Author | Zhou Lei, Editor | Bao Yonggang
In the past decade of cloud computing, a mature ecosystem model supported rapid market expansion: cloud vendors provided standardized computing and storage resources along with foundational technology stacks, while partners across the region acted as sales channels and system integrators, packaging these products into end customers' solutions.
Through long-term communication with ecosystem partners, we have learned that this division of labor was indeed efficient and direct, fitting the characteristics of the cloud market at the time. However, it gradually solidified the positions of both parties, and new contradictions began to emerge—the core value of partners was often simplified to "customer relationships" and "integration capabilities." Their revenue highly depended on resource price differences and vendor rebates in projects, essentially earning hard money from information asymmetry and localized services.
As price wars in cloud computing continue to escalate, this "price difference" margin is continuously compressed, challenging single-source revenue. Meanwhile, partners are also worried: will any company with similar customer relationships and delivery capabilities easily replace them?
01. "Growth Anxiety" and "Trust Crisis" in the Cloud Ecosystem
We have mentioned in previous analytical articles on the cloud ecosystem that under this traditional model, ecosystem partners experience a "precarious" growth anxiety. Behind this lies a deeper sense of insecurity: they worry that the solutions they have poured their hearts into, refined based on industry knowledge, might ultimately be absorbed and standardized by the platform, thus having the rug pulled out from under them. They are also anxious about the platform's frequent adjustments to channel policies and assessment metrics, which make long-term investments and business planning feel at a loss.
This concern over "blurred boundaries" and "uncertain rules" is essentially a trust crisis derived from solidified roles and shallow moats. This is also closely related to the strategic ambiguity of some cloud vendors: they claim to treat partners as a "community" for co-creating industrial value, but in actual implementation, they still focus on sales targets.
However, the rapid development of AI technology and the urgent need for industrial implementation, along with the emergence of large models and intelligent agent technologies, mean that the rules of the game for the entire ecosystem have the opportunity to be rewritten. The old survival logic is failing, and a new value network based on deep cognition, scenario innovation, and operational services is gradually emerging. This also points out a breakthrough path for those partners mired in anxiety: instead of struggling in the old model, it is better to rely on their unique industry "strong suits" to take the lead in establishing an irreplaceable position in a brand-new arena that emphasizes professional division of labor.
Therefore, the key to breaking the game lies not only in the simple upgrade of technology or products, but more importantly, in whether cloud vendors can fundamentally change their roles, build a new type of relationship based on clear value division of labor and mutual growth, and rebuild a new trust for ecosystem partners. Baidu AI Cloud's partner cooperation practice provides a new paradigm for the industry on how cloud vendors and partners can co-build and win together.
In the past, when cloud vendors reviewed their cooperation with partners, they sometimes easily focused on a few indicators, such as the total contract amount or gross profit margin of projects, or the proportion of self-developed products. Today, advanced cloud vendors like Baidu AI Cloud are beginning to realize that focusing on providing powerful and easy-to-use AI toolboxes and verified solution prototypes, allowing partners to focus on what they do best—transforming industry knowledge into specific product optimization, service design, and business closed loops, clarifying each other's boundaries and forming consensus—is the only way to take a further step towards true value co-creation.
In this way, the test for cloud vendors escalates to: do they already possess an efficient, reliable, complete, and open "AI capability matrix"? Have their large model training and inference capabilities, intelligent agent construction frameworks, and reusable industry algorithm models formed a "capability toolbox" that can be flexibly combined and deeply customized according to the endowment advantages of different partners?
For different types of partners, they indeed possess a deep understanding of vertical industries, proprietary data for segmented scenarios, service networks rooted in regions, and continuous operational capabilities facing customers directly. Can they all find the most suitable access points and amplification methods for their own strong suits within this AI capability matrix?
02. A "Self-Funded" Exploration: The Market Breakthrough Battle of Traditional Vendors
The sense of crisis among ecosystem partners is real and agonizing. But some are already facing this test head-on.
Xiong Mengfei, Deputy General Manager of Taiyue Satellite, told us that in the first half of 2025, there was an intense internal debate over a matter concerning the company's transformation and future.
As a "veteran" deeply engaged in the field of smart transportation in Hubei for nearly ten years, Taiyue Satellite has participated in urban traffic congestion management since 2016. But what the company's management feels most acutely now is a profound transformation anxiety: even if current business performance is decent, the wave of technology is already approaching—how can they create something eye-catching and seize the initiative?
Coincidentally, Taiyue Satellite established contact with Baidu AI Cloud in 2024 and gained an in-depth understanding of Baidu AI Cloud's traffic signal control SaaS platform in 2025. At that time, Taiyue noticed that the platform could obtain real-time road condition information, analyze congestion points at intersections, and optimize signal control to alleviate traffic congestion. The congestion mitigation indicator data for each intersection could be quantitatively analyzed and displayed through the system. This intuitive effect provided solid data support for internal reporting or public announcements to citizens.
Consequently, the chairman proposed taking the initiative to invest substantial funds to build the province's first Baidu AI Cloud traffic signal control SaaS pilot site in Ezhou.
This proposal to "self-fund the pilot" sparked huge controversy internally. The reasons for opposition were also very realistic: first, the company had to pay out of its own pocket and mobilize its own resources to negotiate with multiple parties such as traffic police and road administration; second, the local traffic signal control facilities in Ezhou were relatively complete, so the improvement effect might not be significant. If the expected results could not be achieved, the loss would be considerable.
The deeper concern was: if the pilot succeeded, would Baidu AI Cloud bypass them and cooperate directly with the customer, leaving Taiyue Satellite to "do the groundwork for others to reap the benefits"?
Taiyue Satellite ultimately took this step.
On the one hand, it was their own sense of crisis stemming from the instinct for survival. The management of Taiyue Satellite clearly realized that instead of clinging to traditional projects, it was better to have the courage to build a pilot and let the results speak for themselves to win the market; there was no third way. Moreover, Ezhou happens to be one of the transportation hub cities in Hubei, with a large cargo airport. If a successful pilot could be built here, it would inevitably create opportunities for Taiyue to open up the provincial market in the future.
On the other hand, two outcomes from Baidu AI Cloud provided confidence for Taiyue Satellite.
First was the technically verified certainty. Before the formal decision, Taiyue Satellite selected two intersections for a small-scale POC test. After the Baidu traffic signal control SaaS platform went online, the average vehicle delay decreased by about 8%. Xiong Mengfei admitted that the test data solidly proved the effectiveness of the technical path.
Second, and more decisively, Baidu AI Cloud committed to signing an ecosystem cooperation agreement, explicitly designating Taiyue Satellite as the core delivery partner in the Hubei market. Subsequent projects would be led and implemented by Taiyue, with full-process technical support provided.
This agreement dispelled the partner's biggest concern about "being replaced by the platform," drawing clear boundaries of capabilities and responsibilities for both parties—Baidu AI Cloud focuses on providing and iterating core AI algorithms and SaaS products, while market development, localized delivery, and continuous services are the irreplaceable value territories of the partners.
This exploration ultimately turned into an excellent strategic breakthrough. Taiyue Satellite proactively invested manpower and material resources, successfully creating the province's first benchmark case in Ezhou. The results exceeded expectations: the average vehicle delay on the pilot trunk road decreased by over 8%, with a maximum drop of 14.8%. The effect received public recognition from the Ezhou Traffic Police Detachment: "(The platform) greatly improved efficiency... and we can intuitively see the effects from the platform; it is a very useful tool."
Xiong Mengfei told us that it was summer in Hubei at the time, with weather alternating between sun and rain. Engineers from Baidu AI Cloud and the team from Taiyue Satellite ran all over the streets of Ezhou together, meticulously recording traffic flow data under different scenarios, weather conditions, and time periods, and then returning to the conference room for in-depth analysis and discussion. Through the joint efforts of both parties, this benchmark case took only about two to three weeks from decision to completion.
He revealed that next, Taiyue Satellite will promote and implement this solution in Zhijiang City and other counties and cities under the jurisdiction of Yichang, Hubei.
This cooperation between Taiyue Satellite and Baidu AI Cloud is not just the smooth delivery of a single project; it also signifies a fundamental shift in the market role of such ecosystem partners. The two parties evolved from a Party A-Party B relationship based on single projects into a "regional community" based on long-term common interests and clear rules. Taiyue Satellite itself also transformed into an "evangelist" and deep delivery partner for Baidu's AI capabilities in the Hubei region.
More importantly, a new paradigm of "joint market expansion" has emerged in the business model of ecosystem partners. That is, the market expansion logic is expected to upgrade from the previous relationship-driven single-project gaming to effect-driven rapid replication of standardized products. Xiong Mengfei told us that the Ezhou pilot completely changed their sales methods and sales cycles, becoming a stepping stone to knock on the doors of more than ten other prefecture-level cities in Hubei.
"In the past, we just took technical solutions to talk about some advantages, which was mere theoretical talk, and customers would be half-believing and half-doubting. Now we can bring the real data from Ezhou and the platform data that customers can see in real-time to quickly advance the project. This also dispels customers' doubts, and our entire communication cycle will be reduced." We learned that under the original sales model, it took an average of about half a year from communication to project implementation; now with the Ezhou pilot, this figure can be shortened to three months.
03. Smart Elderly Care "Hubei Sample": Co-creating a Brand New Business Model
If the case of Taiyue Satellite is about breaking through with "joint market expansion" on an existing battlefield, then the story of Ruilian Kangda reveals another completely different possibility: through "model co-creation," jointly defining a brand new service paradigm and business closed loop in a blue ocean market that has not yet been fully digitized.
Unlike the transformation anxiety of Taiyue Satellite, the anxiety of Lu Xiaodan, Chairman of Ruilian Kangda, stems from a more fundamental industry dilemma. The current state of elderly care in China is that "90% of the elderly wish to age in place, but very few have actually experienced relevant services." The market has long hovered at the primary stage of "policy-driven, project-based procurement," and elderly care service providers often operate around government tenders rather than the real needs of the elderly.
She sighed during the exchange that this dilemma is very specific in daily scenarios. For example, many elderly people in their sixties and seventies need to go out to buy heavy items like rice and oil and carry them upstairs themselves, as old residential areas lack door-to-door delivery; when elderly people living alone encounter minor illnesses, they are afraid to disturb their children's work and often seek medical treatment alone, and even when undergoing examinations that require general anesthesia, such as colonoscopies, they cannot proceed because they cannot find accompanying personnel.
The team at Ruilian Kangda, with its deep technical genes, once tried to break through with pure self-research and even made a hardware demo. But Lu Xiaodan soon realized that in the field of elderly care, which is extremely operation- and service-heavy, "R&D is only the first step." The massive investment required for subsequent continuous iteration, service guarantee, and ecosystem construction is a more cruel test for them.
The turning point began with a chance encounter, which Lu Xiaodan frankly described as "the perfect alignment of timing, location, and human effort." Ruilian Kangda and Baidu AI Cloud's Qianfan Large Model Innovation Center are located in the same Hanjiang Bay AI Industrial Park. During a park event in October 2025, the two teams got to know each other.
At that time, Ruilian Kangda had already investigated various general-purpose smart hardware on the market but felt deeply that they lacked designs deeply integrated with industrialized and localized services. When Baidu showcased its mature Resident Service Agent (RSA) platform and "Xiaodu" hardware, Ruilian Kangda was immediately impressed.
Upon in-depth understanding, we found that the "Xiaodu" smart screen, as a terminal deeply penetrating households, provides an excellent elderly-friendly interactive interface. For example, all services can be invoked through natural voice interaction; the device is portable and can follow the elderly from the living room to the kitchen and bedroom; the integrated camera and SOS emergency call functions solve the extreme pain point of "not being able to reach the phone after a fall," becoming a safety guardian at critical moments.
On the other hand, the RSA platform provided by Baidu AI Cloud acts as a digital foundation for "connection" and "empowerment," providing a plug-and-play integration framework for Ruilian Kangda's years of accumulated industry understanding and operational resources. Baidu AI Cloud can handle general connection capabilities, order routing, and data channels. Ruilian Kangda, without needing to build everything from scratch, can integrate its accumulated local elderly care service provider network, unique service closed-loop management system (including supervision and assessment of service personnel, star ratings), and independently designed operational processes into this platform to grow.
Their co-creation is actually directly aimed at "running through a service model" and building an online hub for a "10-minute life circle": allowing the elderly to link to community canteens, book home services, and access health management with a single call of "Xiaodu," transforming low-frequency, occasional offline service contacts into high-frequency, convenient online interactions.
Lu Xiaodan vividly described this change using her own mother as an example: the smart screen is used for up to 6-8 hours a day, becoming a natural part of the elderly's life from entertainment to safety care. More importantly, operators can clearly perceive the state of the elderly through data, shifting services from "blind surprise attacks" to "precise responses." The service radius of an elderly care butler can expand from 30 people to 300 people, achieving a tenfold increase in efficiency.
Compared with the previous "procurement-integration" ecosystem relationship, this cooperation is a deeper coupling of "joint product R&D - joint operation sharing - joint revenue sharing." Baidu AI Cloud even provided the rare "secondary development authorization," enabling Ruilian Kangda to deeply integrate its professional elderly care service management system with Baidu's platform, utilizing the platform's capabilities while maintaining business autonomy.
The value of Ruilian Kangda as an ecosystem partner is clarified here: its core assets are the locally deeply cultivated service ecosystem and continuously operated user relationships. At the same time, revenue sources have upgraded from government project orders to sustainable service commissions, value-added product sales revenue, and potential membership fees. Profits are expected to be deeply bound to user activity and service usage frequency, entering a virtuous cycle.
This successful model gave Lu Xiaodan the confidence to push the "Hubei Sample" to a broader market. Ruilian Kangda is standardizing and packaging the verified operational model into a rapidly replicable solution package. She told us that the solution has been successfully run in Wuhan and Huangshi. The next core task is to find more like-minded partners to scale this up together. In such an industry that values long-term, steady growth, it is not easy to find patient companions. What Baidu AI Cloud provides is not only a technical foundation but also a "confidence foundation" that has been verified and can effectively reduce the complexity and initial investment of partners' cross-regional expansion.
04. The Confidence and Courage of Pioneers
The two cases of Ruilian Kangda and Taiyue Satellite coincidentally interpret how ecosystem partners with different directional advantages and accumulations fully leverage their own characteristics based on Baidu AI Cloud's AI capability matrix.
In the story of Taiyue Satellite, Baidu AI Cloud, with the verifiable hard power of technical products and self-restraint in the form of contracts, clarifies "what is the irreplaceable value of partners" and jointly explores the market with them. In the story of Ruilian Kangda, Baidu AI Cloud provides an operable technical soil, allowing and encouraging partners to integrate their strongest industry knowledge and core strengths into it, together growing a business model that completely fits the flesh and blood of the scenario.
As a cloud vendor that has welded "intelligence" into its body, Baidu AI Cloud has been targeting the AI cloud market for many years. In this rapidly developing market, Baidu AI Cloud can transform cutting-edge AI technology into a set of ready-to-use and deeply combinable "business capability modules," directly empowering partners to open up new markets, new products, and new profit spaces that were previously inaccessible.
The core supporting Baidu AI Cloud to realize the new AI ecosystem stems from Baidu's full-stack self-developed AI layout from underlying chips, AI cloud, and large models to upper-layer applications. This extremely rare layout in the industry ensures deep optimization and efficient collaboration at all levels of the technology stack, thereby thoroughly productizing and standardizing AI Infra and Agent Infra, forming an open and stable "innovation foundation" and providing a layered set of capabilities that can be deeply understood and called.
We previously reported that Baidu AI Cloud has built an AI Infra system centered on its self-developed Kunlun chips and the Bage AI computing platform, achieving efficient release of computing power through hardware-software collaboration. Relying on the large-scale implementation practice of 10,000-card and 30,000-card clusters, Baidu AI Cloud continues to lead in key indicators such as training efficiency, stability, and energy efficiency control, providing a sustainable, low-cost, and replicable computing power path for large model training and inference.
On the industrial scenario side, Baidu AI Cloud possesses the Qianfan large model platform, building a full-stack Agent Infra system covering models, tools, development, data, and runtime environments, providing enterprises with a low-threshold, highly reliable, and scalable path for intelligent agent implementation. Relying on over 150 full-scenario models including Wenxin 5.0, combined with rich tool capabilities such as Baidu Search, Maps, Wenku, and Netdisk, as well as the open MCP ecosystem, Qianfan enables enterprise intelligent agents to have stronger understanding, execution, and collaboration capabilities.
Baidu AI Cloud provides the industry's rare full-stack self-developed AI cloud layout. Partners do not need to build technical high-rises from scratch; instead, they can precisely insert their core strengths in industry knowledge, scenario data, localized services, or deep operations into this foundation like calling "Lego modules," jointly defining new products and models.
The role of Baidu AI Cloud is to provide verified and reusable "industry solution prototypes," such as the Resident Service Agent platform for smart elderly care or the traffic signal control SaaS platform for smart transportation, thereby rapidly transforming partners' core strengths into scalable and sustainable new business models. In other words, the "stage" provided by Baidu AI Cloud is large enough, and the lighting and stage configuration are rich enough to allow ecosystem partners to choreograph their most proficient "scripts."
The commercial effectiveness of this new paradigm of value co-creation has been preliminarily verified in multiple industries. In the financial industry alone, which has extremely stringent requirements for security and stability, Baidu AI Cloud ranked first in the number of winning bids for large model projects with 38 projects.
The 2025 annual ecosystem report card released by Baidu AI Cloud further shows that its core partners' revenue grew by over 100% year-on-year, and the annual revenue of channel-side partners reaching the tens of millions level achieved more than three times growth; among all its implemented projects, delivery partners achieved 100% participation. At the same time, product cooperation covers most core scenarios, and the number of solutions co-created with partners and entering the replication stage has reached hundreds.
Behind these numbers are partners like Ruilian Kangda and Taiyue Satellite, who have transformed from "project executors" to "value co-creators," releasing growth momentum. It proves that when ecosystem cooperation moves from shallow "resource superposition" to deep "capability integration," the growth created is multiplicative.
Yin Yingli, Vice President of Baidu AI Cloud and General Manager of the Channel Ecosystem Department, stated that in 2026, the company will continue to increase ecosystem investment. The core directions are to provide a richer product shelf, build more open collaborative policies, and create more integrated platform capabilities. All measures are for only one goal—to share more abundant growth dividends with partners in 2026.
When true opportunities arrive, they are often not like the red light at an intersection with a clear countdown. It is more like the moment the green light turns on—only those who are fully prepared and have stepped on the gas pedal can rush out at the first moment and seize the initiative.
For partners, this preparation is years of industry accumulation, keen insight into local markets, and the courage to dare to transform. For Baidu AI Cloud, this preparation is the full-stack self-research from chips to models, the continuous solving of industrial problems, and the determination to transform capabilities into an open "foundation." When the wave of industrial intelligence truly hits the shore, only the most fully prepared core strengths can mesh precisely and leap together to the crest of new value.
This road has just begun, but the direction is already clear.