Wall Street doesn't get software

Welcome to Runtime! Today: Why Monday's software-stock selloff and Tuesday's software-stock rally shows how investors don't understand how and why enterprises use software, OpenAI's Project Stargate will be lucky to settle for the moon, and the latest funding rounds in enterprise tech.

Please forward this email to a friend or colleague! If it was forwarded to you, sign up here to get Runtime each week, and if you value independent enterprise tech journalism, click the button below and become a Runtime supporter today.

How long must we sing this song

At Runtime, we tend to leave stock-market analysis to the financial news community given that it tends to have little impact on how enterprise tech buying decisions are made. But the biggest enterprise tech story of 2026 so far has been the massive sell-off in enterprise software stocks and how divorced that reaction is from the foreseeable future.

It was IBM's turn to take a beating on Monday, right after Anthropic freaked out security stocks on Friday with the research preview release of Claude Code Security. This time, all Anthropic had to do was point out on Monday that Claude Code can be used to help modernize apps built around older programming languages like COBOL to send IBM's stock down 13% on the day.

• Never mind that legacy app modernization has been one of the most touted use cases for large-language models in software development for years; Amazon CEO Andy Jassy talked about using these tools to upgrade older Java applications in August 2024.
• Claude Code certainly can accelerate that process with languages like COBOL, as Tribe AI's Tommy Pawelski pointed out a year later in August 2025.
• And while IBM will always be closely associated with COBOL migrations thanks to its mainframe legacy, AWS, Microsoft, and Google Cloud all also make money "digitally transforming" COBOL applications, not to mention the dozens of consulting firms and integration partners that have been working on that problem for years.

Then on Tuesday, all it took to send enterprise software stocks surging was the mere acknowledgment by Anthropic that the world of enterprise software is deeply interconnected. It announced that IT administrators can now link Claude to all the other tools they rely upon through "connectors."

• As Anthropic put it in a blog post, "Connectors bring Claude closer to where your work lives — helping you move faster across tools, not replace the expertise or software you rely on."
• That concept — that modern enterprise tools should be designed to talk to other enterprise tools because there's no single tool that gets everything done — has been one of the biggest animating principles behind enterprise software design over the 15-plus years of the SaaS boom.
• And it also underscores that Anthropic understands that it will take a long time before businesses feel comfortable ditching their existing tools, even if another unhinged AI science-fiction blog post comes along next week.
• "The reality is that these new AI tools will not rip and replace existing software ecosystems and data environments with these AI tools only as useful as the data it can reach," Wedbush Securities analysts wrote Tuesday, according to CNBC.

There is no question that AI tools like Claude Code and OpenAI's Codex have forever changed the way software will be constructed. But it will take years before we fully understand how those developments will change the way massive enterprise companies buy and manage software.

• Even when revealing that Databricks customers are building new apps at a pace its veteran executives have never seen, co-founder Reynold Xin told Runtime earlier this month that those customers are at the cutting edge of software trends and pointed out that it took nearly a decade for cloud computing to become mainstream.
• After years of hype, it's become safe to say that generative AI is an enterprise platform shift, and some companies built during the previous era will not make it in the new one.
• But if you think Silicon Valley is dominated by a herd mentality, check out Lower Manhattan some time.

Fault's in their Stars

When Sam Altman, Larry Ellison, and Masayoshi Son joined President Trump in the Oval Office the day after he returned to the White House last year to announce Project Stargate, it wasn't a stretch to assume an undertaking that massive championed by four of the biggest blowhards of our time would fall short of its goals. More than a year later, that appears to be exactly what happened.

The Information reported Saturday that "the Stargate joint venture has not staffed up and is not developing any of OpenAI’s data centers," which forced OpenAI to work with Oracle and Softbank separately to deploy servers and raise capital. It also signed deals with AWS and Google Cloud to find more computing capacity but as of the end of the year was only able to deploy 7.5 gigawatts of computing power after hoping to get 10 gigawatts up and running in 2025, according to the report.

Couple that the news late last week that OpenAI recently told investors that it only plans to spend $600 billion on computing infrastructure by 2030 compared to earlier commitments of $1.4 trillion, and it's more clear than ever that building massive data centers in a short period of time is very hard. It's not going to get easier.

Enterprise funding

Humand raised $66 million in Series A funding for its HR app, which was designed to connect employees who work in sectors like retail or manufacturing and don't sit behind a desk trading memes on Slack all day.

Nimble landed $47 million in Series B funding as it builds out a platform of AI agents that can take over browsers to search the web for enterprise data.

Cogent Security scored $42 million in Series B funding for its vulnerability management software, which uses agents to detect flaws and take action to plug those holes.

Letter AI raised $40 million in Series B funding and launched a product called Letter Compass, which helps companies create personalized marketing collateral used by sales teams to pitch specific clients.

Venice Security emerged from stealth with $33 million in total funding for its privileged access management technology, which can identify humans and AI agents on networks.

Selector landed $32 million in new funding as it builds a customizable observability service that uses AI to assemble a picture of network activity across operating environments.

The Runtime roundup

Workday added fuel to the "SaaS is doomed" narrative after projecting slightly less revenue than expected during the current quarter, despite beating analyst estimates for revenue and profit during its fourth quarter.

Another worm targeting developers using npm software packages is on the loose, according to CSO, this time using "typosquatting" techniques to make malicious packages appear legit.

Thanks for reading — see you Thursday!