Sun Valley Releases 2025 Financial Report: Bitcoin Mining Revenue Reaches $670 Million, Accelerating Transformation to AI Infrastructure Platform
On March 16, 2026, in Dallas, Texas, USA, CanGu Company (New York Stock Exchange code: CANG, hereinafter referred to as "CanGu" or the "Company") today announced its unaudited financial performance for the fourth quarter and full year ended December 31, 2025. As a btc-42">bitcoin mining enterprise relying on a globally operated layout and dedicated to building an integrated energy and AI computing power platform, CanGu is actively advancing its business transformation and infrastructure development.
• Financial Performance:
Total revenue for the full year 2025 was $688.1 million, with $179.5 million in the fourth quarter.
Bitcoin mining business revenue for the full year was $675.5 million, with $172.4 million in the fourth quarter.
Full-year adjusted EBITDA was $24.5 million, while the fourth quarter was -$156.3 million.
• Mining Operations and Costs:
A total of 6,594.6 bitcoins were mined throughout the year, averaging 18.07 bitcoins per day; of which 1,718.3 bitcoins were mined in the fourth quarter, averaging 18.68 bitcoins per day.
The average mining cost for the full year (excluding miner depreciation) was $79,707 per bitcoin, and for the fourth quarter, it was $84,552;
The all-in sustaining costs were $97,272 and $106,251 per bitcoin, respectively.
As of the end of December 2025, the company has cumulatively produced 7,528.4 bitcoins since entering the bitcoin mining business.
• Strategic Progress:
The company has completed the termination of the American Depositary Receipt (ADR) program and transitioned to a direct listing on the NYSE to enhance information transparency and align with its strategic direction, with a long-term goal of expanding its investor base.
CEO Paul Yu stated: "2025 marked the company's first full year as a bitcoin mining enterprise, characterized by rapid execution and structural reshaping. We completed a comprehensive adjustment of our asset system and established a globally distributed mining network. Additionally, the company introduced a new management team, further strengthening our capabilities and competitive advantage in the digital asset and energy infrastructure space. The completion of the NYSE direct listing and USD pricing also signifies our transformation into a global AI infrastructure company."
"As we enter 2026, the company will continue to optimize its balance sheet structure and enhance operational efficiency and cost resilience through adjustments to the miner portfolio. At the same time, we are advancing our strategic transformation into an AI infrastructure provider. Leveraging EcoHash, we will utilize our capabilities in scalable computing power and energy networks to provide cost-effective AI inference solutions. The relevant site transformations and product development are progressing simultaneously, and the company is well-positioned to sustain its execution in the new phase."
The company's Chief Financial Officer, Michael Zhang, stated: "By 2025, the company is expected to achieve significant revenue growth through its scaled mining operations. Despite recording a net loss of $452.8 million from ongoing operations, mainly due to one-time transformation costs and market-driven fair value adjustments, the company, from a financial perspective, will reduce its leverage, optimize its Bitcoin reserve strategy and liquidity management, introduce new capital to strengthen its financial position, and seize investment opportunities in high-potential areas such as AI infrastructure while navigating market volatility."
The total revenue for the fourth quarter was $1.795 billion. Of this, the Bitcoin mining business contributed $1.724 billion in revenue, generating 1,718.3 Bitcoins during the quarter. Revenue from the international automobile trading business was $4.8 million.
The total operating costs and expenses for the fourth quarter amounted to $4.56 billion, primarily attributed to expenses related to the Bitcoin mining business, as well as impairment of mining machines and fair value losses on Bitcoin collateral receivables.
This includes:
· Cost of Revenue (excluding depreciation): $1.553 billion
· Cost of Revenue (depreciation): $38.1 million
· Operating Expenses: $9.9 million (including related-party expenses of $1.1 million)
· Mining Machine Impairment Loss: $81.4 million
· Fair Value Loss on Bitcoin Collateral Receivables: $171.4 million
The operating loss for the fourth quarter was $276.6 million, a significant increase from a loss of $0.7 million in the same period of 2024, primarily due to the downward trend in Bitcoin prices.
The net loss from ongoing operations was $285 million, compared to a net profit of $2.4 million in the same period last year.
The adjusted EBITDA was -$156.3 million, compared to $2.4 million in the same period last year.
The total revenue for the full year was $6.881 billion. Of this, the revenue from the Bitcoin mining business was $6.755 billion, with a total output of 6,594.6 Bitcoins for the year. Revenue from the international automobile trading business was $9.8 million.
The total annual operating costs and expenses amount to $1.1 billion.
Specifically, they include:
· Revenue Cost (excluding depreciation): $543.3 million
· Revenue Cost (depreciation): $116.6 million
· Operating Expenses: $28.9 million (including related-party expenses of $1.1 million)
· Miner Impairment Loss: $338.3 million
· Bitcoin Collateral Receivable Fair Value Change Loss: $96.5 million
The full-year operating loss is $437.1 million. The continuing operations net loss is $452.8 million, while in 2024, there was a net profit of $4.8 million.
The 2025 non-GAAP adjusted net profit is $24.5 million (compared to $5.7 million in 2024). This measure does not include share-based compensation expenses; refer to "Use of Non-GAAP Financial Measures" for details.
As of December 31, 2025, the company's key assets and liabilities are as follows:
· Cash and Cash Equivalents: $41.2 million
· Bitcoin Collateral Receivable (Non-current, related party): $663.0 million
· Miner Net Value: $248.7 million
· Long-Term Debt (related party): $557.6 million
In February 2026, the company sold 4,451 bitcoins and repaid a portion of related-party long-term debt to reduce financial leverage and optimize the asset-liability structure.
As per the stock repurchase plan disclosed on March 13, 2025, as of December 31, 2025, the company had repurchased a total of 890,155 shares of Class A common stock for approximately $1.2 million.

The US AI Startup Is Loving China's Open Source Model

Three Weeks of the US-Iran War: Who's Making Money, Who's Paying the Bill?

Interpreting Polymarket's Major Update Last Night: Fee Expansion, Self-Regulation, and New Incentives

From Human Application to Intelligent Collaboration: How GOAT Network Builds the Next Generation Digital Economy

CZ Washington Dialogue: Crypto Entrepreneurs are Accelerating Their Return to the United States

Morning Report | Strategy increased its holdings by 1,031 bitcoins last week; Katana Blockchain acquires IDEX; NYSE completes rule change to eliminate trading limits on crypto ETF options

Electric Capital: Tracking 501 types of yield-generating RWA assets, we discovered these patterns

Those who are cut off by AI will not disappear; they will become the creators of the next round of the economy

Stablecoins reshaping cross-border payments in Asia? Strategic panorama and investment opportunity analysis

Zuckerberg is building an AI agent to help him as CEO

