Software Bidding Is Reaching Its Breaking Point
A couple of days ago I had dinner with a friend who works in government-enterprise software. He said he's submitted seven or eight bids this year, won one, and the quoted price was pressed down to a third of what it used to be. It's not that the client is deliberately lowballing—there really are people bidding at that price, and what they're offering looks decent enough.
He asked me what's going on. I told him to go look up how fast AI can write bids and build software now.
This got me thinking for a long time. The software industry's bidding system probably can't hold on much longer.
What the Old System Ran On
China's government procurement market is massive. In 2024, total government procurement nationwide exceeded 3.37 trillion yuan, with open tenders accounting for 76.63%. IT-related government procurement was roughly 143.3 billion yuan. Although this dropped 4.5% year over year, the number of projects actually grew 21.5%—nearly 79,000 projects in total. In other words, there are more projects, but each one is getting smaller.
In the past, software bidding operated stably because of a few barriers to entry.
First, qualifications. These ensure bidding companies are legitimate, with corresponding capabilities and experience. But this barrier has never blocked too many people—companies that want in always find a way. Make it too strict and you lose the point of competition.
Then there's information asymmetry. Many companies only find out about a tender after it's been issued, leaving insufficient preparation time, or they never hear about it at all. The window between issuance and deadline is limited, which objectively keeps some competitors out. Even if they do know, producing a decent bid document takes significant time and manpower. Companies that aren't fully prepared lose out during the technical evaluation.
Finally, commoditized comparison. This is the core design logic of bidding: treat software as a standardized commodity, write out the functional parameters clearly, and score item by item. If you haven't built similar systems before, you can't produce many of the required functions, you can't explain the technical details, and you have nothing to demo.
Combined, these barriers kept competition within a relatively controllable range. For a long time, the system worked reasonably well.
The Barriers Are Crumbling at the Same Time
AI is simultaneously loosening all of these barriers.
Qualifications need no explanation—they were never a real moat to begin with.
The fastest change is in information asymmetry. Quite a few companies are already using AI tools to scan tender information across the entire web. Bailian Intelligence's "Zhiliao Biaoxun" covers over 100,000 tendering websites nationwide, accumulating more than 300 million tender records and updating 20 million daily, with AI-predicted win rates exceeding 70%. Qianlima Tendering Network integrated large models like DeepSeek in 2025, updating 300,000 tender notices daily, with AI monitoring the whole process from project approval to tender announcement and automatically filtering high-potential projects.
You might never have known a tender existed before; now AI is watching constantly. Don't have time to prepare the bid document? AI can write that for you too.
In the bid-document generation space, the market grew 230% year over year in 2025. iFLYTEK's "Spark Bidding" claims to compress bid production from 30 days to 3 days, improve document compliance by 90%, and increase win rates by 40%. Tai Biaoshu says it can generate a thousand-page bid in 30 minutes and extract over 200 key elements in 3 minutes. Kuaibiaoshu AI directly advertises rapid generation in 10 minutes.
I can't fully verify these vendors' numbers. But the direction is clear: the time and manpower needed to prepare a bid document are dropping sharply. The gap between well-prepared and poorly prepared bidders used to be huge; now it's shrinking fast.
The most deadly blow is to that final barrier.
Bidding requires clear functional parameters, then compares finished or near-finished products. Previously, if you hadn't built a similar system, you truly had nothing to show. But now, the clearer the functional description, the faster AI builds it. Throw the parameter requirements from the tender document into a coding agent, spend a few thousand yuan and a few days, and you can produce something demonstrable. Screenshots, feature points, even a working system you can show the evaluation experts.
In Y Combinator's Winter 2025 batch, 25% of startups had 95% of their code generated by LLMs. YC CEO Garry Tan put it bluntly: you no longer need a 50-to-100-person engineering team. This is the change brought by vibe coding—describe requirements in natural language, and AI turns them into runnable software. Cursor's daily active users exceeded one million, and by early 2026 its annualized revenue surpassed $2 billion. 84% of developers are using or planning to use AI coding tools, and 41% of code already involves AI generation.
Is that enough to handle a tender? Absolutely. The cost of subsequent refinement and delivery is also dropping significantly.
What Used to Be Worth 500K, Now Enters at 50K
In the past, a 500,000-yuan quote for a software project seemed normal—development and testing alone required considerable manpower and time. Now those costs are compressed to one-tenth or even less. Some dare to enter at 50,000 yuan: the bid is AI-written, the product is coding-agent-built, and their real cost burden is minimal.
To make matters worse, China's enterprise SaaS industry was already struggling. EY's 2024 report shows that the average net profit margin of listed Chinese enterprise SaaS companies over the past four and a half years has been negative. Gross margins are below 60%, sales expense ratios are 30%, and R&D expense ratios are 20%. The industry as a whole is still losing money.
The software outsourcing space is even worse. The industry's real state in 2025 has been summarized as a "three-piece set": unfinished projects, wage arrears, and layoff waves. SaaS products for SMEs are so homogenized that they compete on manpower alone. IBISWorld estimates the industry's profit margin in 2025 is just 12.5%.
Against this backdrop, bidding prices keep getting pushed lower. The whole system is starting to look a bit absurd.
The Path Large Models Took Last Year
In fact, large models already went through the exact same journey last year.
In early 2025, DeepSeek R1 was released, matching or even exceeding OpenAI o1 on key benchmarks, with API pricing at roughly 4% of OpenAI's. The same inference task cost $100 on OpenAI o1 but only about $3.60 on DeepSeek. Training costs were reportedly around $6 million.
A price war then broke out in the Chinese market. ByteDance's Doubao priced its Pro-32k model at 0.0008 yuan per thousand tokens—99.3% below the industry average. By July 2024, daily token usage exceeded 500 billion, a 22-fold increase from May. Alibaba's Tongyi Qianwen slashed Qwen-Long's input price from 0.02 to 0.0005 yuan, a 97% cut. Baidu directly announced ERNIE Speed and ERNIE Lite were free, and starting April 2025, all Wenxin models became free.
A RAND Corporation report found that Chinese AI models cost roughly one-quarter to one-sixth of comparable U.S. systems.
Top-tier models have all gone open source; paying for a model itself no longer makes sense. The eventual outcome everyone saw was that the model business transformed into services around models. Helping enterprises use models well, doing post-training, doing industry adaptation—these things have value. The model itself is no longer the object of transaction.
Software Is Walking the Same Path
Now, looking back at software.
When production costs fall low enough, the demand side will eventually ask: why go through such a complex procurement process to buy this thing?
The logic is identical to models. When supply-side costs approach zero, organizing complex transactions specifically for it loses meaning.
How can the bidding system respond? One direction is to stop comparing functional parameters and instead compare experience and case studies. How many clients have you served? How long has the system been running? Do you have records of large-scale usage? But this goes against the original purpose of bidding. The whole point was to compare things as standardized commodities so that competition could be fair. If it ends up coming down to experience and connections, what's the difference from skipping the tender and directly appointing someone?
This is where the problem gets stuck. Compare functions, and AI can help you build them fast—no differentiation. Compare experience and resources, and that's not what bidding is supposed to do.
Not Far Off
Many people think AI's impact on their industry is still a few years away. Bidding may not be able to wait that long.
In 2024, the number of IT government procurement projects grew 21.5%, but total value fell 4.5%. More projects, less money. Layer AI's compression of software costs on top of that, and the speed to the tipping point will be faster than most people expect.
The last time a similar situation was seen was in the large-model market in early 2025. From DeepSeek's open-source release to all-out price wars among major vendors, only a few months passed. Everyone was forced to transform: stop selling the model itself, and sell services around the model instead.
The software industry will most likely follow the same path. The transactional value of software itself will keep shrinking. What can really be sold are the services around software: helping clients clarify requirements, ongoing operations and maintenance, data migration, process transformation. The open-source community is going through the same shift—when production and distribution costs both approach zero, the original business model can't hold up.
Back to my friend's confusion at the beginning. He's thinking about how to adapt to the new price competition. But perhaps what he should be thinking about isn't how to win on price—it's how much longer this game itself can be played.
Originally published at https://guanjiawei.ai/en/blog/software-bidding-is-broken