05.30.2025

ForwardThinking 🤔

Data management is having a moment

While this week has been a bit quieter than last when it comes to technology news, we did have one bombshell drop on Tuesday when Salesforce announced its intent to acquire Informatica for $8 billion. It’s a big chunk of cash for a company born in the 1990s, but Salesforce saw a chance to get an enterprise-scale data management tool in the fold and it didn’t flinch. It’s only cash after all.

Data has taken a central role for the enterprise in the age of generative AI. As we’ve discussed often in this space, good data is key to the success of AI projects and garbage in/garbage out, the notion that if you build applications based on bad data, you will have bad outcomes, still applies.

What’s more, companies like IBM have been talking about smaller, more focused models lately that take advantage of company-specific data. As IBM CEO Arvind Krishna put it at IBM Think earlier this month, “To win, you are going to need to build special purpose models, that are much smaller, tailored for a particular use case, and that can ingest the enterprise data and then work.”

Screens hanging from ceiling with various types of data.

Photo by Leif Christoph Gottwald on Unsplash

Perhaps it’s not a coincidence that IBM closed a deal to acquire DataStax this week, a data management company built on the open source Apache Cassandra project. As DataStax CEO Chet Kapoor put it on LinkedIn announcing the deal had closed, “Our mission remains the same -- empowering enterprises and developers to build and scale mission-critical applications.”

A pattern emerges

It’s part of a pattern we are seeing where large companies like IBM and Salesforce are using their wallets to scoop up data management tooling at a time it’s becoming increasingly important to find ways to access that mission-critical data Kapoor referred to, data that’s often trapped in legacy systems. The idea is to take this data in older systems and put it in modern formats that large language models can consume and understand.

Thomas Squeo, chief technology officer at Thoughtworks, a technology consultancy, says his company has been working to help clients unlock legacy data, which he believes is going to be essential for many organizations moving forward. “AI has made data lock-in an existential threat for some enterprises because they're going to be out-competed by new players that are not encumbered by that,” he said. He’s talking about newer companies that were built from the ground up on top of a more modern data stack.

“AI has made data lock-in an existential threat for some enterprises because they're going to be out-competed by new players that are not encumbered by that."
~Thomas Squeo, Thoughtworks CTO

But those organizations dealing with legacy data problems need to think more deeply about, and put more resources into getting their data in order. Perhaps companies like IBM and Salesforce see this as a way to help companies sitting on large legacy data stores to position themselves better for an AI future.

What if they're right?

Jack Gold, principal at J. Gold Associates, certainly sees it that way. Writing about the Salesforce-Informatica deal on LinkedIn, he said, “Even if you believe that legacy data management isn't the future, it will be in place for a very long time, and it's a critical stepping stone to AI.”

Photo by Nana Smirnova on Unsplash

If he’s right, Salesforce’s $8 billion bet might pay off handsomely by making it easier for customers to access all that data locked in those older systems. What’s more, pairing Informatica with MuleSoft, a tool built to connect to legacy systems via APIs, fits neatly into Salesforce’s broader agentic vision where Informatica handles the data management, while MuleSoft helps build the connectors the agents rely on to move between systems.

Regardless, it feels like we’ve reached a point where large companies are recognizing that it’s going to be critical for them to help customers get at data wherever it lives, and we could see data management tools increasingly in demand as M&A targets as a result. Perhaps Salesforce and IBM are just ahead of the game.

~Ron

What's new on the blog 📰

Klaviyo’s CTO keeps company firmly focused on the data

Klaviyo CTO Surabhi Gupta's background includes stints as director of engineering at Airbnb and SVP of engineering at Robinhood. We talked about how data has been a constant across her entire career, and how Klaviyo's focus on managing customer data was a a big draw for her.

"With Klaviyo, I felt that everything started with data, and the fact that you're storing all of this data for customers means that there's an infinite set of possibilities of what you can do. So in terms of the problem space and potential, I felt it was ripe for us to be doing a lot more."

Read the full article >>

Salesforce opts for legacy data management with $8 billion Informatica acquisition

As we noted in this week's ForwardThinking commentary, Salesforce is back to its big deal-acquisitive ways. This time it's legacy data management company Informatica, and the price tag is a hefty $8 billion. I can see their thinking: AI feeds off data and rather than relying on a third-party data management tool, they bought one of their very own. It's only money after all, but is a legacy data management the answer for modern AI? I'm not so sure. Time will surely tell.

Read the full article >>

Photo by Ron Miller

Walter Sun is helping SAP build AI that understands tabular data, not just language

Walter Sun joined SAP in 2023 after almost two decades at Microsoft to head up AI at the company. It was a pivotal time, as SAP began its generative AI shift. He's helping address a number of challenges including finding a way for AI models to understand tabular data, a format in which you'll find a lot of crucial business data, especially SAP's.

“We're building an SAP foundation model, which is different from large language models,” he said. “It's going to supplement LLMs and allow SAP customers to access and leverage business data. This business data, which we consider mostly tabular data, is more difficult to process,” he said.

Read the full article >>

With Neon, Databricks added a crucial database piece to the data lakehouse puzzle

When Databricks announced it was acquiring Neon earlier this month, it filled in a crucial missing piece for the company and set it up for more dynamic agentic use cases in the future. I spoke to Sanjeev Mohan and Devin Pratt to get their perspectives on the deal.

Read the full article >>

Photo courtesy of Databricks

News of the Week 📣

Zscaler acquires Red Canary for $675 Million to boost threat Detection capabilities

While Salesforce stole the headlines this week with its hefty $8 billion acquisition of Informatica, it wasn't the only company adding capabilities through M&A. Zscaler, the security operations company, announced it was buying Red Canary, a managed threat detection tool.

CRN reports that the deal was worth $675 million in cash and equity.

Zscaler CEO Jay Chaudhry sees the combination of the two companies making a more powerful security operations platform. "The proposed acquisition of Red Canary is a natural expansion of our capabilities into managed detection and response and threat intelligence to accelerate our vision of AI-powered SOC of the future," he said in a statement.

Photo by Peter Conrad on Unsplash

For Red Canary, it's a matter of getting access to more resources than it could on its own, a familiar refrain from an acquired company. CEO and co-founder Brian Beyer talked about the scale of Zscaler's Zero Trust Exchange, which processes 500 billion transactions a day. He believes all that data will help make them better and allow them to expand into other areas. "This will significantly enhance our ability to detect threats faster and more accurately. The innovation this will bring is going to be incredible," Beyer wrote in a blog post announcing the deal.

Whether the combined organizations will really enhance each other and increase innovation remains to be seen, but optimism reigns for the time being.

The deal is expected to close by August, assuming it meets all of the closing conditions.

Oracle will reportedly buy $40 billion worth of Nvidia chips to fuel OpenAI stargate data center

When OpenAI announced the $500 billion Stargate project in January, it was met with some skepticism in the industry in terms of where all the money would come from. Last week, The Financial Times reported that Oracle plans to spend a whopping $40 billion on advanced Nvidia chip technology for a single data center currently under construction in Abilene, Texas.

To reach that kind of figure, the Times reported that it would require 400,000 (yes, 400,000) of Nvidia’s GB200 super chips, which reportedly cost between $60,000 and $70,000 apiece.

Photo courtesy of Nvidia

It's a hard number to believe, and the reports indicated that neither company has actually confirmed the purchase yet. It's even harder to believe when you consider that Oracle has just $17 billion in cash on hand. While that's still a lot of money, it's not nearly enough to pay for an investment of that magnitude without a massive round of financing, and putting other internal projects at risk.

Even if it is accurate, The Register reports that it's highly unlikely Oracle could ever take full advantage of all of these chips because of the resource requirements around water, cooling and electricity to run them all at the same time.

If it's true, however, it will create a massive computing center for OpenAI to operate with as it looks to expand its capabilities and its market reach.

Grammarly scores $1 billion investment from General Catalyst

Grammarly has been around for 14 years now, helping writers check their grammar, spelling and punctuation as they write. Over time, of course, they have moved into AI to help with those tasks. This week, the company announced an unusual $1 billion funding deal from General Catalyst.

The investment is designed to help Grammarly expand their sales and marketing efforts and to make strategic acquisitions without diluting the value of the company.

Grammarly acquired Coda, an AI-driven collaborative workspace tool earlier this year, and has quickly moved to incorporate it into the broader Grammarly platform. The company hopes to expand into other areas with this investment. "Since integrating with Coda earlier this year, we've evolved into an AI productivity platform serving millions daily. This investment will help us accelerate our growth," CEO Shishir Mehrotra wrote on LinkedIn.

TechCrunch's Marina Temkin reports that GC isn't buying an equity stake in the company. Instead, it's a kind of loan. "Grammarly will repay the capital along with a fixed, capped percentage of revenue it generates from the use of General Catalyst’s funds," Temkin wrote.

Adventures in AI 🤖

Photo by Andy Kelly on Unsplash

I often drop a draft of my articles into one of the large language models and ask for feedback like how well the piece flows, whether there are spelling, grammar, or punctuation mistakes, or if I overused a particular word, which is surprisingly easy to do without noticing in your own writing.

I try to rotate between different models like Claude, ChatGPT and Perplexity, but one thing I've noticed is that regardless of which model I use, they're all obsessed with the em dash. They constantly suggest replacing perfectly good commas with em dashes when frankly, the commas work just fine.

I think I'm going to add a "no em dash suggestion" note to my prompt next time and see if that reduces this particular recommendation.

How are you using AI in your job? I would love to hear from you at ron@boldstart.vc.

What we're reading 📚

Person reading book, but all you can see is their hands holding the book. They are sitting on a hardwood floor steeped in shadow, creating an interesting look.

LivPhoto by Blaz Photo on Unsplash

The Register

Forgotten Turing treasure trove rescued from attic

Richard Currie

TechCrunch

Khosla Ventures among VCs experimenting with AI-infused roll-ups of mature companies

Marina Temkin

Six Colors

Sam and Jony and skepticism

Jason Snell

CNBC

3 key takeaways from Nvidia’s earnings: China blow, cloud strength and AI future

Samantha Subin and Kif Leswing

Look who's talking 👄

"I’ve come to believe the idea of a “long term moat” is mostly a myth, especially in this market. Moats aren’t permanent. They’re time-bound. At best, they function as a bridge. And companies either use that bridge to reach the next defensible position, or watch their moat get breached."

- Jamin Ball in this week' s Clouded Judgement Newsletter

Please subscribe to the newsletter and feel free to share.

3350 Virginia St, Suite 262, Miami, FL 33133

ForwardThinking 🤔

Data management is having a moment

A pattern emerges

What if they're right?

What's new on the blog 📰

Klaviyo’s CTO keeps company firmly focused on the data

Salesforce opts for legacy data management with $8 billion Informatica acquisition

Walter Sun is helping SAP build AI that understands tabular data, not just language

With Neon, Databricks added a crucial database piece to the data lakehouse puzzle

News of the Week 📣

Zscaler acquires Red Canary for $675 Million to boost threat Detection capabilities

Oracle will reportedly buy $40 billion worth of Nvidia chips to fuel OpenAI stargate data center

Grammarly scores $1 billion investment from General Catalyst

Adventures in AI 🤖

What we're reading 📚

The Register

Forgotten Turing treasure trove rescued from attic

TechCrunch

Khosla Ventures among VCs experimenting with AI-infused roll-ups of mature companies

Six Colors

Sam and Jony and skepticism

CNBC

3 key takeaways from Nvidia’s earnings: China blow, cloud strength and AI future

Look who's talking 👄

Subscribe

FastForward #24: Data management is having a moment

You may also like...