Today: As is tradition, AWS released all the news that won't make the re:Invent keynote ahead of time, the Allen Institute for AI introduces a powerful and truly open-source AI model, and the quote of the week.
This era of enterprise software is either the dawn of a new era of corporate productivity or the most hyped money pit since the metaverse. ServiceNow's Amit Zavery talks about the impact of generative AI, how SaaS companies should think about AI models, and his decision to leave Google Cloud.
Today: Zoom is increasing the use of its own infrastructure after relying on the cloud during the pandemic, Databricks says it is growing faster than its larger rival, and this week's enterprise moves.
Welcome to Runtime! Today: Zoom is increasing the use of its own infrastructure after relying on the cloud during the pandemic, Databricks says it is growing faster than its larger rival, and this week's enterprise moves.
(Was this email forwarded to you?Sign up here to get Runtime each week.)
Repatriation games
A clear pattern is emerging among enterprise tech companies that are rethinking their use of cloud infrastructure services; they have a deep understanding of the behavior of their workloads, and they are growing at slow, predictable rates. After scrambling for cloud servers during the early days of the pandemic, Zoom plans to rely more heavily on its own infrastructure going forward.
On Wednesday DCD spotted a paragraph in Zoom's 2023 ESG report that discussed how the company relies on a mix of self-managed data centers and cloud providers — AWS and Oracle — to serve those hour-long video calls during which half the attendees are paying attention. In the future, however, Zoom wants to rely more heavily on its own infrastructure for production customers.
"Our data center strategy is to migrate more into the leased colocation data centers, decreasing our dependence on cloud services to support our customers," Zoom said, noting that it will still have some level of presence on cloud servers as needed.
At the end of 2023, Zoom maintained 29 data centers across the world.
Zoom doesn't construct its own data centers; instead, it rents space from "top-tier global data center providers" like Equnix and Digital Reality, with Equinix hosting around half of that co-located infrastructure.
It is also in the process of launching "NextGen" data centers, five of which have arrived to date to replace older infrastructure.
Cloud providers crowed about their integral role supporting Zoom as the world locked down while trying and failing to prevent the spread of Covid-19 in 2020, but the reality was a little different.
Brendan Ittleson, who was CTO at Zoom during those frantic days, told me in late 2020 that Zoom had 19 co-located data centers running at 50% capacity serving 10 million customers a day in December 2019, heading into the pandemic. (Please donate to the Internet Archive.)
By March 2020 Zoom was serving 200 million customers a day, and while Ittelson said Zoom's application was designed to scale quickly on its own hardware, it still needed some help and turned to previous partner AWS and new partner Oracle.
But Zoom depended on its own data centers to host most production traffic, and it still does: Zoom only turns to cloud providers when it needs "a specific capacity," it said in the ESG report, calling AWS its "preferred provider."
That makes Oracle co-founder Larry Ellison's 2020 claim that "the majority of Zoom meetings — these tens of millions of Zoom meetings that happen every day — happen in the Oracle cloud" look even more ridiculous in retrospect.
Zoom beat Wall Street expectations for revenue growth last week, but it's only growing at 3.1% these days and operates at margins far lower than the cloud providers.
Dropbox wrote the playbook for Zoom's approach back in 2015, when it realized that it could save millions by standing up its own servers thanks to a deep understanding of the needs of its application and a relatively predictable growth pattern.
More recently, venture capitalists tired of paying the AWS Tax have tried to convince startups under their influence to adopt their own infrastructure, otherwise known as "cloud repatriation."
Zoom's story is a little different, given that it has always relied heavily on its own infrastructure.
But as DCD noted, this week Zoom CFO Kelly Steckelberg might have laid out the new infrastructure strategy for established enterprise tech companies: "We have cloud providers that we work with, especially when you launch something new to make sure there's enough capacity, but then as quickly as possible getting that back into our data centers."
A MESSAGE FROM CANVA
As enterprises rush to embrace AI, CIOs are grappling with how to merge their enterprise’s IT past and future. Increasingly, in-house technology leaders are tasked with the seemingly impossible mandate of reaping the benefits of next-generation systems while simultaneously reducing legacy technical debt and costs and managing risk. Read more aboutThe CIO Paradox on Runtime.
Waiting on IPO
Databricks doesn't need much more to prove it will be an enterprise tech company for the long haul, given its pivotal role helping companies and researchers manage the explosion of application data even before the generative AI boom kicked in. But after years of teasing a public offering, CEO Ali Ghodsi still isn't quite ready to pull the trigger.
According to the Wall Street Journal, Databricks recorded $1.6 billion in revenue during its last fiscal year, a 50% increase compared to the previous year. That's a faster rate of growth than the 38% clip Snowflake posted last year, but its already public rival pulled in $2.7 billion during its last fiscal year.
Asked about its IPO plans, Ghodsi told the WSJ that "we’re certainly ready as a company: The way we’re operating, the way we’re doing our audits, the way our financials are, the CFO, the board structure." But enterprise tech companies on the verge of IPOs keep eyeing each other from a distance, waiting for somebody else to go first and break the logjam.
Salesforce's low-code Einstein Studio platform is now known as Einstein 1 Studio, and now comes with a generative AI feature that can create code from prompt engineering.
Tom Krazit has covered the technology industry for over 20 years, focused on enterprise technology during the rise of cloud computing over the last ten years at Gigaom, Structure and Protocol.
Today: As is tradition, AWS released all the news that won't make the re:Invent keynote ahead of time, the Allen Institute for AI introduces a powerful and truly open-source AI model, and the quote of the week.
Today: Microsoft shores up its AI strategy heading into a pivotal year, Meta is getting into the AI SaaS business with the former leader of Salesforce's AI division, and the latest enterprise funding.
Today: OpenAI would rather ChatGPT users spend more time using its tool than other "copilots," HPE rolls out a new supercomputer design, and the quote of the week.