On July 15, Alibaba and SAP officially announced a strategic partnership.
This cooperation is not the usual "we signed a framework agreement" — it involves three specific actions. First, Alibaba Cloud has joined SAP's IaaS certification program, becoming a hyperscale cloud service provider supporting SAP workloads. Second, SAP is exploring integrating Alibaba's Tongyi Qianwen (Qwen) large model into SAP's generative AI hub in China. Third, Alibaba's own infrastructure will deploy SAP's cloud ERP private cloud edition.
Every July is the final warm-up period before SAP's earnings season. Dropping such a heavyweight partnership at this point sends a clear signal: SAP's AI strategy in China is no longer just about "localizing global products" — it has begun seeking true local AI partners for deep integration.
I took a careful look at this message and feel there are several layers worth discussing in detail.
Three layers of cooperation, each layer breaking down walls
Looking at this collaboration between SAP and Alibaba Cloud in detail, each layer addresses different pain points.
First layer: Alibaba Cloud becomes an SAP IaaS-certified cloud service provider. In plain terms, this means: SAP is telling its Chinese enterprise customers that they can now run SAP systems on Alibaba Cloud, and it is officially certified. Previously, the primary cloud infrastructures promoted by SAP in the Chinese market were AWS and Microsoft Azure, but many Chinese enterprises had limited acceptance of these two due to concerns over data compliance, network latency, and localized services. Alibaba Cloud's entry provides SAP's Chinese customers with an additional local option, and potentially a more suitable one.
Layer 2: Integrating the Qwen Large Model into SAP Applications. This layer is the most interesting. SAP is exploring making Qwen a callable model within its generative AI hub. This means that when Chinese enterprise users utilize SAP systems, they can use Qwen to generate report summaries, perform data analysis, and assist in decision-making — all without data leaving the country.
Previously, when SAP was promoting AI features in China, it encountered an awkward situation: most of the AI models used globally by SAP were Anthropic's Claude and OpenAI's GPT series, but these models had issues with availability and compliance in mainland China. After connecting with Qwen, this problem was at least half solved.
Layer 3: Alibaba itself uses SAP. This layer seems the most ordinary, but it may be the most convincing. Alibaba itself decided to deploy the SAP cloud ERP private cloud version in its infrastructure — meaning one of SAP's largest potential customers is also using this solution. In the procurement logic of enterprise-level software, this carries considerable weight.
Sapphire 2026's 7 AI Tools: SAP's AI Offensive Comes Prepared
The Alibaba partnership is not SAP's only major move in the near term. Sapphire 2026 preview information revealed in the same week shows that SAP is about to launch seven new AI-assisted tools, comprehensively upgrading its Cloud ALM cloud application lifecycle management system.
I have seen specific tool directions: enterprise process automation, intelligent system operation and maintenance alerts, intelligent data analysis, and business decision support. These tools are no longer just helping users "answer questions," but are directly integrated into the operation of ERP systems—the system itself analyzes data, identifies issues, issues alerts, and even provides decision-making recommendations.
SAP Asia Pacific management has also undergone a simultaneous reshuffle: Verena Siow has been appointed President of SAP Asia Pacific, based in Singapore, with the core task of coordinating customer success and AI solution implementation in the region. With Alibaba Cloud as a catalyst, SAP's AI pace in Asia Pacific is clearly accelerating.
Billing model changed: from "buying a seat" to "buying AI thinking"
Another move by SAP is worth elaborating on. According to SAP's weekly news report, SAP is advancing a new billing model for RISE and GROW with SAP, with the core logic being "pay per use, pay per value." Simply put: the more AI features you use, the more you pay; the more effective the AI features you use, the stronger the justification for payment.
The pressure behind this shift comes from two directions: first, cloud vendors' infrastructure costs are rising, and SAP needs a new pricing system to hedge against this; second, customers are becoming increasingly sensitive to the value of AI and are unwilling to pay a fixed subscription fee for a feature that is "useful but whose value is hard to quantify."
In my previous article, I discussed the major trend of ERP shifting from selling licenses to selling "AI thinking." SAP's latest billing reform is aligned with its product investments in AI — the seven AI tools are not given away for free, but customers can pay based on actual usage rather than being forced to purchase a full package.
Siemens+IFS Alliance: The ERP market landscape is beginning to shift
Another piece of news this week is also worth mentioning: Siemens and IFS have formed an industry alliance, breaking the existing landscape of the high-end enterprise ERP market — in simple terms, it challenges SAP's exclusive position in the manufacturing sector.
IFS specializes in EAM (Enterprise Asset Management) and FSM (Field Service Management), while Siemens has a foundation in industrial automation and PLM. Both companies indeed overlap with SAP in their manufacturing coverage. The direct consequence of this alliance is that purchasers of ERP for high-end manufacturing now have a second reliable option.
For Party A, this is a good thing. No matter how fast SAP's AI transformation pace is, with competitors nearby, prices and service terms will be more reasonable.
What does this mean for Chinese enterprises' ERP selection
Writing this far, I want to connect a few things together:
First signal: Running SAP in China finally has a true cloud foundation. After Alibaba Cloud certification, SAP now has a "compliant + localized + low-latency" cloud deployment path in China. Previously, many enterprises could only run SAP on self-built data centers; now there is at least one more option worth serious consideration.
Second signal: The choice of AI models is returning to customers. SAP can integrate Qwen, meaning that at the AI layer, ERP vendors are no longer tied to a single model. A couple of days ago, Odoo did something similar—supporting both OpenAI and Gemini in its AI Provider configuration. The multi-model strategy is shifting from "optional" to "standard."
Third signal: ERP pricing logic is being reshaped by AI. SAP has shifted from fixed subscriptions to AI usage-based billing, Odoo uses an all-inclusive price but offers MCP Server for free, and Workday adopts a hybrid model of per-user plus AI usage billing. Although the three models differ, they share a common direction: AI capabilities are no longer "paid add-on features" but are becoming part of the core billing of ERP.
At the end of July, SAP will release its second-quarter financial report. By then, we will be able to see the actual impact of this wave of AI investment and billing reform on SAP's revenue. I will update another article at that time.
Finally, to put it bluntly: the collaboration between SAP and Alibaba Cloud carries more weight than most press releases suggest. It is not an ordinary channel partnership, but a deep integration at the product level. From the underlying cloud infrastructure to the top-tier AI models, SAP is building a complete localized technology stack in the Chinese market.
For Chinese enterprise managers currently selecting an ERP system, what deserves attention in the second half of the year is not only the deep integration of Odoo 20 and SAP Alibaba, but also the pace of domestic ERP vendors (Kingdee, Yonyou) catching up in the AI layer. In this ERP+AI competition, the second half of 2026 will be much more intense than the first half.
