Are Z.ai and MiniMax heading down opposite financial paths months after Hong Kong IPOs?

While Z.ai and MiniMax, two prominent Chinese AI companies, shared a common objective of capitalizing on the burgeoning demand for artificial intelligence domestically and internationally following their Hong Kong IPOs in January, their financial performances following the IPOs appear to diverge significantly. Z.ai, headquartered in Beijing, has been garnering praise from market analysts due to its significant revenue growth and superior model performance.

Conversely, MiniMax, its Shanghai-based counterpart, is grappling with skepticism over its lagging technical benchmarks and projected growth. In the first half of the year, Z.ai reported a staggering 400% increase in revenue to 953.9 million yuan (approximately USD 142 million), while MiniMax's revenue expanded by 283% to USD 116.6 million.

Upon closer inspection, the disparity is even more pronounced when evaluated using annual recurring revenue (ARR), a standard metric utilized by software firms to forecast revenue over a 12-month period based on present monthly subscriptions. Z.ai's ARR reached an impressive USD 1.6 billion based on the results of August 2023, whereas MiniMax's ARR stood at USD 800 million – precisely half of Z.ai's.

However, MiniMax's founder and CEO, Yan Junjie, later admitted that the ARR figure was derived from a single week in August multiplied by 52, a method that, if applied to Z.ai's figures, would yield an ARR of USD 2 billion. Technology and industrial policy analyst Tilly Zhang labeled MiniMax's calculations as "creative," suggesting that the company might have emphasized a traffic surge following the launch of its flagship H3 video model on July 31.

Doubts about MiniMax's prospects have been further exacerbated by its model performance. According to the Artificial Analysis Intelligence Index, MiniMax's latest flagship M3 model scored 45, trailing behind both Z.ai's GLM-5.3 and Chinese startup Moonshot AI's Kimi K3, which both scored 60. HSBC analysts, including Ritchie Sun, drastically reduced their price target for MiniMax from HK$760 (USD 96.92) to HK$330, citing the need for the company to invest more to remain competitive.

Similarly, JP Morgan analysts, including Olivia Xu, described MiniMax as "in catch-up mode" concerning model capabilities. Market sentiment towards Z.ai, on the other hand, has been more favorable. CMB International analysts, including He Saiyi, raised their price target for Z.ai by 32% to HK$1,985, highlighting a balanced combination of model consistency, raw intelligence, and task cost.

Macquarie analysts, including Ellie Jiang, also recognized Z.ai's advanced capabilities and frontier positioning, albeit reducing their price target by 21.5% to HK$1,693, taking into account the broader market volatility. Despite this divergent financial trajectory, both firms continue to incur significant losses, primarily attributed to their investments in model development and compute capacity.

Z.ai's research and development expenditures increased by 33.6% to 2.13 billion yuan, while its gross profit margin shrank to 26.4% from 50% the previous year, resulting in an adjusted net loss that widened by 12.1% to 1.96 billion yuan. MiniMax, on the other hand, saw its research and development expenses more than double to USD 296.9 million as it endeavored to establish an "independent and controllable" computing cluster, partially powered by Chinese chips.

While its gross margin improved to 17.9% from 12.1%, its adjusted net loss surged by 111.2% to USD 293 million. Both companies are also contending with broader structural challenges, such as a "very real compute shortage," which has compelled them to embark on extensive data center capital expenditure plans. Although they lack the competitive edge or resources to directly compete with Chinese tech giants like Tencent Holdings, ByteDance, and Alibaba Group Holding, which own the South China Morning Post, they have demonstrated "meaningful commercial scale" and differentiated themselves through "faster execution, deeper focus, and stronger product iteration," according to internet and media analyst Alex Liu of Bank of America Global Research.