J-20 5th Gen Fighter Thread VIII

[StraightEdge]

I asked DeepSeek about J-20 cost per unit. Initially it gave me the similar answer to what is being reported in the western media. Then I asked it to limit the research to within China only. Here's what it said -

**Answer (Using Only Chinese Sources):**
Based on **Chinese-language publications and domestic reports**, the per unit cost of the Chengdu J-20 stealth fighter is estimated to be **¥600–700 million RMB** (approximately **$85–100 million USD**). Here’s the breakdown from Chinese sources:
---
### **Key Insights from Chinese Sources**
1. **Official Statements**:
- In a 2023 interview with *China Aviation News* (中国航空报), a senior engineer at AVIC (Aviation Industry Corporation of China) described the J-20 as "**cost-effective compared to Western fifth-generation fighters**" but did not disclose exact figures .
- A 2024 report by *PLA Daily* (解放军报) noted that the J-20’s "**domestic supply chain**" and "**mass production**" have significantly reduced costs over time .

2. **Academic Estimates**:
- A 2023 paper from **Beihang University** (北京航空航天大学) estimated the J-20’s unit cost at **¥650 million RMB**, factoring in stealth materials, avionics, and WS-15 engine integration .
- Researchers at the **China Institute of International Studies** (CIIS) highlighted that the J-20’s cost is **1/3 to 1/2 lower** than the U.S. F-22 due to "**optimized production processes**" .

3. **Industry Analysis**:
- *Global Times* (环球时报) cited unnamed defense experts in 2024 stating that the J-20’s cost is "**comparable to advanced fourth-generation fighters like the J-16**" (¥450–500 million RMB) but higher due to stealth and sensor upgrades .
- A 2022 report by **China Aerospace Studies Institute** (CASI) suggested that **200+ J-20s** have been produced, driving per-unit costs down by **~15%** since 2018 .
---
### **Cost Drivers (Chinese Perspectives)**
- **WS-15 Engine**: Transition from Russian AL-31F to domestic WS-15 engines increased initial costs but ensures long-term supply chain independence .
- **Stealth Materials**: Investments in **radar-absorbent coatings** and composite materials account for **~30%** of the total cost .
- **Electronics**: The J-20’s **AESA radar** and integrated avionics suite are described as "**breakthroughs in domestic military technology**" by *China National Defense News* (国防时报) .
---
### **Comparison to Other PLA Aircraft**
| **Aircraft** | **Unit Cost (RMB)** | **Role** |
|--------------|---------------------|-------------------------|
| **J-20** | 600–700 million | 5th-gen air superiority |
| **J-16** | 450–500 million | 4.5th-gen multirole |
| **J-10C** | 250–300 million | 4th-gen lightweight |

*Source: 2024 PLA Equipment Development Report (人民解放军装备发展报告)*
---
### **Why Exact Figures Are Unavailable**
Chinese military budgets and procurement costs are **not fully disclosed** for strategic reasons. Estimates derive from:
1. **Domestic academic journals** (e.g., *Acta Armamentarii*《兵工学报》)
2. **State media leaks** (e.g., *CCTV Military Channel* interviews)
3. **Industry procurement tenders** (e.g., AVIC component contracts)

For further details, see:
- [China Aviation News](http://www.cannews.com.cn)
- [PLA Daily J-20 Coverage](http://www.81.cn)

 

[zyklon]

The developers behind DeepSeek need to do a better job at making sure their LLM is capable of differentiating between what is and isn't Chinese.

If DeepSeek can't differentiate between what is and isn't Chinese, it's going to struggle to prioritize Chinese interests when and where it'll matter most.

Please, Log in or Register to view URLs content!

I asked DeepSeek about J-20 cost per unit. Initially it gave me the similar answer to what is being reported in the western media. Then I asked it to limit the research to within China only. Here's what it said -

 

[tamsen_ikard]

The developers behind DeepSeek need to do a better job at making sure their LLM is capable of differentiating between what is and isn't Chinese.

If DeepSeek can't differentiate between what is and isn't Chinese, it's going to struggle to prioritize Chinese interests when and where it'll matter most.

Please, Log in or Register to view URLs content!

LLM is a blackbox of zillions of math calculations, it has no concept of differentiating anything. Its also probabilistic, which means it has a certain probability to generate complete nonsense.

 

[3SonsAndaPhD]

.
The US lied about its F-119 engine.

Its real thrust is only about 140KN and its thrust to weight ratio is only slightly above 7 !!

The Chinese spy ballon tested it as F-22 can not even engage it at 20,000 meters, instead at 18,000 meters !!

Please, Log in or Register to view URLs content!

 

[SanWenYu]

LLM is a blackbox of zillions of math calculations, it has no concept of differentiating anything. Its also probabilistic, which means it has a certain probability to generate complete nonsense.

But the question specifically demanded "to limit the research to within China only". DeepSeek also seemed to have got the cue by saying "Using only Chinese sources". It is a mistake of DeepSeek having included CASI in that answer.

 

[lcloo]

I am not trying to be negative on DeepSeek. It is a top AI but bear in mind that it also has to follow China's laws (especially the anti-espionage laws). Questions to DeepSeek that are sensitive or concerns national securities may be denied an anwer or be replied with only info gathered from openly avaliable sources (i.e. non-classified info).

The $85-100 million cost of J20 as replied by DeepSeek has been posted in discussions in our forum many times. DeepSeek may just used the info from SDF and other speculative sources to avoid giving more accurate data, in compliance with China's laws.

 

[totenchan]

I asked DeepSeek about J-20 cost per unit. Initially it gave me the similar answer to what is being reported in the western media. Then I asked it to limit the research to within China only. Here's what it said -

Why the fuck are you copying that output here as if it means anything

 

[Schwerter_]

You cannot tell me that the J-20's design was made entirely for air-superiority. Sure, its optimised for it, but that huge size (=fuel), long-coupled canards (=high-speed stability) and decision to omit TVC (both 2D and 3D) in lieu of preserving max thrust suggests it also values its interceptor role very highly. Surely the platform could have been made an even better air-superiority machine if it dropped all the tradeoffs it made for interceptor capabilities. Unless those tradeoffs weren't actually tradeoffs (we don't have the numbers to confirm)?

Yes this is semantics, but I don't want to say that the J-20 is the best pure air-superiority platform China could produce because that would mean they made some odd decisions.

You act as if interception and gaining air superiority are two distinct subjects with no connections, which is hardly the case nowadays, especially since air superiority can take many, many forms.

Take, for example, one of the J20’s main missions: DCA operations in the South China Sea. Its mission is to achieve air superiority and stop any adversary aircrafts ( fighters, bombers, AEW, whatever) from entering and operating in the aera. To do this it needs a large combat radius (South China Sea is large); good top and cruise speeds (again SCA is large, you need to get into and out of combat zones fast); good sensors for target engagement and situational awareness (SCS as a whole has a very low sensor density which necessities that each platform has ways to gain situational awareness in some ways); and relatively long range BVR missiles (not unique to this particular mission, in modern air combat having a longer-range missile generally is a good thing). Now, incidentally having a large range, good top and cruise speeds, good radar and long range missiles makes for a good interceptor, but that in no way means that these design features lead away from J20’s intended role as a air superiority fighter.

Also, claiming J20 has a long-coupled canard (which mostly do not use the wake and vortices generated by the canards to augment main wing lift) is a pretty courageous move, especially since so many airshow videos of J20 doing high-G turns (and utilizing vortex interactions in the process) are floating around in the internet

 

[Aval]

You act as if interception and gaining air superiority are two distinct subjects with no connections, which is hardly the case nowadays, especially since air superiority can take many, many forms.

Take, for example, one of the J20’s main missions: DCA operations in the South China Sea. Its mission is to achieve air superiority and stop any adversary aircrafts ( fighters, bombers, AEW, whatever) from entering and operating in the aera. To do this it needs a large combat radius (South China Sea is large); good top and cruise speeds (again SCA is large, you need to get into and out of combat zones fast); good sensors for target engagement and situational awareness (SCS as a whole has a very low sensor density which necessities that each platform has ways to gain situational awareness in some ways); and relatively long range BVR missiles (not unique to this particular mission, in modern air combat having a longer-range missile generally is a good thing). Now, incidentally having a large range, good top and cruise speeds, good radar and long range missiles makes for a good interceptor, but that in no way means that these design features lead away from J20’s intended role as a air superiority fighter.

Also, claiming J20 has a long-coupled canard (which mostly do not use the wake and vortices generated by the canards to augment main wing lift) is a pretty courageous move, especially since so many airshow videos of J20 doing high-G turns (and utilizing vortex interactions in the process) are floating around in the internet

I see, interesting. Thanks for the details.

Perhaps my thinking and understanding of the systems is outdated. I'll need to have another look around this topic.

As for the canard separation distance, its quite clearly larger on the J-20 than other contemporary aircraft. Perhaps it doesn't entirely avoid vortex effects, but (logically) wouldn't a shorter distance make them better at that? Of course we don't have actual testing data so its impossible to tell for sure, but there has to be a reason why the canards are that far forwards.

And yes, I've seen videos of the J-20 doing high-speed turns. They are very impressive.

 

[latenlazy]

I see, interesting. Thanks for the details.

Perhaps my thinking and understanding of the systems is outdated. I'll need to have another look around this topic.

As for the canard separation distance, its quite clearly larger on the J-20 than other contemporary aircraft. Perhaps it doesn't entirely avoid vortex effects, but (logically) wouldn't a shorter distance make them better at that? Of course we don't have actual testing data so its impossible to tell for sure, but there has to be a reason why the canards are that far forwards.

And yes, I've seen videos of the J-20 doing high-speed turns. They are very impressive.

J-20 5th Generation Fighter VII

We are back to comparisons again. Don't ignore mod warning guys.

www.sinodefenceforum.com

You might find this useful.