Borrow conditions in Knowledge Atlas Technology (2513 HK) have tightened rapidly as the stock’s post-IPO rally drew heavy short positioning. EquiLend data shows utilization moved from a negligible 0.3% on February 24 to 70% within three trading days and has reached as high as 96.8%, having held above 90% since the start of April.

EquiLend lending activity shows demand building quickly across market channels. In parallel with the surge in lender-to-broker activity, EquiLend Broker-to-Broker Borrow Quantity has climbed to around 7.5 million shares, more than 2x the lender-to-broker count. As a result of sustained demand, EquiLend Cost to Borrow has remained around 3,000 bps throughout, underscoring the intensity of investor demand to attain short positions.

Knowledge Atlas debuted in Hong Kong on January 8 at HK$116.20, marking China’s first major large language model developer to go public. The stock then staged a remarkable post-IPO rally, aided by heightened AI momentum, including a public endorsement from Nvidia CEO, Jensen Huang, tied to the company’s OpenClaw platform in March, and a first earnings report as a public company indicating revenue doubling year-on-year. The stock reached HK$1,150 on May 13, representing a 780% gain from its IPO price.

As the rally extended, short sellers pressed into the name, and investor focus shifted toward fundamentals and governance signals. In addition to the heightened positioning since late February, bears have pointed to a director resignation on May 6 and persistent net losses of CN¥1.9 billion as ongoing sources of concern.

With lendable supply tightening, the setup has become increasingly sensitive. With only 290,000 shares remaining of a 3.2 million share lendable pool, EquiLend data suggests borrow conditions remain near capacity and volatility risk may stay elevated as the story develops.