Wednesday, November 4, 2020

How does Snowflake, whom Buffett has chosen, profoundly changed the cloud computing industry?

 As we all know, Buffett has never touched technology stocks for a long time. But people always change. He started holding Apple for the first time in 2016 and has repeatedly increased his holdings since then. Apple is currently Berkshire's largest holding stock.

Not long ago, Buffett took a stake in another technology company in his life, Snowflake, and made a new subscription at the IPO stage, and said that he will purchase more than 4 million shares in the later transaction, which may cost more than US$320 million. The company went public on September 17th in the United States, with an increase of 130% on the day of listing, and its market value once exceeded $70 billion.

Buffett himself publicly stated in May last year: "Berkshire has never participated in an IPO in the past 54 years." Even for many Chinese value investors, IPOs have never been an option worth considering.

Who is Snowflake? What is unique about it that makes Buffett worthy of breaking the ring, and twice?

With cloud native and data blessing, Snowflake is growing rapidly

"From the perspective of cloud computing, one of Snowflake's greatest features is cloud native." Huachuang Capital Investment Manager Xie Jia said to Yiou.

Snowflake proposed the concept of data-warehouse-as-a-service (DaaS), which is a SaaS service that is native to the cloud and focuses on data warehouses.

Traditional architecture is often difficult to solve the various data requirements and problems faced by modern enterprises: Shared-disk architecture is easy to cause system crashes under multi-user access, and it is also difficult to meet the needs of high-frequency read and write, data replication and migration.

In recent years, the more mainstream architecture Shared-nothing, although it can guarantee scalability and concurrency, but because node resources do not separate storage and computing, when enterprises need to upgrade and expand data architecture, a large amount of data needs to be migrated to end users The use of it has an impact.

Snowflake completely separates storage and computing in its technical architecture, essentially solving the pain points of previous architectures, and has great advantages in performance, concurrency, and ease of use, maximizing the characteristics of cloud-native architecture.

From the perspective of big data, Snowflake proved the great value of data analysis for the first time.

"The core value of Snowflake lies in data analysis and not in the cloud," said Ma Xiaodong, CEO of National Cloud Data, a domestic cloud data warehouse startup that targets Snowflake, to Yiou.

He believes that because Snowflake itself does not directly provide corresponding cloud services, but provides services based on IaaS vendors such as Amazon AWS and Microsoft Azure. Therefore, in essence, Snowflake is a data analysis company. It is deployed and processed into data products for customers."

"It has sounded the clarion call for the value of data," Xie Jia said. "In the past few years, everyone has been talking about the value of data analysis, but Snowflake has verified this large-scale explosion of demand commercially for the first time. The data management layer in the computing infrastructure can obtain such high growth, high income, and high recognition from customers."

Snowflake's growth data is indeed exciting. According to the prospectus:

In terms of revenue growth, in fiscal year 2020, Snowflake’s operating revenue was US$265 million, a year-on-year increase of 174%, which was 1/3 higher than Zoom’s 108% growth rate over the same period.

Snowflake's number of customers has also grown rapidly, from 1,547 customers on July 31, 2019, to 3,117 customers on January 31, 2020. As of July 31, 2020, the number of customers that contributed more than $1 million in revenue in the past 12 months increased from 22 to 56, indicating that Snowflake's recognition among large customers is high and continues to increase.

NRR (Net Revenue Retention Rate) is an important indicator to measure income health and customer stickiness. Snowflake's NRR is 158%, ranking first in the SaaS field, which means that its existing customers are highly satisfied and loyal, and have a strong ability to pay continuously.

In addition, although Snowflake is not yet profitable, its losses are gradually narrowing: at the end of January 2020, the company's net loss was 350 million U.S. dollars. As of the end of July, the net loss fell to 180 million U.S. dollars, and the gross profit margin also increased from 46.5% in January last year. Rose to 61.6%.

Snowflake is a consumer company

But cloud computing, big data, and high growth, these attributes that many technology companies have, are not the reason Snowflake attracts Buffett. The most important thing is that Snowflake has turned "data insights" into a "fast-moving consumer goods" for the first time.

In an interview with CNBC, Buffett once said that he held a position in Apple because he regarded Apple as a consumer company. “Although Apple has strong technological attributes, it is still a consumer product to a certain extent.”

"The stickiness of Apple products is so good. I took my great-grandchildren to eat DQ ice cream and watched each of them hold an iPhone, and so did their friends. The lives of these young people revolve around Apple. New machines, they will be swarming with new ones."

In fact, in the industry and sector classification, Apple has always been classified as a "consumer product." Dow Jones Industry Classification Standard, Apple are "consumer goods" sector; in the S & P global lines at industry classification standard, Apple is "Consumer Discretionary" "consumer electronics" in the.

After heavy storage of Apple, Buffett's investment logic has not substantially changed, he just chose another Coca-Cola, another oil company.

In the new Snowflake, Buffett’s logic may still be to regard Snowflake as a consumer company.

Data has become the oil of the intelligent age. If land drives the development of the agricultural era, and oil drives the development of the industrial era, then big data will drive the development of the information and intelligent era. In the intelligent era, data becomes the most basic means of production and the infrastructure for intelligent economic activities.

But just as oil can enter all links of industrial production and realize its value after being mined and refined, the value of data can only be reflected after cleaning, sorting, analyzing, insight, and applying to all links of production activities. .

"In the early days of the Internet, acquiring new users was the most important thing for companies. But now that Internet traffic is gradually peaking, the industry generally gradually enters the stock market, and gaining insights from existing user interaction data has become more important. "He Pei, the founding partner of Jingya Capital, who focuses on the field of corporate services, said to Yiou.

Just as Ford Motor has applied the assembly line before cars become mass consumer products, Snowflake's "pay-as-you-go" business model builds data insight into this service and builds it into a "fast-moving consumer product."

Before Snowflake, companies had two ways to analyze data.

One is to build a data processing platform by itself. This requires companies to first purchase computers that can store massive amounts of information, then purchase data analysis software, and hire IT technicians for system operation and maintenance, and data analysts for analysis.

Companies need to pay for hardware, software, and labor costs, but they don’t use them at all times. “It’s like you need to travel and buy a private car, but you only use it for commuting every day. Most of the time the car is lying in the garage. But the maintenance and maintenance costs will not be less at all.” Ma Xiaodong compared.

Another way is to use subscription-based SaaS software, with quarterly or annual one-time billing, "Similar to you renting a car for one year."

Snowflake’s pay-as-you-go model is equivalent to the current Didi Taxi, which is billed according to the amount of data the company needs to analyze and the length of computing time. Enterprise customers do not need to pay for resources and services that are not actually used.

The prospectus shows that 95% of Snowflake's revenue is charged based on the number of times of service, and is not prepaid, but is settled after the fact. Large customers mainly use monthly and quarterly accounts.

Similar to investing in Apple, Buffett may not value Snowflake's technological attributes, but the necessities of production and life in the intelligent age-the consumption attributes of data analysis, but the payer has become an enterprise.

Snowflake barriers to competition

The technical structure and business model constitute Snowflake's core competitiveness, and "neutrality" constitutes Snowflake's competitive barriers.

The cloud-native data warehouse model is not the first of Snowflake. In fact, there are many giants on this track. Competitors include Amazon's AWS Redshift, Google's Google BigQuery and Microsoft's Azure Data Warehouse. According to HG Insights data, Redshift has four times more customers than Snowflake, and Big Query has two times more customers.

But Snowflake has an advantage that giants can never replicate: neutrality.

"It is the only cloud-native data warehouse that can run on the three major public cloud vendors, which caters to the current general trend of multi-cloud." He Pei said.

In recent years, for data security considerations, "multi-cloud deployment" has become the choice of more and more enterprises. According to the "RightScale 2019 Cloud State Survey", on average, overseas companies will run applications in a combination of 3-4 public and private clouds. Overall, 84% of companies adopt a multi-cloud strategy.

In the domestic market where the cloud computing market is relatively immature, multi-cloud deployment is gradually emerging.

12306. Many large enterprises such as China Telecom and China Guangfa Bank have already purchased multi-cloud services. In the government cloud market with a large IT share, government cloud platforms in Beijing, Chongqing and other places have been jointly bid by many cloud computing vendors.

"Under the multi-cloud deployment model, just like the'Switzerland' in the cloud computing industry, third-party independent vendors are more suitable for unified services than IaaS vendors themselves." He Pei said, data security, operation and maintenance, analysis and other services need to be deepened from the outside Within each IaaS vendor, no one will undoubtedly let its competitors deepen.

In terms of business relationships, only third-party independent vendors can provide effective services in a cross-cloud model. It is this neutrality that fundamentally builds Snowflake's unbreakable barriers to competition.

In addition to this level of competition, Snowflake also benefits more cloud computing vendors at the IaaS layer.

As of now, Snowflake's storage layer supports AWS and Azure, and Google Storage will be compatible in the future. In other words, Snowflake is a customer of IaaS vendors, and its business growth can bring massive revenue to IaaS vendors at the same time. For example, Snowflake promises to spend 1.2 billion US dollars on AWS within 5 years.

"Snowflake is a milestone in the cloud computing industry. It makes it easier for enterprises to obtain cloud data warehouse services. This will encourage more enterprises to go to the cloud and use cloud computing resources. The final result is equivalent to making the market cake of the IaaS layer even better. Big," Xie Jia said.

Ma Xiaodong continues to use metaphors to describe the relationship between the two: "Snowflake is equivalent to a central kitchen, and IaaS manufacturers are equivalent to gas. In the past, you had to buy a large amount of gas at one time if you want to cook, but now in the central kitchen, the gas can be used as needed. The frequency and usage are priced, so while the central kitchen attracts more people to cook, it also increases gas usage."

Taking Snowflake as the boundary, the "data value" may no longer be expected, but closer and closer to reality.

No comments:

Post a Comment

Introducing Your Go-To Source for the Latest in AI Technology News

Introducing Your Go-To Source for the Latest in AI Technology News AI technology is rapidly advancing and changing the way we liv...