October 1, 2025
Cloud Cost News Roundup for September
8 min read
Cloud costs and AI investments are dominating headlines and budgets. September was packed with big moves in cloud cost management and strategy. Microsoft struck billion-dollar government deals, Alibaba unveiled software-driven efficiency breakthroughs, and enterprises worldwide continued to wrestle with the ever rising costs of AI. At the same time, hyperscalers like Oracle, AWS, and SAP doubled down on multi-cloud, generative AI, and sovereign cloud initiatives.
Amid these shifts, Amnic launched its Enterprise FinOps Acceleration Kit to help organizations benchmark their FinOps maturity, close governance gaps, and prioritize cost-saving actions with easy-to-use diagnostics, providing a clearer path to control and optimize cloud spend.
In this cloud cost news roundup for September, we break down the most important cloud cost stories so you don’t miss the trends shaping technology, budgets, and business strategy.
Key Cloud Cost Stories from September
Amnic launches FinOps Acceleration Kit to nenchmark your FinOps maturity

Amnic’s Enterprise FinOps Acceleration Kit is a practical toolset designed to help organizations assess their current position in the FinOps lifecycle. The kit includes four ready-to-use worksheets that enable teams to benchmark maturity, identify governance gaps, and prioritize high-impact cost actions, all with minimal effort.
Built for engineering, finance, and business stakeholders, the worksheets can be completed in under 20 minutes. They help set quarterly outcomes tied to a 24-month roadmap covering visibility, governance, allocation, forecasting, and optimization, driving clarity and alignment across teams to accelerate FinOps progress.
Alibaba unveils new tools to cut outages and cloud costs
Alibaba has shared new research showing how it’s cutting outages, improving efficiency, and lowering cloud costs through smarter software. The company detailed three systems that will be presented at the upcoming SIGCOMM conference:
ZooRoute: A failure recovery service that keeps cloud networks running smoothly during outages. By constantly checking for alternate paths, it can reroute traffic instantly, reducing outage time by over 92% in production use.
Hermes: A smarter approach to load balancing using eBPF technology. It helps distribute traffic more evenly across servers, cutting CPU imbalances by 90% and reducing load balancer costs by 19%.
Nezha: A system for balancing workloads across SmartNICs (network cards with processors). It shifts tasks from overloaded cards to underused ones, improving performance at a fraction of the cost of new hardware.
Together, these innovations show how cloud providers like Alibaba are using software-driven techniques to improve reliability and cut infrastructure costs, key factors in both cloud operations and FinOps practices.
Microsoft offers free Copilot & discounts for U.S. government agencies

Microsoft has struck a major deal with the U.S. General Services Administration (GSA) to make its cloud and AI tools more affordable for federal agencies. Under the new OneGov agreement, Microsoft will:
Provide Copilot (its AI assistant) for free to Microsoft 365 government customers for the next 12 months.
Offer discounted pricing on Microsoft 365, Azure Cloud Services, Dynamics 365, and security tools for up to 36 months.
Give some agencies Dynamics 365 for free for one year if they qualify.
The GSA estimates these discounts could save agencies $3.1 billion in the first year alone.
Microsoft is also investing $20 million in workshops and support services to help agencies adopt these tools effectively. Over three years, the company expects the program to deliver $6 billion in value.
This move comes as rivals like OpenAI, Anthropic, and Google also roll out ultra-low-cost government AI deals, with offers starting as low as $1 per year. Microsoft, however, has a head start thanks to its long track record of working with government cloud services under FedRAMP authorization.
SAP commits $23 billion to Europe's sovereign cloud
SAP has announced a massive €20 billion ($23.3 billion) investment over the next decade to expand its sovereign cloud services in Europe. The move reflects growing demand from European businesses and regulators to keep sensitive data stored and processed within local borders.
SAP’s Sovereign Cloud portfolio offers:
A full technology stack hosted entirely in EU-based data centers.
The option to run infrastructure directly in a customer’s chosen data center.
Delos Cloud, a Germany-specific solution built for local compliance.
This push comes as data sovereignty becomes a hot topic in Europe, driven by GDPR and the upcoming EU AI Act. Major players are also responding: AWS, Microsoft, and Google have all recently launched or announced Europe-focused sovereign cloud services to address stricter privacy and compliance needs.
For enterprises, this trend means more options to meet compliance without sacrificing cloud scalability, but often at a higher cost compared to standard public cloud offerings.
AI cost management challenges grow as 85% of companies miss spending forecasts and face margin pressure

2025 State of AI Cost Governance report reveals that most companies are struggling to accurately forecast and control their AI infrastructure costs. In fact, 85% of organizations miss their AI spend forecasts by more than 10%, leading to significant profit margin erosion. The rapid adoption of AI technologies like large language models and GPU computing is outpacing companies’ ability to manage the associated costs effectively.
Key findings from the report include:
84% of companies report gross margin losses due to AI infrastructure costs, with 26% seeing a margin impact of 16% or higher.
67% of organizations are planning to repatriate AI workloads back on-premises, and 61% already run hybrid AI infrastructure (combining public and private environments).
Many companies have blind spots in cost reporting; only 35% include on-premises AI costs in their financial tracking.
Unexpected cost surprises often come from:
Data platforms (56% report unexpected spend here)
Network access costs (52% face surprises in this area)
AI cost forecasting is largely ineffective:
80% miss their AI cost forecasts by more than 25%
Only 15% can forecast costs within a 10% margin of error
Lack of visibility is the biggest barrier to managing AI costs, with only 34% of companies reporting mature cost management practices.
Organizations that charge for AI usage show twice the level of cost maturity and better discipline in attribution.
The report emphasizes the urgent need for improved AI cost governance, better visibility, and more accurate forecasting if companies want to protect profitability as AI investments grow.
94% of IT leaders struggle to optimize cloud costs
A recent global report shows that 94% of IT leaders struggle to manage and optimize their cloud costs, making it the biggest challenge in IT today. The study surveyed 2,300 IT decision-makers worldwide and found that almost half of IT budgets are going into expanding cloud capabilities, especially to support AI workloads.
Security and compliance concerns are driving some organizations to move part of their IT infrastructure back on-premises or to a hybrid model. However, many still face issues with unclear cloud spending and unexpected cost spikes, despite using cloud cost management tools.
FinOps is becoming essential to help companies forecast budgets, measure success, and manage risks. As AI grows in importance, combining FinOps with AI governance offers new ways to improve both technology use and financial control.
Oracle Cloud set to surpass half a trillion dollars in booked orders as demand for AI and multi-cloud grows

Oracle announced strong growth in its cloud business, expecting booked cloud orders to surpass half a trillion dollars soon. This surge is driven by rising demand for Oracle Cloud Infrastructure (OCI), especially as companies seek cost-effective cloud and AI tools.
In its recent quarter, Oracle’s booked revenue jumped 359% to $455 billion, and the company expects more multi-billion-dollar deals to push this past $500 billion. Oracle forecasts OCI revenue to grow 77% to $18 billion this fiscal year and reach $144 billion over the next four years.
Oracle is attracting customers by integrating advanced AI models like ChatGPT directly with their databases in the Oracle Cloud. It has also partnered with Amazon, Google, and Microsoft to run OCI services within their cloud platforms, with revenue from these collaborations growing over 1,500%.
With plans to add more data centers and expand its multi-cloud offerings, Oracle is positioning itself as a competitive, flexible player in the cloud space. The company expects overall revenue growth of 12-14% next quarter, with cloud revenue growing even faster at 32-36%.
Media companies shift from fixed infrastructure to cloud-based, dynamic supply chains
Media organizations are moving away from costly, fixed on-premises infrastructure toward elastic cloud-based and hybrid supply chains. Traditional setups often left resources underutilized, while cloud architectures offer real-time scalability and consumption-based pricing that better align costs with actual workloads.
Hybrid models are emerging as the standard, giving companies operational control over sensitive functions while using cloud resources for demand spikes. Migration strategies typically start with less critical functions before scaling to core operations, supported by strong change management and training.
High-profile examples, like Tubi, utilized its cloud-based infrastructure built with Kubernetes on Amazon Web Services to stream the Super Bowl to 15.5 million concurrent viewers. Vendors report measurable results, including up to 60% workflow efficiency improvements. Analytics and monitoring tools are also key to optimizing scaling and cost management.
AWS strengthens cloud leadership with generative AI surge

Amazon Web Services (AWS) continues to lead the global cloud market with a 30-31% share, utilizing its broad range of cloud services and now strategically focusing on the booming generative AI sector. With 94% of businesses expected to adopt generative AI soon, AWS is rapidly growing by hosting complex AI workloads that need powerful computing and storage.
AWS’s generative AI offerings include Amazon Bedrock, a managed service accessing multiple advanced AI models; custom AI chips like Trainium for training and Inferentia for inference; and AI tools such as Amazon Q and SageMaker. These solutions help businesses build, deploy, and scale AI applications efficiently and cost-effectively.
Key highlights:
AWS expects generative AI to drive a 20% annual growth in cloud revenue through 2030.
Partnerships with leading AI developers and investments in AI chips boost AWS’s market strength.
AWS is expanding its data centers and investing over $100 billion in AI technologies and infrastructure in 2025.
Competitors like Microsoft Azure and Google Cloud remain strong, but AWS’s ecosystem and scale give it an edge.
AWS supports enterprises, startups, and hardware makers benefiting from the AI cloud boom.
Challenges include intense competition, regulatory scrutiny over market dominance, and the need to upskill workers for AI adoption.
AWS plans to train two million people globally in generative AI skills this year.
AWS’s dominant position in generative AI hosting is reshaping cloud services from basic computing to full AI ecosystems, enabling faster innovation and new business opportunities across industries while driving significant economic impact.
Summing up
Cloud costs are now at the center of business strategy, government policy, and AI innovation. September’s developments made it clear: whether it’s Microsoft driving government adoption, Alibaba using software to cut costs, or enterprises struggling with AI budgets, the cloud economy is evolving fast.
As we head into the next quarter, the winners will be the organizations that pair innovation with financial discipline, adopting FinOps practices, gaining visibility into AI spend, and staying ahead of the shifts in multi-cloud and sovereign cloud.
With Amnic, teams can grow confidently, knowing their cloud investments are aligned with business outcomes, not lost in runaway bills.
[Sign up for a free 30-day trial]
[Check out Amnic’s context-aware AI Agents]
[Download the free FinOps Resources]