Inside the company that helped build China’s equity options market
Fintech firm Bachelier Technology on the challenges of creating a trading platform for China's unique OTC derivatives market
From Wall Street to Shanghai: The Story of Bachelier Technology
Ten years ago, in February 2015, a group of friends working as traders and quants in New York saw a lightbulb moment. The Shanghai Stock Exchange had just launched China’s first listed options contract, the SSE China 50 ETF, and a bull market was brewing. They foresaw an opportunity.
“We were talking about Chinese capital markets,” says Tuohua Wu, vice-president at Bachelier Technology, at the time a derivatives trader at Citigroup. “There was a bull market coming on by the end of 2014 and, more importantly, there was news regarding the first derivative listed product.”
The group reasoned that a listed derivatives market in China would eventually lead to a larger over-the-counter (OTC) market. However, Chinese banks and securities firms lacked the technology and data infrastructure to support it.
“We thought we can leverage our experience in this industry to help China build this market by providing IT solutions,” says Baolin Shao, Bachelier’s chief architect, then a vice-president at Morgan Stanley.
A year later, the friends had returned to China and founded Bachelier Technology (known as Tongyu in China) in Shanghai, aiming to provide the essential pricing and trading services for China's burgeoning OTC derivatives market.
Fueling the Derivatives Boom
Fast forward to today, and China's equity options market is booming. Since launching its first product in 2017, Bachelier Technology's suite of front-to-back trading and risk solutions has become a technological backbone for the market's development.
Beginning in 2018, the OTC equity derivatives market in China, including a retail structured products market centered on "snowballs" (equity index-linked autocallable products), entered a period of rapid growth. This was enabled by a new regulatory framework from the Securities Association of China.
Today, Bachelier Technology's client list is a who's who of the market:
All eight tier-1 securities houses in China permitted to offer single-stock options are clients.
Other notable clients include the China securities ventures of UBS and Morgan Stanley.
Now, the company is looking to expand beyond mainland China, starting with Hong Kong.
“Hong Kong is our first stop,” Shao says. “We’ve acquired a dozen clients in Hong Kong already, both sell side and buy side, and we’re helping them to do their cross-border businesses.”
A 'Tailored for China' Strategy
Bachelier believes its edge over larger, established tech vendors lies in its focus on the specific needs of Chinese institutions.
Shao explains that while global vendors focus on pricing and risk management, the Chinese market is at an earlier stage.
“In China... people’s focus is more on operation and workflow. That comes first, before pricing and risk,” he says.
The company also sees an advantage in China’s Xinchuang initiative (Information Technology Application Innovation, or ITAI), a government-led plan to domesticate IT infrastructure and reduce dependence on foreign technologies.
“Xinchuang requires China’s homemade servers, operating systems, database, and of course those global vendors won’t do any customisation for that,” Shao says.
Because Bachelier's technology is built mainly on open-source technologies and uses databases like Postgres, MySQL, and various local Chinese databases, it is less impacted by these guidelines.
While facing onshore competition from firms like Hundsun Technologies, Shao believes the bigger competitors are the in-house IT teams of Chinese financial institutions. He argues that Bachelier's comprehensive grasp of market requirements will help overcome the bias towards in-house development.
Bachelier now has a team of about 200 people, mostly developers, across Shanghai, Hangzhou, Beijing, and Shenzhen.
Solving the 'Data Gap'
From its inception, Bachelier recognized a major missing piece for the OTC market: data. After spending three years building its IT infrastructure, the firm set out to create a derivatives data platform, replicating the model of Markit Totem (now part of S&P Global).
The business model involved aggregating OTC derivatives pricing data from participant banks and selling benchmark prices back to them. Convincing local firms to share their data was initially a challenge.
“When we first started to talk about OTC data, some of the local broker-dealers didn’t understand," Wu admits. "They said, ‘You ask me to contribute to the data, and then I pay for the data subscription? That’s a strange business model.’ We explained to them how this logic works, and they will get to understand the importance of OTC derivative data.”
Today, the firm harvests, cleans, and analyzes pricing information from over 20 dealers. This service allows contributors to benchmark their pricing and provides the market with crucial insights.
“We can back out the whole volatility surface for major underlyings like equity indices, as well as most of the commodities," says Wu. "That’s helpful not only to the dealers, but also to buy-side companies."
Looking Ahead: AI and Buy-Side Solutions
Bachelier is continuing to innovate. Currently in development is a portfolio management system for the buy-side, an area Shao believes is generally underserved by fintech in China.
“Our goal is, can we provide to buy-side clients a better product, good service, and at a low cost? That’s our angle to the global market,” Shao says.
The company has also integrated artificial intelligence into its workflow, using models like DeepSeek’s R1 and OpenAI’s o1. It developed a ChatGPT-style interface that allows traders to make complex requests using natural language.
“A trader can say, ‘I want to calculate the delta and gamma from my book by shifting the vol surface by 2%’. This chat can directly translate into the API of our pricing system and execute that,” Shao explains.
Shao notes that for coding tasks, they don't need to post-train the models. Instead, they provide the AI with specific requirements and relevant code examples. For instance, to integrate a new product like an autocallable structure, they give the AI a vanilla option implementation as a template.
“AI is basically a fancy data interpolation machine,” Shao says. “For this type of capability, you don’t need any post-training.”