As cloud spending on Internet enterprise software tightens, slowing growth is becoming a dark cloud over the heads of cloud vendors.

ChatGPT has come out of nowhere to break this bottleneck, and AI will reinvent software. Software companies, the customers of cloud vendors, are aggressively embedding AI capabilities from big models into existing workflows to accomplish higher-order automation.

With new cloud customers drying up, software companies are no longer going to the cloud for the sake of going to the cloud, but are seeking to improve productivity with AI. "This is the biggest increment in the cloud computing market over the next decade. Algorithmic infrastructure is the absolute dividend beneficiary of the big model." a cloud computing industry insider who has been in the business for more than a decade elaborated to Geek Park.

With such a prospect, several overseas cloud service giants-Microsoft, Amazon, Google, and Oracle-have quickly made changes. In the past few months, the cloud giants have been dropping real money to develop big models, strategic investments, and self-research AI chips ...... The era of big models is on the rise, and they have targeted a new generation of AI software customers.

The former mountains are far from unbreakable, the cloud market is rapidly reshuffling, the giants have opened a whole new curtain of competition.

After all, the fall of the big brother in the mobile Internet era is just around the corner, and Nokia went from 70% of the cell phone market in its heyday to no one's business in a few years, just a matter of making the wrong decision. And for the big model, the cloud industry is quickly forming a consensus: AI is by no means a small variable this time around, and the current leading players could be left behind from the speed of the industry's day-to-day development.

Halfway through 2023, this article will be around several overseas cloud giants to sort out what is the key to competition among cloud vendors today?

01Develop AI-specific chips, can not give "life" all to Nvidia

After the advent of the era of big models, the most scarce resource for cloud providers today is computing power, or AI chips. Investing in the lowest level of infrastructure - AI acceleration chips - has also become the first focus of cloud vendors' competition today.

Scarcity and high cost are seen as the primary reasons for cloud vendors to accelerate their own chip development. Even a powerful man in the tech world like Musk commented that "this thing (Nvidia GPU) is harder to get than drugs," and secretly bought 10,000 cards from Nvidia for his AI company X.AI, as well as a lot of idle equity from Oracle.

Such a degree of scarcity, reflected in the business of the cloud giant, corresponds directly to the loss of business from the "neck". Even Microsoft, which was the first to strike, has been exposed to rumors of a GPU rationing system for its internal AI R&D team due to a GPU shortage, delays in various new programs, and months of queuing for new customers to get on Azure.

Even venture capital institutions to grab the project, are dependent on the hands of Nvidia chip inventory. In order to N-card, all the forces to the point of "do whatever it takes".

Another name for scarcity, called expensive. Considering that the demand for arithmetic power increases tenfold for large models, the card will only get more expensive. Recently, an investor said to Geek Park, "A100 single card of 80,000 at the beginning of the year has now been speculated to 160,000, and still can not get. Accordingly, the tens of thousands of cards of the cloud giants to pay the "Nvidia tax" will only be an astronomical figure.

"The most popular Microsoft has the most to say about how it feels to have your life in someone else's hands. A month ago, The information exclusively reported that Microsoft set up a 300-member "team" to accelerate the pace of its own AI chip research, codenamed Cascade server chip may be launched as early as next year.

Not only because of the "neck", cloud vendors to develop their own chips, there is another layer of reference - the GPU is not necessarily the most suitable for running AI chips, self-developed version may be optimized for specific AI tasks.

It's true that most of today's advanced AI models are powered by GPUs because they are better at running machine learning workloads than general-purpose processors. However, GPUs are still seen as general-purpose chips and not really a processing platform native to AI computing. As the Farrer Institute's "A Crack in the Nvidia Empire" points out, GPUs are not built to train neural networks, and the faster AI grows, the more these problems are exposed. Relying on CUDA and various technologies to "magic change" scene by scene is one option, but not the optimal solution.

Amazon, Google and Microsoft have been developing chips called ASICs - application-specific integrated circuits - that are better suited to AI. typically perform these tasks faster and with less power.

As shown above: Amazon, Microsoft and Google have all elevated the importance of silicon to the level of in-house in-house research, developing two types of chips for their data center divisions: standard computing chips and chips specifically designed to train and run machine learning models that can power chatbots like ChatGPT.

Currently, Amazon and Google have developed custom versions of ASICs for key on-premise products and are already making these chips available to customers via the cloud. Microsoft has also been working since 2019 to develop custom ASIC chips to power large language models.

Some of the chips developed by these cloud providers, such as Amazon's Graviton server chip and the AI-specific chips released by Amazon and Google, already match the performance of chips from traditional chipmakers, according to performance data released by cloud customers and Microsoft. Google TPU v4 is 1.2 - 1.7 times faster than Nvidia A100 computation, while consuming 1.3 - 1.9 times less power.

02 Strategic investment race: giants spend money to "buy customers

In addition to the development of chips, the second key point of competition between several overseas cloud giants is the foreign strategic investment to grab AI customers and AI projects.

Compared with venture capital, the giants' war investment has an absolute advantage, and the collaboration between OpenAI and Microsoft serves as an excellent model, opening the precedent of holding hands with big models and war investment. This is because the resource barriers required for big models and related applications are extremely high, and only money, limited money, is simply not enough to grab AI projects. After all, Google, Microsoft, AWS, Oracle or Nvidia can not only write huge checks, but also provide scarce resources such as cloud credits and GPUs.

From this perspective, the grab for projects and customers is happening among the cloud giants, with no other rivals. They are playing a new game - seeking a promise from AI companies that they will use their cloud services instead of their competitors'.

Microsoft sits as OpenAI's exclusive cloud service provider, paying OpenAI's huge cloud bill in exchange for a host of enviable benefits such as equity in OpenAI and first access to its products.

Microsoft's competitors are also scrambling to win over other AI customers. These cloud providers are offering AI companies steep discounts and credits (credits) to win their business. Some critics have pointed out that this is akin to buying customers, although the practice of holding equity in future or current customers is not uncommon in the enterprise software space.

Oracle has also offered hundreds of thousands of dollars worth of computing credits as an incentive for AI startups to rent Oracle's cloud servers, The Information reported earlier.

Google may be the most aggressive of these big cloud vendors, offering AI startups a combination of cash and Google Cloud credits in exchange for equity. Earlier this year, Google invested $400 million in Anthropic, one of the main startup challengers to OpenAI. Google Cloud said in February that it had become Anthropic's 'preferred' cloud provider.

Recently, Google invested $100 million in Runway, an AI company in the "Bunsen burner video" space. But before that, Amazon AWS touted Runway as a key AI startup customer. In March, AWS and Runway announced a long-term strategic partnership to become its "cloud provider of choice". Now Runway appears to be one of Google's "pawns" in its showdown with Amazon, as Runway is also expected to rent cloud servers from Google.

Earlier, Google Cloud also announced partnerships with two other popular AI companies: Midjourney, in the area of Venn diagrams, and Chatbot App Character.ai, which was formerly a key cloud customer of Oracle.

It's too early to tell if these deals will help Google catch up with its bigger cloud competitors - AWS and Microsoft - but Google Cloud is on a roll.

Of the 75 (AI) software companies in The information database, Google provides some cloud services to at least 17 companies, more than any other cloud provider. Amazon is a close second, with at least 15 companies using AWS for cloud computing. Microsoft and Oracle, on the other hand, provide cloud services to six and four companies, respectively. Of course, using multiple clouds is also customary in the industry, with at least 12 of these 75 companies using a mix of cloud providers.

03 Big models, the real key to winning and losing

The early stages of the cloud war are about algorithmic power and war chests. But in the long run, the big model is the real key to winning or losing the market.

Microsoft's partnership with OpenAI has been instrumental in making Microsoft a leader, and the team's excellent engineering capabilities enabled GPT-4 to be embedded in the Microsoft "family" within a few months. In the past six months, Microsoft first used the priority access to OpenAI products and reduced prices for enterprise software products to capture more of the cloud market. Then it relied on the price increase of the product line upgraded to Microsoft 365 Copilot to gain more revenue.

According to the research of CloudQ Capital, Microsoft's underlying model basically relies on OpenAI, and after accessing the big model, Microsoft began to sell application layer products such as Teams, Power BI, Azure, etc. at a lower price package.

Microsoft Chief Financial Officer Amy Hood told investors in April that OpenAI will generate revenue for Azure as more people start using its services.

The latest reports suggest that Microsoft is charging some Office 365 customers an additional 40 percent to test AI features -- which automate tasks such as writing text in Word documents and creating PowerPoint slides -- and that at least 100 customers have paid a fixed fee of up to $100,000. The data shows that less than a month after launch, Microsoft has earned more than $60 million in revenue from Microsoft 365 Copilot's AI capabilities.

In stark contrast to Microsoft, the once-leading Amazon cloud, which is one step behind and one step behind on the big model, is facing an even tougher challenge today.

AWS had been an early developer of AI cloud services, having been in place since about 2016. But customers didn't find these services very useful, including facial recognition, converting text to realistic speech, and chatbots in their raw form for tasks such as customer service.AWS also had launched SagaMaker, an AI digital tool for use by the engineer community that helps them develop and use machine learning models, in 2017, which at one point became AWS' primary AI products.

But AWS' AI products have failed to keep up with the wave of big language models in the years since, and since November 2021, Microsoft has been selling AI products based on the GPT family of models developed for use by enterprise customers. Google, meanwhile, has grabbed major AI startups as cloud customers and is selling proprietary AI software to its cloud customers. Even Oracle, the laggard of cloud computing, has its own advantages when it comes to providing computing resources to AI startups.

As cloud spending on Internet enterprise software tightens, slowing growth is becoming a dark cloud over the heads of cloud vendors.

ChatGPT has come out of nowhere to break this bottleneck, and AI will reinvent software. Software companies, the customers of cloud vendors, are aggressively embedding AI capabilities from big models into existing workflows to accomplish higher-order automation.

With new cloud customers drying up, software companies are no longer going to the cloud for the sake of going to the cloud, but are seeking to improve productivity with AI. "This is the biggest increment in the cloud computing market over the next decade. Algorithmic infrastructure is the absolute dividend beneficiary of the big model." a cloud computing industry insider who has been in the business for more than a decade elaborated to Geek Park.

With such a prospect, several overseas cloud service giants-Microsoft, Amazon, Google, and Oracle-have quickly made changes. In the past few months, the cloud giants have been dropping real money to develop big models, strategic investments, and self-research AI chips ...... The era of big models is on the rise, and they have targeted a new generation of AI software customers.

The former mountains are far from unbreakable, the cloud market is rapidly reshuffling, the giants have opened a whole new curtain of competition.

After all, the fall of the big brother in the mobile Internet era is just around the corner, and Nokia went from 70% of the cell phone market in its heyday to no one's business in a few years, just a matter of making the wrong decision. And for the big model, the cloud industry is quickly forming a consensus: AI is by no means a small variable this time around, and the current leading players could be left behind from the speed of the industry's day-to-day development.

Halfway through 2023, this article will be around several overseas cloud giants to sort out what is the key to competition among cloud vendors today?

01Develop AI-specific chips, can not give "life" all to Nvidia

After the advent of the era of big models, the most scarce resource for cloud providers today is computing power, or AI chips. Investing in the lowest level of infrastructure - AI acceleration chips - has also become the first focus of cloud vendors' competition today.

Scarcity and high cost are seen as the primary reasons for cloud vendors to accelerate their own chip development. Even a powerful man in the tech world like Musk commented that "this thing (Nvidia GPU) is harder to get than drugs," and secretly bought 10,000 cards from Nvidia for his AI company X.AI, as well as a lot of idle equity from Oracle.

Such a degree of scarcity, reflected in the business of the cloud giant, corresponds directly to the loss of business from the "neck". Even Microsoft, which was the first to strike, has been exposed to rumors of a GPU rationing system for its internal AI R&D team due to a GPU shortage, delays in various new programs, and months of queuing for new customers to get on Azure.

Even venture capital institutions to grab the project, are dependent on the hands of Nvidia chip inventory. In order to N-card, all the forces to the point of "do whatever it takes".

Another name for scarcity, called expensive. Considering that the demand for arithmetic power increases tenfold for large models, the card will only get more expensive. Recently, an investor said to Geek Park, "A100 single card of 80,000 at the beginning of the year has now been speculated to 160,000, and still can not get. Accordingly, the tens of thousands of cards of the cloud giants to pay the "Nvidia tax" will only be an astronomical figure.

"The most popular Microsoft has the most to say about how it feels to have your life in someone else's hands. A month ago, The information exclusively reported that Microsoft set up a 300-member "team" to accelerate the pace of its own AI chip research, codenamed Cascade server chip may be launched as early as next year.

Not only because of the "neck", cloud vendors to develop their own chips, there is another layer of reference - the GPU is not necessarily the most suitable for running AI chips, self-developed version may be optimized for specific AI tasks.

It's true that most of today's advanced AI models are powered by GPUs because they are better at running machine learning workloads than general-purpose processors. However, GPUs are still seen as general-purpose chips and not really a processing platform native to AI computing. As the Farrer Institute's "A Crack in the Nvidia Empire" points out, GPUs are not built to train neural networks, and the faster AI grows, the more these problems are exposed. Relying on CUDA and various technologies to "magic change" scene by scene is one option, but not the optimal solution.

Amazon, Google and Microsoft have been developing chips called ASICs - application-specific integrated circuits - that are better suited to AI. typically perform these tasks faster and with less power.

As shown above: Amazon, Microsoft and Google have all elevated the importance of silicon to the level of in-house in-house research, developing two types of chips for their data center divisions: standard computing chips and chips specifically designed to train and run machine learning models that can power chatbots like ChatGPT.

Currently, Amazon and Google have developed custom versions of ASICs for key on-premise products and are already making these chips available to customers via the cloud. Microsoft has also been working since 2019 to develop custom ASIC chips to power large language models.

Some of the chips developed by these cloud providers, such as Amazon's Graviton server chip and the AI-specific chips released by Amazon and Google, already match the performance of chips from traditional chipmakers, according to performance data released by cloud customers and Microsoft. Google TPU v4 is 1.2 - 1.7 times faster than Nvidia A100 computation, while consuming 1.3 - 1.9 times less power.

02 Strategic investment race: giants spend money to "buy customers

In addition to the development of chips, the second key point of competition between several overseas cloud giants is the foreign strategic investment to grab AI customers and AI projects.

Compared with venture capital, the giants' war investment has an absolute advantage, and the collaboration between OpenAI and Microsoft serves as an excellent model, opening the precedent of holding hands with big models and war investment. This is because the resource barriers required for big models and related applications are extremely high, and only money, limited money, is simply not enough to grab AI projects. After all, Google, Microsoft, AWS, Oracle or Nvidia can not only write huge checks, but also provide scarce resources such as cloud credits and GPUs.

From this perspective, the grab for projects and customers is happening among the cloud giants, with no other rivals. They are playing a new game - seeking a promise from AI companies that they will use their cloud services instead of their competitors'.

Microsoft sits as OpenAI's exclusive cloud service provider, paying OpenAI's huge cloud bill in exchange for a host of enviable benefits such as equity in OpenAI and first access to its products.

Microsoft's competitors are also scrambling to win over other AI customers. These cloud providers are offering AI companies steep discounts and credits (credits) to win their business. Some critics have pointed out that this is akin to buying customers, although the practice of holding equity in future or current customers is not uncommon in the enterprise software space.

Oracle has also offered hundreds of thousands of dollars worth of computing credits as an incentive for AI startups to rent Oracle's cloud servers, The Information reported earlier.

Google may be the most aggressive of these big cloud vendors, offering AI startups a combination of cash and Google Cloud credits in exchange for equity. Earlier this year, Google invested $400 million in Anthropic, one of the main startup challengers to OpenAI. Google Cloud said in February that it had become Anthropic's 'preferred' cloud provider.

Recently, Google invested $100 million in Runway, an AI company in the "Bunsen burner video" space. But before that, Amazon AWS touted Runway as a key AI startup customer. In March, AWS and Runway announced a long-term strategic partnership to become its "cloud provider of choice". Now Runway appears to be one of Google's "pawns" in its showdown with Amazon, as Runway is also expected to rent cloud servers from Google.

Earlier, Google Cloud also announced partnerships with two other popular AI companies: Midjourney, in the area of Venn diagrams, and Chatbot App Character.ai, which was formerly a key cloud customer of Oracle.

It's too early to tell if these deals will help Google catch up with its bigger cloud competitors - AWS and Microsoft - but Google Cloud is on a roll.

Of the 75 (AI) software companies in The information database, Google provides some cloud services to at least 17 companies, more than any other cloud provider. Amazon is a close second, with at least 15 companies using AWS for cloud computing. Microsoft and Oracle, on the other hand, provide cloud services to six and four companies, respectively. Of course, using multiple clouds is also customary in the industry, with at least 12 of these 75 companies using a mix of cloud providers.

03 Big models, the real key to winning and losing

The early stages of the cloud war are about algorithmic power and war chests. But in the long run, the big model is the real key to winning or losing the market.

Microsoft's partnership with OpenAI has been instrumental in making Microsoft a leader, and the team's excellent engineering capabilities enabled GPT-4 to be embedded in the Microsoft "family" within a few months. In the past six months, Microsoft first used the priority access to OpenAI products and reduced prices for enterprise software products to capture more of the cloud market. Then it relied on the price increase of the product line upgraded to Microsoft 365 Copilot to gain more revenue.

According to the research of CloudQ Capital, Microsoft's underlying model basically relies on OpenAI, and after accessing the big model, Microsoft began to sell application layer products such as Teams, Power BI, Azure, etc. at a lower price package.

Microsoft Chief Financial Officer Amy Hood told investors in April that OpenAI will generate revenue for Azure as more people start using its services.

The latest reports suggest that Microsoft is charging some Office 365 customers an additional 40 percent to test AI features -- which automate tasks such as writing text in Word documents and creating PowerPoint slides -- and that at least 100 customers have paid a fixed fee of up to $100,000. The data shows that less than a month after launch, Microsoft has earned more than $60 million in revenue from Microsoft 365 Copilot's AI capabilities.

In stark contrast to Microsoft, the once-leading Amazon cloud, which is one step behind and one step behind on the big model, is facing an even tougher challenge today.

AWS had been an early developer of AI cloud services, having been in place since about 2016. But customers didn't find these services very useful, including facial recognition, converting text to realistic speech, and chatbots in their raw form for tasks such as customer service.AWS also had launched SagaMaker, an AI digital tool for use by the engineer community that helps them develop and use machine learning models, in 2017, which at one point became AWS' primary AI products.

But AWS' AI products have failed to keep up with the wave of big language models in the years since, and since November 2021, Microsoft has been selling AI products based on the GPT family of models developed for use by enterprise customers. Google, meanwhile, has grabbed major AI startups as cloud customers and is selling proprietary AI software to its cloud customers. Even Oracle, the laggard of cloud computing, has its own advantages when it comes to providing computing resources to AI startups.