The global race for the OS of the future

  • Post by Rachel
  • Sep 07, 2020

Artificial Intelligence (AI) is at the heart of a fierce arms race taking place among the biggest public cloud providers, Amazon, Google & Microsoft.

According to the MIT Technology Review whoever dominates “Cloud AI” will own the operating system of the next era of tech and become the most powerful company in history. No wonder that, aside from the biggest cloud players, others like Apple, IBM, Oracle, Adobe, Salesforce, SAP, Alibaba, Baidu and Tencent have joined the fray.

AI was once an obscure corner of computer science. But over the past decade, an increase in digital data availability, exponential growth in compute power and algorithm breakthroughs have fuelled global excitement about AI. AI (and machine learning in particular) has been one of the fastest growing technologies in recent years, poised to unleash the next wave of digital disruption.

“Whoever manages to dominate the business of providing artificial-intelligence services through cloud computing may have the operating system of the future and reign over what may be the most lucrative technology mega-trend yet.”

-“How the AI Cloud Could Produce the Richest Companies Ever” by Peter Burrows in MIT Technology Review (March 2018)

Despite the COVID-fuelled slowdown, worldwide revenues for the AI market (including software, hardware and services) are expected to total $156.5 billion in 2020, an increase of 12.3% over 2019. BY 2024, total revenues are expected to surpass $300 billion. According to industry reports, the global “cloud AI” market is projected to reach $53.06 billion by 2026, with a CAGR of 35.4% from 2019 to 2026.

What’s Cloud AI?

Today, big corporations across every sector, from retail and manufacturing to healthcare and agriculture, are moving from AI experimentation to industrialisation and trying to integrate machine learning into their products and operations. However, while the technical and economic potential of AI is vast, wide scale AI adoption is still a challenge. Common hurdles such as limited in-house know-how, skills gap, poor data quality and large scale implementation are tricky to overcome.

On the other hand, company leaders know that AI is predicted to have a substantial impact on the world’s economy. According to McKinsey, companies that fully adopt AI could double their cash flow by 2030, while companies that don’t could see a 20% decline. So AI is both a source of major untapped opportunity and an existential threat which companies can’t afford to ignore. Cloud AI (aka “AI-as-a-Service” or simply “AIaaS”) is a neat solution to these problems. It’s been hailed as the next big thing in tech as it is set to play a huge role in AI adoption globally.

“Ultimately, the cloud is how most companies are going to make use of AI and how technology suppliers are going to make money off of it”

-Nick McQuire, CCS Insight

Big tech - Alphabet, Amazon, Apple, Facebook, IBM and Microsoft - have long been leading the AI charge, investing huge sums to develop their own AI capabilities, as have their counterparts in China. It’s difficult to separate tech firms’ investments in AI from other kinds, but McKinsey estimated a couple of years ago that, globally, tech giants spent $20 billion to $30 billion on AI in 2016 alone.

Similarly to cloud computing, where platforms like Amazon’s AWS, Microsoft’s Azure and Google’s GCP enable access to on-demand technology services, such as computing power, storage, and databases, with cloud AI, leading cloud providers are leveraging their infrastructure efficiencies and extensive AI prowess to provide AIaaS to companies and developers, making it easier, faster and more cost-effective to build AI applications and capabilities without the need to have deep AI expertise.

Companies across all sectors - including Disney, T-Mobile, The New York Times, 20th Century Fox, AB InBev, BP, Walgreens Boots Alliance and ASOS - are already using these platforms to tap into capabilities such as image recognition, natural-language processing and voice technologies that power product features currently used by Google, Amazon and Microsoft on their own platforms.

The scale, breadth and potential of these AI platforms mean they are fast emerging as the operating systems of the future, much like AWS became the OS of the internet (after Amazon realised it could leverage the infrastructure expertise it built in order to tackle its own business needs in the early 2000’s).

“AIaaS is gaining momentum precisely because AI-based solutions can be economically used as a service by many companies for many purposes. Those companies that deliver AI-based solutions targeting specific needs understand vertical industries and build sophisticated models to find actionable information with remarkable efficiency. Thanks to the cloud, providers are able to deliver these AI solutions as a service that can be accessed, refined and expanded in ways that were unfathomable in the past.”

-Get ready for the emergence of AI-as-a-Service (January 2020)

The Cloud AI Race

“If you really want machine learning to be expansive across companies, you have to find a way to let everyday developers build machine-learning models and put them in production. We wanted to make that easy for developers to take advantage of because that’s where all the innovation is going to happen . . . We said, how are they going to get hands-on experience and actually try it? I don’t know if it’ll be three years from now or five years from now or 10 years from now but the vast majority of applications will have machine learning and artificial intelligence fuelling them.”

- Andy Jassy, CEO of AWS

For the past few years, the ‘big three’ (Amazon, Google and Microsoft) have been playing a game of leapfrog in the Cloud AI space, with AWS continuously signalling that it had no intention to cede this market to Google or Microsoft.

Amazon’s AI platform push reflects how the bleeding edge of competition in the multibillion-dollar cloud market has evolved from its original business of data centres renting out storage and computing capacity to courting developers and businesses which lack the resources or appetite to develop the underlying machine learning and AI technology themselves.

“It is clear to me we are evolving from a mobile-first to an AI-first world.”

- Google CEO, Sundar Pichai

While Amazon clearly has a first-mover advantage in the $270bn global public cloud space, Microsoft and Google are using their expertise in building applications and developing AI to nip at AWS’s heels. Google has successfully established itself as a major player in the cloud space, with its cloud products increasingly gaining market share despite still lagging Amazon and Microsoft. Google realised that customers are seeking expertise in machine learning, an area where Google remains differentiated, and was quick to capture some business from its competitors.

“Ultimately the cloud is about powering the next generation of applications. It is always the next generation applications that have driven infrastructure and when we look at this current generation of applications that people are building, the thing that is going to define these applications, that characterises these applications, is machine learning and artificial intelligence. Therefore we are building out Azure as the first AI supercomputer."

-Microsoft CEO, Satya Nadella

Amazon first announced the launch of its new Amazon AI platform at its developer event, re:Invent, in Las Vegas in December 2016, promising to bring many of the machine learning smarts Amazon has developed in-house over the years to devs outside the company. Just a week after AWS unveiled an update to its GPU-powered cloud computing service, targeting AI applications that need vast amounts of parallel processing power, Microsoft CEO Satya Nadella discussed how Microsoft’s cloud computing offering underpinned a new wave of applications that use AI technologies. Nadella also said Microsoft was offering higher level services that help build AI services, such as APIs to connect to speech, image, object recognition and natural language processing services.

Google was fast to catch up - in 2017 Google Cloud Machine Learning Engine was launched to help developers with machine learning expertise to easily build ML models. Google also unveiled several AI APIs, including Vision, Speech, NLP, Translation and Dialogflo, which could be used to bring unmatched scale and speed to business applications. In July 2017, Microsoft announced plans to turn Azure into an AI Cloud calling it “a major step toward democratizing AI”. Meanwhile, A few weeks after AWS SageMaker was announced in November 2017, Google introduced Cloud AutoML.

But AWS stole everyone’s the thunder when it took machine learning to the masses in December 2018 with much fanfare when it unveiled DeepRacer, a toy car that developers can buy (on Amazon, for $399) and teach to drive using machine learning models to compete against other cars in a newly created autonomous racing league. Andy Jassy, AWS chief executive, said DeepRacer embodied the goal of making machine-learning technology approachable. More than a dozen machine-learning tools and services followed, promising to make the technology cheaper and easier for anyone to use. In 2019, Amazon launched another 5 new AI services.

Fast forward to March 2020, Google launched Cloud AI Platform Pipelines - in beta - to simplify machine learning development via a service designed to deploy robust, repeatable AI pipelines along with monitoring, auditing, version tracking, and reproducibility in the cloud.

By May 2020, Microsoft announced at its Build developers conference that it has built one of the top five publicly disclosed supercomputers in the world, making new infrastructure available in Azure so its customers can train extremely large artificial intelligence models. This was done in collaboration with OpenAI on their mission to build safe artificial general intelligence.

Google hit back in September 2020 saying it was planning to launch new capabilities for its Cloud AI platform, including new MLOps features to provide automation and monitoring at all steps of ML system construction, including integration, testing, releasing, deployment and infrastructure management and industry-specific solutions for application developers such as Lending DocAI (a new, specialised solution for the mortgage industry, that processes borrowers’ income and asset documents to speed-up loan applications), retail forecasting, anti-money laundering, know your customer, healthcare and media asset management.

Winning the Cloud AI race

“So far, none of this activity has resulted in much in the way of revenue…But that would quickly change for the company that creates the underlying technologies and developer tools to support the widespread commercialization of machine learning. That’s what Microsoft did for the PC, by creating a Windows platform that millions of developers used to build PC programs. Apple did the same with iOS, which spawned the mobile-app era.”

-“How the AI Cloud Could Produce the Richest Companies Ever” by Peter Burrows, MIT Technology Review

It remains to be seen who will emerge as the Cloud AI winner. Especially, as the battle for AI supremacy is more global than you might think. US-based tech giants aren’t the only ones vying for dominance. The Chinese technology ecosystem has become a powerhouse in its own right - China’s multi-billion dollar tech giants Alibaba, Baidu, Tencent, and Huawei Technologies, are each heavily investing in AI.

Alibaba Cloud, Alibaba’s ML cloud platform, branded as “China’s first AI platform” to compete with Amazon and Microsoft, has been around since 2015 and recently partnered with Unilever to enable next generation marketing campaigns.

However, according to recent research, it is Baidu (“China’s Google”) who boasts the highest market share in China’s fast-growing AI public cloud services market, reflecting the company’s continuous investments in building core AI technologies and new infrastructure to power industrial transformation.

The research, published in July 2020, says that Baidu Cloud offers the most overall AI products and AI capabilities, particularly in natural language processing (NLP) and intelligent voice, facial and body recognition, conversational AI, and machine learning. In total, Baidu Cloud has opened up more than 250 AI capabilities that service 1.9 million developers, who cumulatively use these capabilities 1 trillion times per day on average. Much like Google, the sheer amount of data that Baidu has as China’s most popular search engine makes it a very strong player in the AI market.

While Baidu has profound ML knowledge, experience and talent, it’s still not that well known outside China and will require a better go-to-market strategy to accelerate ecosystem expansion, especially in enterprise space.

“As an AI platform company that is focused on empowering other organisations, Baidu hopes that every enterprise, no matter how small, is able to use capabilities and services provided by our platforms, just like they were water and electricity, to quickly and easily realise intelligent transformation”

- Robin Li, Baidu Co-Founder, Chairman and CEO, 2020 World Artificial Intelligence Conference (WAIC), July 9, 2020.

Globally, AWS is likely to be the current leader in the ‘cloud AI’ segment thanks to its global market share in the public cloud arena. But, Microsoft obviously has unparalleled experience in enterprise software, and Google’s major strength is its profound R&D capability - having some of the world’s leading AI researchers (like Geoffrey Hinton) in its ranks definitely helps.

None of the tech giants tends to break out their cloud revenues so it’s hard to tell how their respective AI platforms are doing and what their exact market share is in that space. As of July 2020, it’s estimated that AWS is dominating cloud with 31% of the market ($10.8 billion for Q2 2020 revenues) , Azure at 20%, Google Cloud at 6%, and Alibaba Cloud close behind at 5%.

We know that in 2019 Azure AI had more than 20,000 customers and over 85% of Fortune 100 companies have used it in the previous 12 months so it is clearly making strides. Google Cloud meanwhile generated $3 billion in revenue for the same period – a growth of 43% year-over-year - but again it’s not clear what part of that number is attributable to its AI platform.

Because of the nature of machine learning, the more data the system gets, the better its decisions (and therefore the product’s overall quality), so customers are more likely to stay with their initial ML solution vendor.

Additionally, the cost implications of using multiple AI vendors and complexity around transferring data between cloud vendors and potential impact on the product’s latency and performance, all mean that companies would likely stick to one vendor. Developers would also prefer to build apps and tools on platforms already used by their potential customers.

Ultimately, whoever gets the early lead in the cloud AI space will be very difficult to unseat. One thing’s certain though - it’ll be an interesting race to watch in the next few years. Stay tuned.