The KonMari approach to enterprise software

Welcome to Runtime! Today: why Asana's CIO thinks IT departments that spent wildly during the pandemic are overdue for some spring cleaning, Nvidia consolidates its central role in the AI boom, and the latest funding rounds for enterprise tech startups.

(Was this email forwarded to you? Sign up here to get Runtime each week.)

Take only what you need to survive

Cloud computing sparked an explosion of entrepreneurship in enterprise software, freeing IT customers from the burden of having to implement big expensive packages of enterprise software and allowing them to pick and choose the best products for a given task built by focused vendors. But 15 years later, the situation has gotten out of hand, according to Asana CIO Saket Srivastava.

"Standardizing on key platforms absolutely makes sense," Srivastava said in a recent interview. "We are taking a harder look at which (tools) make better sense for us, which ones are the ones that are giving us more value, and doubling down on those."

Here are a few excerpts from that interview.

Srivastava on Asana's infrastructure strategy:

I don't necessarily know if we're doing something very different from what (other) companies might be doing, because it's pretty much a playbook in many ways. (But) what might have happened, what has certainly happened in our case through COVID, there have been investments that have been made in the name of technology to improve employee experience in this concept of the future of work and we might have created a fractured digital experience.

So now is really a good time for the CIO function to double down and simplify our stack in many ways, wherein we take a harder look at the technology that we have in our environment and double down on technologies that are driving value for us. Also start thinking in more of an enterprise architecture mindset, where there's a cohesiveness around the experience for the employee, and look at the tools and technologies that might be out there that we're not getting value from and then eliminate those.

On coming to terms with "good enough" tools:

In the journey for Asana, and more broadly for the CIO community that I engage with, there was a shift — even in the large companies — when we've probably gone overboard with best of breed. And over time, there is a realization that in some cases, best of breed makes sense. But that's not a good answer without going deep into the "why."

Standardizing on key platforms absolutely makes sense. From an experience standpoint, from a cost standpoint, there might be better value in standardizing on a few platforms, and then looking at best of breed use cases where platforms are falling short and then plugging those holes.

On pandemic-era tool sprawl:

I liken this to maybe spring cleaning. We've got a ton of sprawl, this is a great opportunity for us to sort of control costs by consolidating and standardizing, but do it in a thoughtful way; not because you have the mandate and you can drive this change. This is the time to establish that trust, and take people along on that journey.

This is again a shift for the CIO function, wherein we need to sort of be sort of in lockstep with our employees a lot more. Even for technology companies, our employees, workers, engineers; they are really smarter than us. You can't necessarily be just telling them "this is the answer" and not listening.

Read the full interview on Runtime here.

Nvidia's bid to own the AI data center

As if 2023 wasn't enough for Nvidia, which is selling powerful and expensive GPUs as fast as it can make them, the company shed more light on its strategy to convince cloud providers and data-center operators to go all-in on its technology.

Nvidia CEO Jensen Huang showed off the next version of Grace Hopper, its combo GPU/CPU chip named after the groundbreaking computer scientist, at SIGGRAPH in Los Angeles today. The new version, which the company expects to release by the middle of next year, will employ a faster version of the HBE memory standard to remove one of the biggest bottlenecks in AI computing.

Nvidia's data center CPU ambitions are certainly notable, given that it already has much of the enterprise infrastructure world eating from the palm of its hand this year. It's hard to see the major cloud providers as thrilled about the prospect of coming to rely even more heavily on the company, but if Nvidia can deliver on performance they might not have a choice.

Enterprise funding

Resilience raised $100 million in Series D funding to help bring AI to the world of cybersecurity insurance.

Endor Labs raised a $70 million Series A round (these things are getting ridiculous again) for its work on open-source security, which has been an ongoing trend in this section for several weeks now.

Simon Data landed $54 million in Series D funding to expand its customer-data platform service, as businesses continue to look for ways to bolster their first-party data sets following mobile privacy changes.

MindsDB raised an undisclosed sum from Nvidia's venture capital arm to complete a $46.5 million seed round and further develop its database designed for AI applications.

ConductorOne scored $12 million in new funding to add to an earlier Series A round and increase the capabilities of its identity-management software.

The Runtime roundup

Zoom got into a dust-up over changes to its terms of service regarding the use of customer data to train its AI work.

Microsoft Azure customers can now get their hands on Nvidia's H100 GPUs, a few weeks after AWS opened up its H100 instances to the general public.

Google unveiled a version of Microsoft's popular Visual Studio Code open-source developer environment that uses its own Codey programming assistant tool.

Stability AI launched its own take on the AI-enabled programming assistant called StableCode.

Twilio's stock jumped in after-hours trading after it released earrings results that beat the Street and raised its yearly profit guidance.

Datadog, on the other hand, also beat analyst estimates for revenue and profit during its last quarter but provided revenue guidance below expectations, and the young Gordon Gekkos of the world don't like that.

Bram Moolenaar, author of the legendary vim text editor used by generations of programmers to write software, died last week after a brief illness.

Thanks for reading — see you Thursday!