Skip to main content

Hog Analyst: AI Got Him Wrong in November...But AI Was Right After All?

Five months ago, one of China's leading pork analysts vociferously denied making a dire prediction that Chinese hog prices would crash during 2026. Prices are now tumbling to levels unseen in decades, and the bearish prediction he had disavowed now appears to have been more or less accurate. Why did he disown a bearish forecast? What does this incident say about the China dream of AI-driven agriculture? 

On November 24, 2025, long-time Chinese hog analyst Feng Yonghui posted an article on his soozhu.com web site denying that he had predicted that hog prices would fall below RMB 10-to-11 per kg and stay there during 2026. Mr. Feng disavowed the bearish forecast, calling it a "malicious spreading of rumors."  He blamed an artificial intelligence algorithm for splicing together comments he had made out of context and constructing a misleading assessment that tarnished his reputation. 

Five months later, on March 23, Mr. Feng's site reported that hog prices had broken through the RMB 10 threshold, dropped 26% so far this year, and were still under downward pressure. On March 27 his site reported an average of RMB 9.58 per kg. Prices may well rebound above RMB 11 this year, but the drop in prices is alarming, and there is consensus that hog supply and demand are seriously out of whack. So, sounding an alarm last November was not really irresponsible.

In November Mr. Feng had claimed that the forecast of low prices was inconsistent with supply & demand fundamentals, including signs that sow inventories were being pared back. This week his site reported that the sustained decline in prices is due to a severe imbalance between supply and demand. A high volume of pork is being moved into cold storage because demand is at a seasonal low during the early-Spring, and farms are giving up on expectations of a price rebound and sending hogs to slaughter. Another item this week concluded that large-scale hog companies had consistently failed to meet targets for reducing production issued by authorities. Government officials held meetings with hog companies again this month taking even stronger measures to force them to cut back sow inventories, control the volume of hogs slaughtered, and report company data to authorities. 

In November, Mr. Feng had suggested that production costs would act as a floor preventing a steep drop in prices. This week, Mr. Feng's site said the drop in prices caused a predicament of unremitting losses for the entire sector, with farrow-to-finish operations losing nearly RMB 200 per hog and feeder-to-finish operations losing more than RMB 300 per hog.

Mr. Feng is one of China's most experienced and incisive hog market analysts, and his commentary is usually free of propaganda. So it's out of character for him to get this wrong. It's odd that his November disavowal of the low-price prediction for this year is still featured on his site literally alongside news about this month's plunging hog prices. His November accusation of "spreading rumors" (造谣) -- a favorite phrase of communist apparatchiks -- is a clue that his November diatribe may have been published under pressure from an official behind the scenes intent on squelching bad news. Ironically, it appears that Feng's November essay describing bearish forecasts as false information was itself false information. 

Mr. Feng's November accusation that an AI algorithm assembled his comments "out of context" to spread misleading information gives us another thread to pull on. Soozhu.com also features an article in which New Hope Group founder Liu Yonghao urges hog farms to embrace AI to follow the CCP's "Number One Document's" order to integrate AI in agriculture. But AI's autonomous gathering and analysis of information is at odds with the CCP's obsession with actively controlling the narrative. If the CCP succeeds in scrubbing any fact or statement that is at odds with its accepted narrative (e.g. bad news or evidence that policies failed), an AI algorithm will only be able to utilize factoids and statements that agree with the CCP narrative. Consequently, AI will reproduce the CCP's narrative.
Data from xinfadi.com.cn


Comments

Anonymous said…
I'm curious to see this use of AI as a scapegoat - it's brilliant in a way, since it's a machine that does not have social power over the writer nor does blaming it directly hurt any humans (who might have been blamed in other circumstances.)

I'm not expecting AI to move the needle on official narratives, but would be curious if it becomes a rhetorical device - especially when someone posts a view that's outside the boundaries and deletes it later - acting as a fig leaf when someone's more disruptive opinions leak out.
Dim Sums said…
Great, fascinating insight
Anonymous said…
Blaming AI sounds like the 2026 equivalent of "I've been hacked"

Popular posts from this blog

Xi Jinping's Doctoral Thesis

Xi Jinping is the vice president and presumed next president of China but little is known about him. In this post the dimsums blog offers its contribution to the genre of Xi Jinping-ology by conveying Xi's decade-old views on agricultural markets. Ten years ago Xi Jinping wrote a thesis, "Tentative Study of Agricultural Marketization" (中国农村市场化研究) for a Doctor of Law degree at Tsinghua University in Beijing, a top breeding-ground for Chinese officials. The dimsums blogger has spent several hours poring over the 200-plus page tome to see what it reveals about Dr. Xi. The thesis is remarkably close to what China has been doing lately in agricultural policy, suggesting that Xi (or the person who actually wrote the thesis) has a major say in policy or is at least in agreement with what's being done. There is nothing adventurous, controversial (or insightful) in the thesis. It seems to be the work of a wonkish technocrat who is not prone to talk out of turn or wander from...

China's 2024 Ag Imports Shrank in Value

China's agricultural imports declined 7.9 percent during 2024 to reach $215 billion, according to data posted on the customs administration website. The 2024 value was lower than each of the 3 preceding years. Agricultural exports were up 4.1 percent to reach $103 billion. Source: Data from China Customs Administration December reports. The top two agricultural import categories by value both declined. Soybeans ($52.75 billion in 2024) fell 10.9 percent, and meat ($23.38 billion) fell 15.1 percent. Cereal grain imports ($15 billion) were down 28 percent and fish & shellfish imports ($18.5 billion) were down 6.2 percent. Edible oils imports ($10.6 billion) were down 17.8 percent. Fruit, rubber, cotton and wool and beverage imports were up for the year. The decline in value of imports partly reflected a decline in prices. Customs reported that the volume of soybean imports for calendar year 2024 reached a record 105 million metric tons, up 5.6 million metric tons from the previou...

Feed Boom & Cratering Grain Imports; China Leaves Us Guessing

In the first half of 2025 China increased its meat and egg production by a combined 1.58 million metric tons (mmt) from a year earlier, a moderate increase of 2.5%. Meanwhile, animal feed output during H1 2025 compiled from feed industry association reports increased by 14.5 mmt (+10 percent) from a year ago. China's 14.5-mmt increase feed output growth outpaced the 1.58-mmt growth in meat production by a ratio of 9:1. It's hard to make sense of these inconsistent figures.  [note: The June 2025 feed industry association report has a 7.7% yoy growth rate for feed output which is inconsistent with the 10.1% growth shown here calculated by comparing data from monthly reports issued last year. Growth rates for complete feed were 8.1%, concentrates -1.5%; additives 6.9%. These inconsistencies are common in the feed industry association reports, a reason for doubting the accuracy of this data.] There is no boom in demand for feed ingredients to fuel a huge increase in feed production...