Report:

Magic Quadrant for Cloud AI Developer Services

Top Vendor Products for Cloud AI Developer Services in 2023

• Amazon Web Services
• Google
• IBM
• Microsoft

Cloud AI Developer Services Vendor Product Comparisons 2023

• Amazon Web Services - Leader
  

  Pros:

  • Sales execution and pricing: AWS's approach to sales and pricing has attracted hundreds of thousands of customers to its CAIDS offering. In 2022, AWS introduced the AI Use Case Explorer that augments the AWS Solutions Library and solution architects to help customers and prospects select the right AI solutions for their industry and business-specific needs. AWS's CAIDS offering is competitively priced — there are no minimum fees or upfront commitments — and customers pay only for what they use.
  • Geographical strategy: AWS offers an extensive global cloud infrastructure. Its Availability Zones provide greater coverage than any of its competitors, and AWS makes it easy for customers to deploy their models across multiple Availability Zones. Its CAIDS offering is available in the Americas, EMEA and Asia/Pacific. AWS also offers language and autoML services in China. It plans to expand into Thailand and increase its presence in Australia and New Zealand.
  • Overall viability: AWS is a large multinational company with ample resources to expand its offerings to customers in all markets and regions. AWS's parent company, Amazon, reported that its net sales had increased by 9% to $514.0 billion in 2022, while AWS sales increased by 29% year over year (YoY) to $80.1 billion.

  Cons:

  • Product strategy: Though it has recently announced plans for a multicloud solution, AWS currently offers few options for customers seeking multicloud. If cloud portability is high among customers' priorities, they should consider vendors that can run natively on various clouds.
  • Innovation: AWS CAIDS users may need to augment their efforts with open-source software, leverage the recently announced Hugging Face partnership or visit the AWS Marketplace for popular ML products and services. An increase in research and development by AWS would boost its innovation and provide more visibility.
  • Market responsiveness: Although AWS's new AI Service Cards for Amazon Rekognition face matching, Amazon Textract AnalyzeID and Amazon Transcribe are excellent steps toward strengthening its responsible AI, it only provides three AI Service Cards at the time of writing. Since 2022, AWS has made progress in providing fairness, robustness and explainability capabilities with Amazon SageMaker Clarify and Model Monitor, but customers seeking turnkey responsible AI features may need to look into custom options.


• Google - Leader
  

  Pros:

  • Market responsiveness/record: Google Cloud's Vertex AI continues to gain momentum as the company enhances the capabilities of its services. The addition of more low-code tools makes the platform easier for non-data-scientists to use. GCP also continues to focus on targeted vertical markets and major accounts where Google services are well-received.
  • Overall viability: Google has extensive resources that it can devote to its CAIDS portfolio. The company has built out one of the strongest portfolios of language, vision, structured data and autoML services. Google offers individual prebuilt autoML services for language, structured data and vision, including both image and video. Google's DeepMind and Google Research advance AI via deep learning, unsupervised learning and generative AI, which will position the company to capitalize on advances in AI technology by quickly building them into enterprise-ready services for their customers.
  • Sales execution and pricing: The restructuring of the Google Cloud organization over the past few years is delivering results. The organization is focusing on named accounts and offerings tailored to selected verticals. As a result, GCP's offerings are gaining traction in the market. In addition to its own vertical solutions, Google has partnered with C3.ai and many other partners to deliver a portfolio of partner solutions to its customers.

  Cons:

  • Customer experience: Google's customer experience scores are still maturing as the company continues to invest in its sales and support functions. Recent organizational changes include the hiring of new leaders from other hyperscalers and ISVs.
  • Marketing strategy: Google has been less aggressive than other CAIDS providers in terms of promoting its offering. Google appears to have become more cautious in its approach to marketing as a result of public discourse regarding its AI practices.
  • Vertical/industry strategy: Google has recently expanded the number of industries served by its GCP portfolio inclusive of its CAIDS services. Market awareness is still maturing and Google's success with its new, vertical-specific solutions is yet to be seen.


• IBM - Leader
  

  Pros:

  • Product or service: Watson Discovery and Watson Natural Language Understanding provide strong capabilities for answering questions and generating narratives. IBM's vision services are competitive, especially its video content analysis capabilities. Maximo Visual Inspection includes deep-learning models that learn to analyze video streams for classification and to detect objects, action and anomalies. In addition to autoML support in Watson Studio AutoAI, IBM provides developers with responsible AI capabilities, including data quality, model lineage, model monitoring, drift, bias detection and explainability.
  • Market understanding: IBM's CAIDS portfolio delivers a strong set of capabilities to address a wide range of customer needs. Its user-friendly tools enable developers to deploy embedded AI services in any environment, including public clouds, on-premises and at the edge. IBM's CAIDS offering also provides low-code and no-code tools that enable developers and citizen developers to rapidly build and deploy AI solutions.
  • Operations: IBM's highly trained global technical support team provides 24/7 support. IBM provides best-in-class operational SLAs of 99.995% availability, with response times of less than 15 minutes for high-severity issues.

  Cons:

  • Sales strategy: IBM's sales strategy is driven by its existing partner relationships. In 2022, IBM launched IBM Partner Plus, a program that provides significant access to IBM competitive incentives, greater insider access and enhanced support.
  • Market responsiveness: IBM's offering is fairly comprehensive, but it has some gaps in its image generation and image labeling services. Its offering lacks crowdsourcing and synthetic data creation capabilities, as well as automated building of image and video content analysis models.
  • Sales execution and pricing: IBM offers an extensive range of deployment options that may have unclear costs. Customers should carefully assess the full cost implications for nonstandard deployments.


• Microsoft - Leader
  

  Pros:

  • Business model: Microsoft customers can access and extend a wide range of prebuilt models from Microsoft, OpenAI (via its exclusive agreement) and Hugging Face. It also provides customers with an extensive network of global partners. Microsoft's internal use of CAIDS across its product lines enables it to improve product quality and release features faster than its competitors.
  • Sales execution and pricing: Microsoft offers a free version of Azure AI, a pay-as-you-go tier for small teams, and discounted subscription plans for large enterprises (with fixed monthly or annual commitments). Microsoft offers its autoML capabilities for free, charging only for compute and storage. Microsoft also employs AI solution architects as part of its sales team. These 'AI rangers' create successful, deployable proofs of concept.
  • Market responsiveness/record: Microsoft continues to productize AI innovations for enterprise markets faster than its competitors. In addition to Microsoft's own AI products, customers can leverage marketplace solutions, which grew at a triple-digit rate in 2022. Microsoft also supports developers by providing free online courses, videos, tutorials and documentation. Microsoft is a leading contributor of open-source software (OSS) code, including SDKs for developing AI applications responsibly, and its subsidiary GitHub hosts most of its AI/ML OSS code for free.

  Cons:

  • Vertical/industry strategy: Microsoft has improved its industry coverage, but its model strategy and selection of vertical offerings are less comprehensive than those from other vendors. Customers should evaluate the depth and breadth of Azure AI's prebuilt models in their business domain.
  • Customer experience: Customer ratings for Microsoft's services have improved from last year, but remain lower than those of some other Leaders in this Magic Quadrant. To avoid frustrating its developer base, Microsoft needs to consolidate its CAIDS offerings into a single offering and respond better to requests for new features.
  • Operations: Microsoft's SLAs do not guarantee latency limits for CAIDS. This can be problematic when these services are embedded in mission-critical, real-time applications and when the workloads are computationally expensive. Also, Microsoft provides a maximum of 25% of service credit back for any CAIDS downtime below 99% uptime, while some other Leaders provide 100% of service credit to their customers in this scenario.


• Alibaba Cloud - Challenger
  

  Pros:

  • Innovation: Alibaba Cloud's CAIDS offering is highly customizable, often enabling the use of low-code tools and SDKs. Alibaba's ModelScope platform offers many prebuilt models that developers can easily modify to suit a wide variety of use cases across its language and vision services. Few-shot learning techniques further enhance these capabilities. Its structBERT language model also enhances the company's language services. Alibaba Cloud offers exceptionally strong translation capabilities, with support for more languages than any other provider.
  • Market understanding: Alibaba Cloud's portfolio is clearly driven by customer demand, as evidenced by its vision offering that focuses on major verticals, such as retail, media and logistics. The company continues to enhance its CAIDS portfolio and makes it easy for customers to tailor services to their use cases without needing data science skills. For example, Alibaba Cloud's image generation capabilities are easy to use and can customize a sample image to fit specific use cases.
  • Overall viability: Alibaba Cloud is a wholly owned subsidiary of Alibaba, a multinational conglomerate. The company has extensive resources to enter any market and pursue any opportunity. It also has the discipline to ensure that it can offer a comprehensive set of products and services before expanding into new markets or regions.

  Cons:

  • Product strategy: The platform can be challenging for developers to navigate. The comprehensiveness of the Alibaba Cloud CAIDS offering adds a level of complexity.
  • Geographical strategy: Alibaba Cloud has limited coverage outside China, the Middle East and Asia/Pacific. This narrow geographical focus is also reflected in some of the product limitations in its language services. Although its translation capabilities are very strong, its other language services and its product documentation support a relatively limited number of languages.
  • Marketing execution: Alibaba Cloud has strong marketing execution in China, but a minimal marketing presence outside the country. Though this strategy is a conscious choice by Alibaba Cloud, customers outside China may lack awareness of Alibaba's offerings.


• Baidu - Challenger
  

  Pros:

  • Innovation: Baidu plans to launch a generative AI and ChatGPT-style product called Ernie Bot for users in China. Baidu provides a wide range of innovative features, including autonomous driving, multimodal foundation models, quantum computing, graph computing and metaverse.
  • Sales execution and pricing: Baidu offers competitive prices compared with its competitors. It also offers comprehensive learning courses, an open-source platform (PaddlePaddle), data and free computing power to developers so that they can explore ideas via AI Studio. These features have led to a 60% increase in the number of paid users during 2022.
  • Business model: With its 'Believing Everyone Can AI' vision, Baidu aims to provide every developer with access to open-source AI technologies. The company is supporting this vision by fostering strong developer communities for PaddlePaddle (with more than 5 million developers) and via a partnership with the AI company Hugging Face, which began in 2022.

  Cons:

  • Geographical strategy: Baidu's AI solutions are geared primarily to developers in China. Some of its natural language offerings, like text analytics and sentiment analysis, are available only in Chinese. Its lack of non-Chinese language support may limit its appeal to customers outside China.
  • Product strategy: Baidu's business scale of data management solutions sold to the enterprises, such as database and data lake, is not as big as its direct competitors in China. Baidu's relatively low data gravity may make it difficult for the vendor to stand out in this market.
  • Marketing strategy: Baidu is not actively promoting responsible AI in its marketing messages, and thus has limited visibility to customers. Its planned launch of Ernie Bot, which includes enhanced bias detection and model explainability capabilities, will help, but is late compared with Leaders in this Magic Quadrant.


• Clarifai - Niche Player
  

  Pros:

  • Market understanding: As a pure-play CAIDS company, Clarifai has established two clear focuses that align with market demand. First, Clarifai provides developers with features to build AI-enabled business applications, including generative AI, APIs, accelerators, customizable pretrained models and end-to-end machine learning operations (MLOps) features. Second, Clarifai's offering helps to increase developer productivity through automation.
  • Product strategy: Clarifai plans to strengthen its offering further in 2023. Its roadmap includes ChatGPT's derived generative language services, workflows that combine text-to-audio with other NLP services, and joint embedding of vision and language for search and multimodal tasks. It also plans to release synthetic data generation, distributed training capabilities for deep-trained models and one-click training with autopilot mode.
  • Vertical/industry strategy: In 2022, Clarifai attained new certifications for its CAIDS offering, including Service Organization Control (SOC) 2/SOC 3 and General Data Protection Requirement (GDPR). This demonstrates a further maturation of its product to support regulated industries. Across verticals, Clarifai's customer growth has averaged 16%. This is reinforced by its investments in industry-focused solutions, such as intelligence surveillance and reconnaissance, visual inspection, video analysis for commercial real estate, and content moderation for media and entertainment.

  Cons:

  • Overall viability: Clarifai is a relatively small, venture-funded company that is competing with the cloud giants. Although Clarifai plans to expand its employee presence in North America, EMEA and Asia/Pacific in 2023, it must accelerate its investments in sales and support to scale up to the needs of large enterprises.
  • Geographical strategy: Clarifai's geographic reach is not comparable with Leaders in this Magic Quadrant. Its modest investment will not be sufficient to effectively build customer awareness in all regions.
  • Operations: Clarifai's employee headcount is low compared with that of its competitors. Large enterprises must work closely with Clarifai to ensure they get the level of support they expect.


• H2O.ai - Visionary
  

  Pros:

  • Market responsiveness/record: H2O.ai continues to innovate its CAIDS offerings. Over the past 12 months, H2O.ai has enhanced its core AI engines: Hydrogen Torch for deep learning, Document AI for unstructured data use cases and Driverless AI for greater accommodation of developers and data scientists. H2O.ai also emphasizes responsible AI by offering guardrails to protect against common ML pitfalls, including automatic model validation and model monitoring to ensure trust.
  • Market understanding: H2O.ai co-creates AI products with its customers. It brings its technology, AI expertise, go-to-market support, training and enablement, while customers contribute their domain acumen and data. For example, H2O.ai has co-created, with AT&T, the AI Feature Store, an in-database model deployment process with Snowflake, a solution for generating synthetic training data to protect privacy with Subsalt and fintech apps with the Commonwealth Bank of Australia. H2O.ai has also prioritized vertical solutions for healthcare, financial services, business operations and customer service.
  • Offering (product) strategy: H2O.ai helps developers to quickly build accurate, transparent AI models and applications with features such as model validation, a no-code deep learning framework in H2O Hydrogen Torch, and zero-shot and embedding models for unsupervised image and text data exploration.

  Cons:

  • Marketing execution: Most H2O.ai customers are data scientists, and only 25% are developers. H2O.ai's offering should be attractive to developers due to its easy-to-use SDK, interface, workflow and API experience, but its marketing efforts are not sufficient to attract more developers.
  • Operations: H2O.ai has invested to match the growth it is experiencing in Latin America, Australia and Singapore. H2O.ai's community support and self-service work well for users who are competent in ML, but most developers need additional guidance.
  • Sales execution: The vendor's Cloud AI Units are designed to accommodate the difference in consumption for training and inference, but this model could be confusing for customers. However, customers have an option of server-based licensing, and the ever-evolving H2O.ai open-source platform is free.


• Huawei Cloud - Niche Player
  

  Pros:

  • Overall viability: Huawei Cloud runs a profitable CAIDS business, and most of its sales result from organic customer growth. It continued to expand its partner relationships in 2022 and now has hundreds of partners, with plans to continue to expand in 2023. Its revenue grew by double digits in 2022, with a similar growth rate projected for 2023.
  • Marketing execution: Huawei Cloud leads developer-focused events such as domestic and international conferences, exhibitions and university programs. Its focus on a seamless, end-to-end developer experience is driving a high rate of adoption among engineers. Its thought leadership on enterprise knowledge management and acceleration of decision making using its AI and ML capabilities are resonating with its potential customers.
  • Operations: Huawei Cloud offers world-class operational SLAs of 99.9% availability, with response times as low as 15 minutes, depending upon the scale and severity of the incident. It has hundreds of support and operations personnel dedicated to the platform.

  Cons:

  • Market understanding: Huawei Cloud does not emphasize or advocate for the responsible use of AI as much as Leaders in this Magic Quadrant. Customers should ensure that the Huawei Cloud models they use and develop have acceptable levels of bias, per their organization's requirements.
  • Geographical strategy: Huawei Cloud is available in only a few countries and has a much smaller geographic reach than Leaders in this Magic Quadrant. Furthermore, not all CAIDS (language, vision and autoML services) are available in every country where Huawei Cloud is available. Customers should evaluate which of the three use cases are available in the countries where they need to deploy services.
  • Vertical/industry strategy: Huawei Cloud does not have the breadth of prebuilt industry-specific services and models as Leaders in this research. Customers should evaluate Huawei Cloud's offering to ensure it meets their industry-specific needs.


• Oracle - Niche Player
  

  Pros:

  • Industry/vertical strategy: Oracle offers solid support for prebuilt and industry-specific models. It also offers packaged AI solutions, such as OCI Anomaly Detection, which supports use cases such as fraud detection, marketing effectiveness, threat identification and root cause analysis.
  • Geographical strategy: Over the past year, Oracle has added support for more languages to its services, including sentiment analysis and translation services. Its vision, language and autoML services are available in the Americas, EMEA and Asia/Pacific.
  • Overall viability: Oracle's 2022 revenue exceeded its 2021 revenue, and its 2Q23 revenue is outpacing its 2Q22 revenue. Oracle's acquisition of Cerner Corporation is expected to bolster its position in the healthcare data science market for years to come.

  Cons:

  • Market responsiveness/record: Although Oracle improved its CAIDS offering from last year, some services were not available as of January 2023, including generative capabilities, text-to-speech, and champion/challenger analysis of production models, as examples.
  • Product or service: Although Oracle offers competitive AI services, there are areas that need improvement to better compete in the CAIDS market. To remain competitive, Oracle must quickly expand its language services to include text-to-speech and generative text capabilities and to improve its vision services in the areas of labeling and generative AI.
  • Business model: Oracle's recent Java licensing changes and audits have become problematic for some enterprise IT leaders and their procurement offices. Potential buyers will need to effectively prove the value of Oracle CAIDS to navigate the procurement climate.


• Tencent - Niche Player
  

  Pros:

  • Innovation: Tencent has greatly increased the range and quality of its prebuilt vertical solutions over the past year. The company has added prebuilt models for several industry verticals, including (but not limited to) education, energy, healthcare, transportation and logistics.
  • Overall viability: Tencent has grown in three key areas: partnerships, active customers and revenue. Over the past year, Tencent has considerably expanded its number of partnerships, which is driving increased revenue for Tencent's CAIDS offering. During the same period, Tencent's active customer count has grown by about 37% and its CAIDS product revenue has grown by about 26%.
  • Sales execution and pricing: Tencent offers flexible pricing options, including free trial periods, pay as you go, consumption-based subscriptions and prepaid agreements. Tencent's well-established sales channels and expansive partner ecosystem have boosted its ability to reach and service a growing customer base.

  Cons:

  • Geographical strategy: Tencent offers vision and language services in multiple countries, but its autoML services are almost entirely concentrated in China. Some aspects of its language services, including sentiment analysis and text analytics, are available only in Chinese and English.
  • Product strategy: Tencent's language services can distinguish speakers but currently lack biometric and speaker verification. However, its product roadmap includes plans to add these capabilities in 2023.
  • Product or service: Although Tencent has improved its product offerings in areas such as natural language generation (NLG), it is still lagging behind market leaders in other product areas. These include feature store, model explainability, bias detection and remediation, and model portfolio management.

Reference

• Gartner, Magic Quadrant for Cloud AI Developer Services, 22-May-2023, ID G00773391

View Market Definition
View Vendor Movements