Microsoft Cost Management updates—June 2024

Whether you’re a new student, a thriving startup, or the largest enterprise, you have financial constraints, and you need to know what you’re spending, where it’s being spent, and how to plan for the future. Nobody wants a surprise when it comes to the bill, and this is where Microsoft Cost Management comes in. 

We’re always looking for ways to learn more about your challenges and how Microsoft Cost Management can help you better understand where you’re accruing costs in the cloud, identify and prevent bad spending patterns, and optimize costs to empower you to do more with less. Here are a few of the latest improvements and updates. 

FOCUS 1.0 support in Exports

Cost card in Azure portal

Kubernetes cost views (New entry point)

Pricing updates on Azure.com

New ways to save money with Microsoft Cloud

Documentation updates

Before we dig into details, kudos to the FinOps foundation for successfully hosting FinOps X 2024 in San Diego, California last month. Microsoft participated as a platinum sponsor for a second consecutive year. Our team members enjoyed connecting with customers and getting insights into their FinOps practice. We also shared our vision of simplifying FinOps through AI, demonstrated in this short video—Bring your FinOps practice into the era of AI.

For all our updates from FinOps X 2024, refer to the blog post by my colleague, Michael Flanakin, who also serves in the FinOps Technical Advisory Council. 

FOCUS 1.0 support in exports 

As you may already know, the FinOps foundation announced the general availability of the FinOps Cost and Usage Specification (FOCUS) Version 1 in June 2024. We are thrilled to announce that you can get the newly released version through exports experience in the Microsoft Azure portal or the REST API. You can review the updated schema and the differences from the previous version in this Microsoft Learn article. We will continue to support the ability for you to export the preview version of the FOCUS dataset.

For all the datasets supported through exports and to learn more about the functionality, refer to our documentation.

Cost card in Azure portal 

You have always had the ability to estimate costs for Azure services using the pricing calculator so that you can better plan your expenses. Now, we are excited to announce the estimation capability within the Azure portal itself. Engineers now can quickly get a breakdown of their estimated virtual machine (VM) costs before deploying them and adjust as needed. This new experience is currently available only for VMs running on pay-as-you-go subscriptions and will be expanded in the future. Empowering engineers with cost data without disrupting their workflow enables them to make the right decisions for managing their spending and drives accountability.

Kubernetes cost views (new entry point) 

I had spoken about the Azure Kubernetes Service cost views in our November 2023 blog post. We know how important it is for you to get visibility into the granular costs of running your clusters. To make it even easier to access these cost views, we have added an entry point to the cluster page itself. Engineers and admins who are already on the cluster page potentially making configuration changes or just monitoring their cluster, can now quickly reference the costs as well.

Pricing updates on Azure.com

We’ve been working hard to make some changes to our Azure pricing experiences, and we’re excited to share them with you. These changes will help make it easier for you to estimate the costs of your solutions. 

We’ve expanded our global reach with pricing support for new Azure regions, including Spain Central and Mexico Central. 

We’ve introduced pricing for several new services—enhancing our Azure portfolio—including Trusted Signing, Azure Advanced Container Networking Services, Azure AI Studio, Microsoft Entra External ID, and Azure API Center (now available on the Azure API Management pricing calculator.)

The Azure pricing calculator now supports a new example to help you get started with estimating costs for your Azure Arc enabled servers scenarios.  

Azure AI has seen significant updates with pricing support for Basic Video Indexing Analysis for Azure AI Video Indexer, new GPT-4o models and improved Fine Tuning models for Azure OpenAI Service, the deprecation of S2 to S4 volume discount tiers for Azure AI Translator, and the introduction of standard fast transcription and video dubbing, both in preview, for Azure AI Speech.  

We’re thrilled to announce new features in both preview and general availability stages with Azure flex consumption (preview) for Azure Functions, Advanced messaging (generally available) for Azure Communication Services, and Azure API Center (generally available) for Azure API Management, and AKS Automatic (preview) for Azure Kubernetes.  

We’ve made comprehensive updates to our pricing models to reflect the latest offerings and ensure you have the most accurate information, including changes to

Azure Bastion: Added pricing for premium and developer stock-keeping units (SKUs).

Virtual Machines: Removal of CentOS for Linux, added 5 year reserved instances (RI) pricing for the Hx and HBv4 series, as well as pricing for the new NDsr H100 v5 and E20 v4 series.

Databricks: Added pricing for all-purpose serverless compute jobs.

Azure Communication Gateway: Added pricing for the new “Lab” SKU.

Azure Virtual Desktop for Azure Stack HCI: Pricing added to the Azure Virtual Desktop calculator.

Azure Data Factory: Added RI pricing for Dataflow.

Azure Container Apps: Added pricing for dynamic session feature.

Azure Backup: Added pricing for the new comprehensive Blob Storage data protection feature.

 Azure SQL Database: Added 3 year RI pricing for hyperscale series, zone redundancy pricing for hyperscale elastic pools, and disaster recovery pricing options for single database.

Azure PostgreSQL: Added pricing for Premium SSD v2.

Defender for Cloud: Added pricing for the “Pre-Purchase Plan”.

Azure Stack Hub: Added pricing for site recovery.

Azure Monitor: Added pricing for pricing for workspace replication as well as data restore in the pricing calculator.

We’re constantly working to improve our pricing tools and make them more accessible and user-friendly. We hope you find these changes helpful in estimating the costs for your Azure solutions. If you have any feedback or suggestions for future improvements, please let us know! 

New ways to save money in the Microsoft Cloud 

VM Hibernation is now generally available 

Documentation updates 

Here are a few documentation updates you might be interested in: 

Update: Understand Cost Management data  

Update: Azure Hybrid Benefit documentation 

Update: Automation for partners  

Update: View and download your Microsoft Azure invoice  

Update: Tutorial: Create and manage exported data  

Update: Automatically renew reservations  

Update: Changes to the Azure reservation exchange policy  

Update: Migrate from EA Marketplace Store Charge API

Update: Azure product transfer hub  

Update: Get started with your Microsoft Partner Agreement billing account  

Update: Manage billing across multiple tenants using associated billing tenants

Want to keep an eye on all documentation updates? Check out the Cost Management and Billing documentation change history in the azure-docs repository on GitHub. If you see something missing, select Edit at the top of the document and submit a quick pull request. You can also submit a GitHub issue. We welcome and appreciate all contributions! 

What’s next? 

These are just a few of the big updates from last month. Don’t forget to check out the previous Microsoft Cost Management updates. We’re always listening and making constant improvements based on your feedback, so please keep the feedback coming.  

Follow @MSCostMgmt on X and subscribe to the Microsoft Cost Management YouTube channel for updates, tips, and tricks.
The post Microsoft Cost Management updates—June 2024 appeared first on Azure Blog.
Quelle: Azure

Harnessing the full power of AI in the cloud: The economic impact of migrating to Azure for AI readiness

As the digital landscape rapidly evolves, AI stands at the forefront, driving significant innovation across industries. However, to fully harness the power of AI, businesses must be AI-ready; this means having defined use-cases for their AI apps, being equipped with modernized databases that seamlessly integrate with AI models, and most importantly, having the right infrastructure in place to power and realize their AI ambitions. When we talk to our customers, many have expressed that traditional on-premises systems often fall short in providing the necessary scalability, stability, and flexibility required for modern AI applications.

A recent Forrester study1, commissioned by Microsoft, surveyed over 300 IT leaders and interviewed representatives from organizations globally to learn about their experience migrating to Azure and if that enhanced their AI impact. The results showed that migrating from on-premises infrastructure to Azure can support AI-readiness in organizations, with lower costs to stand up and consume AI services plus improved flexibility and ability to innovate with AI. Here’s what you should know before you start leveraging AI in the cloud.

Challenges faced by customers with on-premises infrastructure

Many organizations who attempted to implement AI on-premises encountered significant challenges with their existing infrastructure. The top challenges with on-premises infrastructure cited were:

Aging and costly infrastructure: Maintaining or replacing aging on-premises systems is both expensive and complex, diverting resources from strategic initiatives.

Infrastructure instability: Unreliable infrastructure impacts business operations and profitability, creating an urgent need for a more stable solution.

Lack of scalability: Traditional systems often lack the scalability required for AI and machine learning (ML) workloads, necessitating substantial investments for infrequent peak capacity needs.

High capital costs: The substantial upfront costs of on-premises infrastructure limit flexibility and can be a barrier to adopting new technologies.

Forrester’s study highlights that migrating to Azure effectively addresses these issues, enabling organizations to focus on innovation and business growth rather than infrastructure maintenance.

Azure AI
Where innovators are creating for the future

Try for free today

Key Benefits

Improved AI-readiness: When asked whether being on Azure helped with AI-readiness, 75% of survey respondents with Azure infrastructure reported that migrating to the cloud was essential or significantly reduced barriers to AI and ML adoption. Interviewees noted that the AI services are readily available in Azure, and colocation of data and infrastructure that is billed only on consumption helps teams test and deploy faster with less upfront costs. This was summarized well by an interviewee who was the head of cloud and DevOps for a banking company:

We didn’t have to go and build an AI capability. It’s up there, and most of our data is in the cloud as well. And from a hardware-specific standpoint, we don’t have to go procure special hardware to run AI models. Azure provides that hardware today.”
—Head of cloud and DevOps for global banking company

Cost Efficiency: Migrating to Azure significantly reduces the initial costs of deploying AI and the cost to maintain AI, compared to on-premises infrastructure. The study estimates that organizations experience financial benefits of USD $500 thousand plus over three years and 15% lower costs to maintain AI/ML in Azure compared to on-premises infrastructure.

Flexibility and scalability to build and maintain AI: As mentioned above, lack of scalability was a common challenge for survey respondents with on-premises infrastructure as well. Respondents with on-premises infrastructure cited lack of scalability with existing systems as a challenge when deploying AI and ML at 1.5 times the rate of those with Azure cloud infrastructure.

Interviewees shared that migrating to Azure gave them easy access to new AI services and the scalability they needed to test and build them out without worrying about infrastructure. 90% of survey respondents with Azure cloud infrastructure agreed or strongly agreed they have the flexibility to build new AI and ML applications. This is compared to 43% of respondents with on-premises infrastructure. A CTO for a healthcare organization said:

After migrating to Azure all the infrastructure problems have disappeared, and that’s generally been the problem when you’re looking at new technologies historically.”
—CTO for a healthcare organization

They explained that now, “The scalability [of Azure] is unsurpassed, so it adds to that scale and reactiveness we can provide to the organization.” They also said: “When we were running on-prem, AI was not as easily accessible as it is from a cloud perspective. It’s a lot more available, accessible, and easy to start consuming as well. It allowed the business to start thinking outside of the box because the capabilities were there.”

Holistic organizational improvement: Beyond the cost and performance benefits, the study found that migration to Azure accelerated innovation with AI by having an impact on the people at all levels of an organization:

Bottoms-up: skilling and reinvestment in employees. Forrester has found that investing in employees to build understanding, skills, and ethics is critical to successfully using AI. Both interviewees and survey respondents expressed difficulty finding skilled resources to support AI and ML initiatives at their organizations.

Migrating to the cloud freed up resources and changed the types of work needed, allowing organizations to upskill employees and reinvest resources in new initiatives like AI. A VP of AI for a financial services organization shared: “As we have gone along this journey, we have not reduced the number of engineers as we have gotten more efficient, but we’re doing more. You could say we’ve invested in AI, but everything we have invested—my entire team—none of these people were new additions. These are people we could redeploy because we’re doing everything else more efficiently.”

Top-down: created a larger culture of innovation at organizations. As new technologies—like AI—disrupt entire industries, companies need to excel at all levels of innovation to succeed, including embracing platforms and ecosystems that help drive innovation. For interviewees, migrating to the cloud meant that new resources and capabilities were readily available, making it easier for organizations to take advantage of new technologies and opportunities with reduced risk.

Survey data indicates that 77% of respondents with Azure cloud infrastructure find it easier to innovate with AI and ML, compared to only 34% of those with on-premises infrastructure. An executive head of cloud and DevOps for a banking organization said: “Migrating to Azure changes the mindset from an organization perspective when it comes to innovation, because services are easily available in the cloud. You don’t have to go out to the market and look for them. If you look at AI, originally only our data space worked on it, whereas today, it’s being used across the organization because we were already in the cloud and it’s readily available.”

Learn more about migrating to Azure for AI-readiness

Forrester’s study underscores the significant economic and strategic advantages of migrating to Azure for be AI-ready. Lower costs, increased innovation, better resource allocation, and improved scalability make migration to Azure a clear choice for organizations looking to thrive in the AI-driven future.

Ready to get started with your migration journey? Here are some resources to learn more:

Read the full Forrester TEI study on migration to Azure for AI-readiness.

The solutions that can support your organization’s migration and modernization goals.

Our hero offerings that provide funding, unique offers, expert support, and best practices for all use-cases, from migration to innovation with AI.

Learn more in our e-book and video on how to migrate to innovate.

Refrences

Forrester Consulting The Total Economic Impact™ Of Migrating to Microsoft Azure For AI-Readiness, commissioned by Microsoft, June 2024

The post Harnessing the full power of AI in the cloud: The economic impact of migrating to Azure for AI readiness appeared first on Azure Blog.
Quelle: Azure

Announcing Phi-3 fine-tuning, new generative AI models, and other Azure AI updates to empower organizations to customize and scale AI applications

AI is transforming every industry and creating new opportunities for innovation and growth. But, developing and deploying AI applications at scale requires a robust and flexible platform that can handle the complex and diverse needs of modern enterprises and allow them to create solutions grounded in their organizational data. That’s why we are excited to announce several updates to help developers quickly create customized AI solutions with greater choice and flexibility leveraging the Azure AI toolchain:

Serverless fine-tuning for Phi-3-mini and Phi-3-medium models enables developers to quickly and easily customize the models for cloud and edge scenarios without having to arrange for compute.

Updates to Phi-3-mini including significant improvement in core quality, instruction-following, and structured output, enabling developers to build with a more performant model without additional cost.

Same day shipping earlier this month of the latest models from OpenAI (GPT-4o mini), Meta (Llama 3.1 405B), Mistral (Large 2) to Azure AI to provide customers greater choice and flexibility.

Unlocking value through model innovation and customization  

In April, we introduced the Phi-3 family of small, open models developed by Microsoft. Phi-3 models are our most capable and cost-effective small language models (SLMs) available, outperforming models of the same size and next size up. As developers look to tailor AI solutions to meet specific business needs and improve quality of responses, fine-tuning a small model is a great alternative without sacrificing performance. Starting today, developers can fine-tune Phi-3-mini and Phi-3-medium with their data to build AI experiences that are more relevant to their users, safely, and economically.

Given their small compute footprint, cloud and edge compatibility, Phi-3 models are well suited for fine-tuning to improve base model performance across a variety of scenarios including learning a new skill or a task (e.g. tutoring) or enhancing consistency and quality of the response (e.g. tone or style of responses in chat/Q&A). We’re already seeing adaptations of Phi-3 for new use cases.

Phi-3 models
A family of powerful, small language models (SLMs) with groundbreaking performance at low cost and low latency

Try today

Microsoft and Khan Academy are working together to help improve solutions for teachers and students across the globe. As part of the collaboration, Khan Academy uses Azure OpenAI Service to power Khanmigo for Teachers, a pilot AI-powered teaching assistant for educators across 44 countries and is experimenting with Phi-3 to improve math tutoring. Khan Academy recently published a research paper highlighting how different AI models perform when evaluating mathematical accuracy in tutoring scenarios, including benchmarks from a fine-tuned version of Phi-3. Initial data shows that when a student makes a mathematical error, Phi-3 outperformed most other leading generative AI models at correcting and identifying student mistakes.

And we’ve fine-tuned Phi-3 for the device too. In June, we introduced Phi Silica to empower developers with a powerful, trustworthy model for building apps with safe, secure AI experiences. Phi Silica builds on the Phi family of models and is designed specifically for the NPUs in Copilot+ PCs. Microsoft Windows is the first platform to have a state-of-the-art small language model (SLM) custom built for the Neural Processing Unit (NPU) and shipping inbox.

You can try fine-tuning for Phi-3 models today in Azure AI.

I am also excited to share that our Models-as-a-Service (serverless endpoint) capability in Azure AI is now generally available. Additionally, Phi-3-small is now available via a serverless endpoint so developers can quickly and easily get started with AI development without having to manage underlying infrastructure. Phi-3-vision, the multi-modal model in the Phi-3 family, was announced at Microsoft Build and is available through Azure AI model catalog. It will soon be available via a serverless endpoint as well. Phi-3-small (7B parameter) is available in two context lengths 128K and 8K whereas Phi-3-vision (4.2B parameter) has also been optimized for chart and diagram understanding and can be used to generate insights and answer questions.

We are seeing great response from the community on Phi-3. We released an update for Phi-3-mini last month that brings significant improvement in core quality and instruction following. The model was re-trained leading to substantial improvement in instruction following and support for structured output. We also improved multi-turn conversation quality, introduced support for <|system|> prompts, and significantly improved reasoning capability.

The table below highlights improvements across instruction following, structured output, and reasoning.

Benchmarks Phi-3-mini-4k Phi-3-mini-128k Apr ’24 release Jun ’24 update Apr ’24 release Jun ’24 update Instruction Extra Hard 5.7 6.0 5.7 5.9 Instruction Hard 4.9 5.1 5 5.2 JSON Structure Output 11.5 52.3 1.9 60.1 XML Structure Output 14.4 49.8 47.8 52.9 GPQA 23.7 30.6 25.9 29.7 MMLU 68.8 70.9 68.1 69.7 Average 21.7 35.8 25.7 37.6 

We continue to make improvements to Phi-3 safety too. A recent research paper highlighted Microsoft’s iterative “break-fix” approach to improving the safety of the Phi-3 models which involved multiple rounds of testing and refinement, red teaming, and vulnerability identification. This method significantly reduced harmful content by 75% and enhanced the models’ performance on responsible AI benchmarks. 

Expanding model choice, now with over 1600 models available in Azure AI

With Azure AI, we’re committed to bringing the most comprehensive selection of open and frontier models and state-of-the-art tooling to help meet customers’ unique cost, latency, and design needs. Last year we launched the Azure AI model catalog where we now have the broadest selection of models with over 1,600 models from providers including AI21, Cohere, Databricks, Hugging Face, Meta, Mistral, Microsoft Research, OpenAI, Snowflake, Stability AI and others. This month we added—OpenAI’s GPT-4o mini through Azure OpenAI Service, Meta Llama 3.1 405B, and Mistral Large 2.

Continuing the momentum today we are excited to share that Cohere Rerank is now available on Azure. Accessing Cohere’s enterprise-ready language models on Azure AI’s robust infrastructure enables businesses to seamlessly, reliably, and safely incorporate cutting-edge semantic search technology into their applications. This integration allows users to leverage the flexibility and scalability of Azure, combined with Cohere’s highly performant and efficient language models, to deliver superior search results in production.

TD Bank Group, one of the largest banks in North America, recently signed an agreement with Cohere to explore its full suite of large language models (LLMs), including Cohere Rerank.

At TD, we’ve seen the transformative potential of AI to deliver more personalized and intuitive experiences for our customers, colleagues and communities, we’re excited to be working alongside Cohere to explore how its language models perform on Microsoft Azure to help support our innovation journey at the Bank.”
Kirsti Racine, VP, AI Technology Lead, TD.

Atomicwork, a digital workplace experience platform and longtime Azure customer, has significantly enhanced its IT service management platform with Cohere Rerank. By integrating the model into their AI digital assistant, Atom AI, Atomicwork has improved search accuracy and relevance, providing faster, more precise answers to complex IT support queries. This integration has streamlined IT operations and boosted productivity across the enterprise. 

The driving force behind Atomicwork’s digital workplace experience solution is Cohere’s Rerank model and Azure AI Studio, which empowers Atom AI, our digital assistant, with the precision and performance required to deliver real-world results. This strategic collaboration underscores our commitment to providing businesses with advanced, secure, and reliable enterprise AI capabilities.”
Vijay Rayapati, CEO of Atomicwork

Command R+, Cohere’s flagship generative model which is also available on Azure AI, is purpose-built to work well with Cohere Rerank within a Retrieval Augmented Generation (RAG) system. Together they are capable of serving some of the most demanding enterprise workloads in production. 

Earlier this week, we announced that Meta Llama 3.1 405B along with the latest fine-tuned Llama 3.1 models, including 8B and 70B, are now available via a serverless endpoint in Azure AI. Llama 3.1 405B can be used for advanced synthetic data generation and distillation, with 405B-Instruct serving as a teacher model and 8B-Instruct/70B-Instruct models acting as student models. Learn more about this announcement here.

Mistral Large 2 is now available on Azure, making Azure the first leading cloud provider to offer this next-gen model. Mistral Large 2 outperforms previous versions in coding, reasoning, and agentic behavior, standing on par with other leading models. Additionally, Mistral Nemo, developed in collaboration with NVIDIA, brings a powerful 12B model that pushes the boundaries of language understanding and generation. Learn More.

And last week, we brought GPT-4o mini to Azure AI alongside other updates to Azure OpenAI Service, enabling customers to expand their range of AI applications at a lower cost and latency with improved safety and data deployment options. We will announce more capabilities for GPT-4o mini in coming weeks. We are also happy to introduce a new feature to deploy chatbots built with Azure OpenAI Service into Microsoft Teams.  

Enabling AI innovation safely and responsibly  

Building AI solutions responsibly is at the core of AI development at Microsoft. We have a robust set of capabilities to help organizations measure, mitigate, and manage AI risks across the AI development lifecycle for traditional machine learning and generative AI applications. Azure AI evaluations enable developers to iteratively assess the quality and safety of models and applications using built-in and custom metrics to inform mitigations. Additional Azure AI Content Safety features—including prompt shields and protected material detection—are now “on by default” in Azure OpenAI Service. These capabilities can be leveraged as content filters with any foundation model included in our model catalog, including Phi-3, Llama, and Mistral. Developers can also integrate these capabilities into their application easily through a single API. Once in production, developers can monitor their application for quality and safety, adversarial prompt attacks, and data integrity, making timely interventions with the help of real-time alerts.

Azure AI uses HiddenLayer Model Scanner to scan third-party and open models for emerging threats, such as cybersecurity vulnerabilities, malware, and other signs of tampering, before onboarding them to the Azure AI model catalog. The resulting verifications from Model Scanner, provided within each model card, can give developer teams greater confidence as they select, fine-tune, and deploy open models for their application. 

We continue to invest across the Azure AI stack to bring state of the art innovation to our customers so you can build, deploy, and scale your AI solutions safely and confidently. We cannot wait to see what you build next.

Stay up to date with more Azure AI news

Watch this video to learn more about Azure AI model catalog.

Listen to the podcast on Phi-3 with lead Microsoft researcher Sebastien Bubeck.

The post Announcing Phi-3 fine-tuning, new generative AI models, and other Azure AI updates to empower organizations to customize and scale AI applications appeared first on Azure Blog.
Quelle: Azure