Cloudera and Hortonworks were bleeding cash as the two data software providers spent years going head-to-head to lure businesses onto their fledgling technology. As of Thursday, they can join forces against a common enemy: Amazon.
The all-stock deal unites the two most prominent vendors of Hadoop open-source software, which customers can use to store, process and analyze many different types of data. Valued at $5.2 billion when the merger was announced in October, the companies are worth just a combined $3 billion as of its official close.
The stocks plunged in the fourth quarter amid the market sell-off and specific concerns about whether Cloudera — the name of the combined entity — has a compelling enough story to take on Amazon Web Services. For all of 2018, a year that was a boon for many cloud stocks, Cloudera shares fell 33 percent and Hortonworks dropped 29 percent.
"The big competitor for us is Amazon, with their house offerings," Cloudera CEO Tom Reilly told CNBC in an interview last month, in preparation of the deal's closure. "We're gearing up to take on Amazon."
The Cloudera-Hortonworks merger fit into a big technology theme in 2018 — the growth of open source. IBM bought Red Hat for $34 billion, Microsoft acquired GitHub for $7.5 billion and Salesforce purchased MuleSoft for $6.5 billion. Elastic, whose software helps companies embed search functions in their apps, had one of the bigger IPOs of the year and is now valued at about $5 billion.
They all are keeping a close eye on AWS. The leader in cloud infrastructure has successfully pulled companies' core workloads into its data centers and is now adding new features, functions and services that are easy to use once a business is already on Amazon's machines.
Cloudera is particularly at risk because Amazon has been investing heavily in databases. AWS' competitive products include Elastic MapReduce, for storing and processing different data types, and Redshift, a data warehouse system. Rishi Jaluria, an analyst at at D.A. Davidson, estimates that EMR generated about $250 million in revenue in 2018, while Redshift reached about $400 million, a small fraction of the $25.5 billion AWS is expected to report for the year, according to analysts polled by FactSet.
At the time of the Cloudera-Hortonworks deal, the companies said they would have combined annual revenue of $720 million and that they expected 2020 revenue to exceed $1 billion.
The beefed-up Cloudera has certain strengths as it stares down Amazon. Reilly said Cloudera is the standard-bearer for companies that want to use Hadoop because it's independent and has a clear open-source model.
Cloudera also works in corporate data centers and across multiple clouds. For years AWS focused exclusively on delivering software to companies from its own remote data centers, and it only recently announced servers that will be able to run Amazon software inside of other companies' existing data centers. Unlike Cloudera, AWS doesn't have a way to let companies use many of its software products on clouds operated by Microsoft or Google.
"Amazon will never be able to deliver multi-cloud," said Reilly, who has run Cloudera since 2013 and took it public in 2017.
An Amazon spokesperson declined to comment.
Deploying Cloudera's software in the cloud is becoming more commonplace. As of June, 26 percent of customers that provide diagnostic information to the company are using the software on a public cloud like Amazon, up from 22 percent one year earlier.
A big advantage for the new Cloudera is it no longer has to spend money taking on Hortonworks, and vice-versa. Cloudera lost $111 million in its previous three quarters, and Hortonworks reported a $114.8 million loss over that stretch. Together, they spent more than 55 percent of revenue on sales and marketing.
Reilly said that by becoming more efficient in sales as well as research and development, within a year Cloudera should have a 15 percent operating cash flow margin, a key measure of profitability.
"Neither of us are at that point right now," he said.
In addition to Amazon, Cloudera has to contend with emerging companies like Databricks and Snowflake.
"The combined entity is better positioned to compete against those players," D.A. Davidson's Jaluria, who has a buy rating on Cloudera, wrote in an email.
While the merger makes smart business sense, there are challenges in pulling together staffs from two companies that were bitter rivals. Reilly said it was hard to build a leadership team with executives from the two sides.
"You have the loyalties," he said. "You have the people you know."
The unification with Hortonworks involved putting together 11 "functional integration teams" with employees from both businesses. The work is wrapping up ahead of schedule, Reilly said.
One thing Cloudera hasn't ruled out is imposing software-licensing changes that could help protect it from larger cloud providers like Amazon. MongoDB, Redis Labs and Confluent all took steps last year to limit what others could do with their software.
Heather Meeker, a lawyer with software licensing experience and a portfolio partner at OSS Capital, said she expects more companies involved with open-source projects to introduce constraints.
"At the end of the day, I think all of these companies are trying to make sure they don't get starved for resources when there are a bunch of big users benefiting from what they're doing, but they're just going down different routes," Meeker said.
Reilly said Cloudera doesn't have plans to tweak the licenses for any of its 26 open-source projects, including the core Apache Hadoop technology. But that could change.
"There are so many different views on this," Reilly said. "We believe that open source is the way to drive innovation and drive standards. We believe some of the cloud players are good about participating in the open-source community. Others are not."
Amazon's track record in that area is poor, Reilly said.