Source: Duke Internet Society (ID: wlyxs888)

The market has once again misinterpreted the signal.

On September 15, news emerged regarding adjustments in the cooperation between Huawei and Seres, causing Seres' A-share price to plummet by 5 points on the same day, wiping out nearly 4 billion in market value, with an even steeper decline in its Hong Kong-listed shares. Public opinion quickly polarized into two camps: one asserting that "Seres has been ousted from the Five Realms," depicting it as a seven-year partnership coming to an end; the other dismissing it as merely a routine adjustment in the division of labor, nothing significant.

Neither perspective accurately captures the situation.

This is neither a breakup farce nor a simple adjustment of rights and responsibilities. It signifies a landmark transition for Huawei Smart Selection after seven years—officially moving from Phase 1.0 of "deep integration for demonstration" to Phase 2.0 of "standardized output for ecosystem building."

The pot remains the same, and the ingredients are unchanged. However, the cooking logic has fundamentally shifted.

【Phase 1.0 Era: Deep Integration, Essentially a Demonstration Project】

To comprehend today's adjustment, we must rewind to the starting point seven years ago.

In 2021, Seres, then known as Sokon, was a local automaker specializing in microvans and low-end SUVs, struggling with two consecutive years of losses and on the verge of delisting. Meanwhile, Huawei had invested tens of billions in its automotive business unit (BU) but was struggling to find a deep partner capable of translating its technology into sales volume. Despite having decent product strength, Arcfox failed to gain market recognition, proving that "technical excellence does not equate to sales success."

Faced with mutual stakes they couldn't afford to lose, the two sides struck a deal that seemed perfect at the time. However, many overlooked that this cooperation was an extremely asymmetric deep integration model from the outset.

Huawei provided a full-stack solution: product definition, styling design, intelligent driving cockpit, brand marketing, and channel sales, even dictating store display standards and sales scripts. Seres contributed the heavy assets: factories, production capacity, supply chain, after-sales service, and all operational risks.

Why deep integration? Because "Huawei doesn't build cars" was a controversial proposition back then, and the Smart Selection model was an untested novelty. To prove to the entire industry that "I don't build cars, but I can help automakers create blockbusters and become sales champions," Huawei had to go all-in, controlling every variable without a single misstep.

In other words, AITO was Huawei's demonstration project from birth.

The results of this model are self-evident. The 2023 M7 facelift achieved over 10,000 monthly sales within a month; the 2024 M9 shattered the price ceiling for luxury SUVs priced above 500,000 yuan, dominating the segment for 21 consecutive months with cumulative deliveries exceeding 280,000 units. Seres soared from the brink of delisting to the first tier of new-force automakers, with annual revenue reaching 165 billion yuan and net profit nearing 6 billion yuan.

But the costs were equally apparent.

According to Seres' prospectus and annual financial reports, from 2022 to 2025, Seres' procurement from the Huawei ecosystem surged from 5.8 billion yuan to 56 billion yuan, with its share of total procurement rising from 14.5% to 33.78%, totaling over 110 billion yuan in four years and accounting for nearly 30% of the same-period total revenue.

Per vehicle, AITO sold approximately 430,000 units in 2025, with about 130,000 yuan flowing to the Huawei ecosystem per unit—covering hardware procurement, technology licensing, and channel service fees. Meanwhile, Seres' net profit excluding non-recurring items stood at only 1.482 billion yuan, with per-unit net profit below 3,500 yuan.

A more fundamental issue was the mismatch between risks and rewards. Huawei settled orders upon delivery, collecting payment for each unit sold without bearing risks like inventory devaluation, price wars, or raw material price hikes; all operational fluctuations were absorbed by Seres. In the first half of 2026, Seres' revenue reached 57.5 billion yuan, down 7.87% year-on-year, with a net loss attributable to shareholders of 1.717 billion yuan, including a quarterly loss of 2.471 billion yuan in the second quarter alone.

This is the essence of the Phase 1.0 model: Huawei traded technology and brand for model validation, while Seres traded assets and risks for growth opportunities. Each got what they needed, but at a price.

Yet this model was destined to be transitional. It could validate a single demonstration but couldn't sustain an entire ecosystem.

【Why Now? Inevitability Driven by Triple Forces】

The adjustment in Q3 2026 wasn't a spur-of-the-moment decision but the culmination of three converging factors reaching a tipping point.

First and foremost, Huawei's resources were stretched thin.

The expansion of Harmony Intelligent Mobility from "One Realm" to "Five Realms" was the game-changer.

SAIC, once adamant about "not surrendering its soul to Huawei," not only launched Shangjie but also invested an additional 1 billion yuan in its Jinqiao exclusive factory. Chery, BAIC, and JAC also joined, forming a five-brand parallel landscape.

However, Huawei's product definition capabilities, channel resources, and marketing traffic all have limits.

In August, Harmony Intelligent Mobility delivered 42,101 units across all models, with AITO accounting for about 22,600 units, or over 53%; the remaining four brands combined sold less than 20,000 units, with Smart Series at 9,237, Shangjie at 7,888, Xiangjie at 1,627, and Zunjie at 749. Except for AITO, the other four were still in their ramp-up phase, all vying for Huawei's resources to boost sales.

If Huawei continued with the deep operations of Phase 1.0, its core team would be overwhelmed by AITO, leaving the other four brands perpetually under-resourced. Harmony Intelligent Mobility would remain just "AITO with different skins," failing to become a true multi-brand ecosystem.

For Huawei, releasing Seres to operate independently and reallocating saved product, marketing, and channel resources to new partners was the inevitable choice to expand the ecosystem's scale.

The second driving force was Huawei's own business model reaching an inflection point.

Huawei's automotive BU was shifting from "hardware-centric" to "software-centric."

In 2025, Yinwang Intelligent achieved 45 billion yuan in annual revenue, up 72% year-on-year, turning a profit for the first time. Software subscriptions and service revenues from Qiankun Intelligent Driving accounted for a rapidly increasing share, with margins far higher than hardware licensing. By July 2026, Qiankun Intelligent Driving had been deployed in over 1.9 million vehicles across more than 25 brands and 50 models.

As intelligent driving transformed from a scarce "luxury" to a standard feature in more models, Huawei's optimal business model shifted from earning hard money through deep operations of a single brand to outputting standardized technical solutions, leveraging economies of scale to reduce R&D costs and earning long-term recurring revenues through software subscriptions.

This resembles Google's approach: developing Nexus phones wasn't about competing with manufacturers but validating the Android system. Once the system proved viable, Google retreated to the background, becoming the underlying platform for all manufacturers to use and earning money from services and the ecosystem across the industry.

Huawei is now following the same path.

The third driving force came from Seres itself.

For Seres, this step was less of a proactive choice than a forced necessity.

The first-half 2026 losses were just the tip of the iceberg. A deeper contradiction lay in the fact that higher sales volumes deepened reliance on Huawei, weakening the profit structure. As the industry entered a downturn marked by price wars and rising costs, this model of "others taking the lion's share while you bear the risks" became unsustainable.

Even more critical was the brand issue. Over the past seven years, consumers bought AITO for Huawei, not Seres. "An AITO without Huawei is just a wild goose," though harsh, reflected the true market perception. Seres had never established its own brand voice.

For an automaker selling hundreds of thousands of vehicles annually with revenue in the hundreds of billions, having product definition, branding, and channel control in others' hands was commercially perilous.

Thus, Seres had long been preparing contingencies:

- At the channel level, it repeatedly mentioned "exclusive franchising" at investor meetings, aiming to separate AITO's sales channels from Huawei's stores.

- At the brand level, it incubated the new brand AIVA, utilizing ByteDance's Doubao large model, Yuanrong Qixing's intelligent driving, and CATL's batteries, completely detaching from Huawei's technology ecosystem.

- At the equity level, after renaming from Landian Technology to Saidou Technology and exiting the consolidated statements, Chongqing State-owned Assets held a 34.5% stake as the largest shareholder, with Seres holding 32.96%, fully isolating financial risks.

Regaining AITO's dominance wasn't an impulsive move but a long-planned step to transform from "Huawei's contract manufacturer" into a truly capable automaker.

【Phase 2.0 Era: Standardized Output, the True Ecosystem】

So, what does Phase 2.0 look like?

The official statement emphasizes "Seres-led, Huawei-enabled." Many focus on the first half but overlook the weight of the latter.

Huawei isn't exiting; it's changing roles from "chief orchestrator" to "technology foundation supplier."

What remains unchanged: Qiankun Intelligent Driving and Harmony Cockpit continue to be fully deployed, with Huawei's technology foundation intact. What changes are operational aspects like product definition, brand marketing, channel retail, and service systems, which revert to the automaker.

Behind this lies a new cooperation logic:

- Huawei outputs standardized technology modules—intelligent driving, cockpit, chassis, electric drive—like Lego bricks for automakers to use as needed.

- Automakers handle product definition, brand operations, manufacturing, and channel sales, bearing all operational risks while enjoying greater profits.

- Huawei's revenue shifts from "per-vehicle procurement sharing" to "technology licensing fees + software subscriptions," earning lighter, steadier, and longer-tail income.

This is the true "Android model."

Google never designed phones, set prices, or managed channels for Samsung or Xiaomi. It focused on the underlying system and services, allowing manufacturers to build their products based on the system. Sales success or failure was their responsibility, with Google earning per-unit licensing fees and subsequent revenues from the app ecosystem.

Huawei Smart Selection 2.0 is the automotive version of this logic.

For Huawei, the benefits are clear:

- Marginal costs plummet. Deep operations for one brand require hundreds of personnel; standardized solutions for ten brands don't scale team size linearly.

- Ecosystem scale explodes. Expanding from one bound partner to five open ones, and potentially dozens in the future, Qiankun's deployment will rise exponentially, accelerating the data flywheel and algorithm iteration.

- Revenue quality improves significantly. Hardware licensing is one-time; software subscriptions are recurring. The former fluctuates with sales; the latter depends on user base, placing them on entirely different levels of business model health.

For Seres, challenges and opportunities are two sides of the same coin.

In the short term, the withdrawal of Huawei's endorsement may erode brand premium, pressuring sales and used-car prices; channel independence and team-building will also increase short-term costs. But in the long run, it finally has the chance to build its product definition capabilities and brand equity, transitioning from "relying on others for meals" to "cooking for itself."

For the industry, this marks the dawn of a new phase:

Only when Huawei retreats to the technology foundation layer and stops competing directly in product development will true competition emerge within the "Huawei ecosystem." With five brands sharing the same technology foundation, what will they compete on? Product understanding, cost control, channel capabilities, and brand operations.

Once technological barriers are leveled, automakers' core competitiveness will return to themselves.

【All Technology-Enabled Cooperations Follow This Law】

The true value of this event lies not in who wins or loses between Huawei and Seres but in the universal business law it reveals.

All "technology-enabled + traditional industry" cooperations likely go through three phases:

Phase 1: Deep Integration, Model Validation

The technology provider goes all-in, deeply binding with one partner, pouring all resources to create a successful demonstration, proving to the industry that "this is feasible." Efficiency trumps all in this phase, with blurred divisions of labor as the technology provider handles both technology and operations.

Phase 2: Standardized Output, Ecosystem Scaling

After model validation, the technology provider gradually withdraws from operations, standardizing and modularizing capabilities for wider partner adoption, leveraging economies of scale to reduce costs and earning through technology licensing and services. Scale becomes the priority, with clear divisions of labor as the technology provider returns to the underlying layer.

Phase 3: Tiered Operations, Barrier Formation

When the ecosystem grows large enough, the technology provider becomes a de facto infrastructure, earning sustained revenues from underlying systems, data services, and app ecosystems. Barriers become the priority, with the technology provider controlling the core rules.

From Windows in the PC era to Android in the smartphone era, and now Huawei Smart Selection in the intelligent automotive era, the underlying logic remains the same.

Seven years ago, Huawei chose Seres because it needed a partner willing to surrender its soul to survive, jointly taking Smart Selection from 0 to 1.

Seven years onward, Huawei has parted ways with Seres, as it seeks to expand its model from a singular instance (1) to multiple applications (N), with the goal of establishing a robust and genuine industrial ecosystem.

This move is not rooted in emotional bonds or their absence; rather, it is a natural consequence dictated by the immutable laws of business.

【Epilogue】

After seven years of sharing the same "pot" (metaphorically referring to a close partnership), the time has come to "split the stoves" (meaning to go separate ways and establish independent operations).

Splitting stoves does not imply a breakup, still less a falling-out. Instead, it signifies that both parties have matured—one is poised to construct a grander kitchen, while the other is ready to learn and master the art of independent cooking.

Huawei's objective is to demonstrate that its Smart Selection strategy can continue to thrive without its direct involvement, and that an ecosystem akin to Android for intelligent vehicles is within reach.

Seres, on the other hand, aims to prove that it can still create blockbuster products, manage its brands effectively, and maintain a strong foothold in the market, all without Huawei's guiding hand.

For all corporate decision-makers observing this development, the more profound question arises:

When collaborating with a technology provider, at which stage of development is your company currently positioned?

Are you merely leveraging technology to ensure survival, or do you aspire to ultimately cultivate your own unique identity and capabilities?

The answers to these questions will chart entirely divergent courses for your company's future.