Tajikistan and Iran have agreed to extend the repayment period for Iran’s investment in the Sangtuda-2 hydroelectric power plant by six years and four months. The extension will allow the Iranian side to recover its investment, after which full ownership of the facility will transfer to the Tajik government.
The amendments to the electricity purchase agreement between Tajikistan’s state-owned power utility Barki Tochik and Iran’s Sangtuda Sangob company were ratified by the lower house of Tajikistan’s parliament in early October. The revised agreement was originally signed on May 29, 2025, in Dushanbe, following high-level negotiations between the two countries.
Under the initial agreement, Iran was to recover its investment by August 2026, after which the hydropower plant would become Tajik state property. However, under the new terms, Iranian management of the plant will continue until the end of 2032.
The total construction cost of the Sangtuda-2 hydroelectric plant was $256 million. Of this, the Iranian government contributed $180 million, Sangob invested $36 million, and the Tajik government provided $40 million.
According to Tajikistan’s Ministry of Energy and Water Resources, between the plant’s launch in 2012 and the end of 2023, Barki Tochik purchased 8.9 billion kilowatt-hours of electricity worth $451.5 million. However, the power plant itself received only $122.5 million, roughly 27% of the total value.
As a result, Tajikistan’s outstanding debt to Sangtuda-2 reached $329 million. Under the new agreement, this debt will be fully written off. In return, Dushanbe has committed to a new payment schedule over the next six years and four months. After this period, control of the plant will transfer to Barki Tochik.
Sangtuda-2 is among the largest joint energy projects between Tajikistan and Iran. Construction began in 2006, with the first hydroelectric unit launched in September 2011 and the second in 2014.
The plant is located in the Danghara district of the Khatlon region, on the Vakhsh River. It is the fifth stage of the Vakhsh cascade of hydroelectric stations.
Raymond Zhao is the CEO and cofounder of Structured AI.
Courtesy of Raymond Zhao
Raymond Zhao thought working at Goldman Sachs was his dream until he interned there.
While studying at Oxford University, he found a new dream: cofounding an AI startup.
Now, his company, StructuredAI, is in Y Combinator’s Fall 2025 batch.
This as-told-to essay is based on a conversation with Raymond Zhao, 23, the cofounder and CEO of Structured AI, which is in Y Combinator’s Fall 2025 Batch. His internship, funding, and acceptance to YC have been verified by Business Insider. Golman Sachs declined to comment when contacted by Business Insider.
This piece has been edited for length and clarity.
AI, on the other hand, feels exciting. Many of the great AI companies of the next decade are being built right now, and I want to be a part of that.
When I spoke to friends trying to break into the industry, however, I kept hearing about this weird dilemma: in some cases, it was easier to raise money for a startup than land an entry-level job in AI.
I skipped the graduate job hunt and cofounded Structured AI. We’re now part of the Y Combinator Fall 2025 Batch and, including the $500,000 we got through YC, have raised about $1 million in pre-seed funding. Every day, I feel energized.
I thought working at Goldman Sachs was my dream
In 2021, I moved to the UK from Australia to study mathematics and statistics at the University of Oxford. I wasn’t sure what career I wanted to pursue. I had wanted to be an architect, but after doing work experience at a firm when I was 18, I realized it wasn’t as exciting as I’d hoped.
Student societies were a great way to get an insight into different industries, and the idea of working in finance sounded attractive: good pay, respect, and smart, hardworking colleagues. As a math student, it felt like the most prestigious route I could follow, and many of the people I looked up to worked at Goldman Sachs.
I applied for the summer 2024 internship at Goldman Sachs in London. It was highly selective, involving an online test, a virtual interview, a superday of back-to-back conversations, and a follow-up interview. When I finally got accepted, I was really excited, but within a month, it became clear that I wasn’t passionate about the work.
I worked long hours, sometimes staying in the office until midnight and occasionally coming in on weekends to finish work. Other interns cared deeply about financial markets and did research in their spare time, but I didn’t.
Raymond Zhao studied for four years at the University of Oxford.
Courtesy of Raymond Zhao
After finishing the internship, I spent some time backpacking in Southeast Asia and thinking about what excited me. In 2024, I returned to Oxford for my Master’s and spent the year figuring out what to do next.
As a member of the venture capital society, I met my future cofounder, Isabel Greenslade, and a network of VCs and founders. I realized I wanted to launch a startup so I could build something myself and steer the boat.
In the final year of my Master’s, I joined the AI society. Isabel and I spent the year ideating and pivoting until we eventually found a working idea. Through the AI society, we met our third cofounder, Brandon Smith, who discovered a problem in his contracting work that sparked the idea of StructuredAI.
We decided to build a copilot for the preconstruction engineering world. Engineers spend a lot of time on very repetitive manual workflows, a lot of which are text-based, like paperwork and building forms. We built AI agents to assist them.
I didn’t expect to raise $500,000 for our AI startup so quickly
When I was close to graduating with my Master’s, I wasn’t job-hunting. I just wanted a startup idea to work. We applied to Y Combinator for the summer 2025 cohort, but weren’t accepted.
During my final week at Oxford in June, I skipped classes and went to San Francisco for an AI startup conference. I hadn’t planned to raise money, but my founder friends encouraged me to try while I was there. I started talking to investors, and we managed to raise about $500,000.
Making Structured AI financially viable so quickly was a huge surprise, and we were able to focus on it full-time. We spoke to dozens of engineers to better understand the problem we were facing and build a sales pipeline. I think that helped us with our second application to YC.
Raymond Zhao, Isabel Greenslade, and Brandon Smith (left to right) behind a Y Combinator sign.
Courtesy of Raymond Zhao
We started YC’s Fall Batch in September. There are a lot of young founders, some under the age of 21, and some who dropped out of university. It’s an amazing, super-ambitious community.
People warned us YC would be intense, and it is. It’s a real sprint, and everyone works around the clock. We’ve turned our living room in our Airbnb in San Francisco into an office, with three monitors and extending desks.
On a usual day, we wake up anytime between 4 a.m. and 9 a.m., get coffee, and work for the whole day. We’re going to conferences, taking calls, and going to events in the evening. I’m not sleeping that much, normally six hours, and I need a solid eight.
At the end of the cohort, in early December, we’ll do Demo Day, where pitch our progress onstage to a group of investors. The pressure is real, but good work comes out of it.
To me, the biggest risk is not taking one
If you have an idea for a startup that’s keeping you up at night, I recommend trying to launch it. If you can make it work and have the appetite for risk, startup life can be very rewarding.
For me, it was a calculation between building something in AI and the returns going directly to me, versus taking a corporate job, where you can wait years for a promotion.
YC is one of the most exclusive clubs in the startup world. I’m living a dream — one that came to me later down the line. Now, the next one is to build a massive company.
