Key ideas from the Sequoia Capital — AI Ascent 2025 conference. The speakers are Konstantin Buhler and Sonya Huang, partners of the foundation.

1. In the coming year, the AI market may potentially surpass the cloud technology market at the time of its rapid growth. Generative models have been integrated into co-pilot scenarios as a new level of participation. AI no longer helps humans — it works instead.

2. Generative AI has achieved an ideal Product Market Fit and demonstrates high user retention rates. AI applications are becoming an integral part of everyday life, which indicates the maturity of the technology and the market's readiness for their large-scale implementation.

3. New physics of propagation. The speed of AI adoption is unprecedented. Previously, large-scale implementation of new technologies took years. Today, everything is different. An example of a ChatGPT product was launched globally at the end of 2022, and in a matter of months it became widespread. This is what the new "physics of propagation" looks like.

4. Economics of agents. AI systems are learning to operate 100% autonomously. They make transactions. They manage resources. They interact with each other. The agent economy is emerging today. This is no longer the future. Here, those who combine the power of AI with deep subject expertise embedded in business processes will benefit. This launches the economy of “vertical agents".

5. The important term is Data Flywheel. Sequoia calls this a loop. What it is: the more users an AI product has → the more data it collects → the better the models/algorithms are trained → the higher the quality of the product → the more new users are attracted → and the cycle repeats in a circle. Sequoia argues that Data Flywheel should not only enhance the AI model, but also instantly improve the product itself based on this model.

6. Significant investments in infrastructure are already needed to realize the agent economy: the construction of new data centers, the development of protocols for secure interaction of AI systems. It is also important to ensure the ability of agents to communicate seamlessly with each other.

7. An AI company is just a company. 95% of AI startups operate according to the classical model: team, product, market. But 5% of companies go the other way. They will define the industry, but there will also be increased demands on them. Their models should not only be of high quality, but also initially cost—effective. The business result generated by the models should be measured by metrics. They need to move away from "vibe revenue" and generate real revenue.

8. There are more and more AI companies. But there are still enough opportunities. The main revenues go to companies that operate "at the application level." Where technology solves specific tasks. Vertical agents and vertical solutions. Deep niche expertise and a clear understanding of its uniqueness.

9. New type of thinking. Not only the market is changing, but also the logic of decision-making. From confidence and knowing how things should be, to managing in complete uncertainty. The future requires stochastic thinking.

10. The future of AI is agents. Programs that operate autonomously, interact with each other, exchange data, and make decisions. This requires new approaches to identification, interaction, and security.

11. Recommendations from Sequoia. - Focus on vertical solutions. Develop AI applications focused on specific industries. This will achieve higher efficiency and meet the specific needs of customers.- Invest in infrastructure. Develop your own scalable and reliable infrastructure.- Ensure trust. Develop security protocols and transparency of AI systems. This will help to gain the trust of users and regulators.The final conclusion. AI is the new logic of economics. And now is the perfect moment to enter it consciously, with an understanding of strategic directions.

12. The expected annual volume of the AI market is $1 trillion.

p.s. What made me happy:

1. The word "future" is increasingly entering the lexicon of the leaders of the largest companies. Foresight ceases to be optional and becomes strategically applied.

2. Sequoia Capital's forecast was released on May 2. Our image of the future of business and education in 2030 was developed at the end of January (for the trendbook with trends in education). Moreover, the coverage and depth are noticeably wider. Although it already needs to be updated again, because AI is developing much faster than predicted.

3. The global market for AI products seems saturated. But the economic system that AI is redesigning is incomparably larger. There is still more than enough space to launch. That's what they say from the stage

#sequoia #AI