OpenAI Acquires io for $6.4 Billion || Tech Field Day News Rundown: May 28, 2025

At t acquires CenturyLink Vast Data reveals an AI operating system. Salesforce buys Informatica for $8 billion to get more ai. Red Hat is pushing Linux for smart vehicles, and Google has a new approach to AI infrastructure.

4 billion? Get all of the news on the tech field day rundown. Welcome to the Tech Field Day rundown, where each time we meet, we run down the IT news with the week with a varying degree of sarcasm.

I'm your host Alistair Cook, and joining me is my co-host Cory. Rodney, welcome to the show. Thanks for having me, Alistair.

It's a joy to have you here with us on National Hamburger Day, and I hope everybody is going to have a nice feed of hamburgers today. It's also National Flip Flop Day, or as we would call it here in New Zealand National Jan Day. Um, but it's funny that World Hunger Day is also national Brisket day.

We have the complete set of interesting flavors to find. Yeah, I I guess if, uh, if you don't like brisket or hamburgers, it, it's set, it's definitely world hunger day for you, uh, while everyone else is downing brisket and hamburgers. Um, all jokes aside, I think my favorite one is the National Flip Flop Day.

I think that's a good excuse to go shopping after this. Yeah, it's a bit rainy outside at the moment here in New Zealand, so, uh, it may not be the best of Jan days for me. 75 billion in cash.

1 million fiber customers across 11 states and access to over 4 million fiber enabled locations, significantly expanding AT&T's fiber footprint in major metro areas like Denver, Seattle, and Orlando. This acquisition, a part of AT&T's strategy to double its fiber network to reach approximately 60 million locations by 2030. Corey, do you think they're gonna make it?

Yeah, I, I think they're gonna make it, I think this is, you know, a clear acceleration of that strategy of AT&T's, um, to double their fiber network. You know, they can do it in one of two ways. They can build it or they can buy it.

In this case, maybe they'll do a little bit of both. But right now, buying it, um, at and t historically has been a Titan, um, and even has had a few side quests with Media and Warner Media, which they've now divested in. But this seems like a very significant res sharpening of focus, uh, to become the number one broadband buyer provider in America.

1 million customers and 4 million locations, it's definitely gonna bolster their progress to expanding high speed fiber across the nation. Um, like I said, builder to buy it. In this case, they're, they're buying a good, a good portion of it.

So for enterprises, what does this mean? Um, more ubiquitous fiber options into key metro areas, like you mentioned, Denver and Orlando. Uh, simplifying support for, for those distributed workforces and solid last mile infrastructure.

On the other side, lumen, uh, many of us still know them as CenturyLink or Quest, um, and even level three before that, uh, seems that they are making a very deliberate, uh, strategic pivot here. Um, they've got a, a very deep history in the enterprise and wholesale networking, particularly with its fast fiber backbone and services geared towards large businesses and government. So by shedding that capital intensive consumer fiber business, you know, I think Lumen ISS going to be able to materially reduce debt and free up some capital to accelerate investments into high growth enterprise solutions.

Um, their CEO explicitly stated that they're sharpening their focus on the enterprise, especially for multi-cloud and everybody's favorite buzzword, AI first world. Um, so I think that means we can expect lum to double down offerings, um, like private connectivity to fabric cloud on-ramp, stuff like that. So to me, this is, we're seeing a clear delineation at t solidifying its its position as a broadband provider and aiming for sheer scale and widespread connectivity across the nation.

And Lumen refining itself into more specialized agile enterprise network powerhouse, uh, specifically for this next wave of, of AI infrastructure. So, goodbye by at and t, uh, and if they wanna throw some money my way, you send 'em on over to me. Next, vast data has introduced its AI operating system designed to unify and simplify global scale AI infrastructure by transitioning from traditional application-based systems to agent driven architectures.

The platform incur encompasses a kernel for cloud services, a runtime for deploying AI agents, real-time event processing, messaging infrastructure, and distributed storage for analytics. Additionally, vast unveiled agent engine, a low-code environment for building intelligent workflows, which includes a prebuilt open source agents tailored for various tasks such as data engineering, compliance, and life sciences research Seems an interesting move to lash a, a big pile of AI on top of your storage, but really, what is this progression? So vice data we saw starting out with this, uh, huge scale out all flash array where they said, there's no need to have any tears, just put everything in here and the the array will look after all.

And then they expanded outwards to being universal storage. So all kinds of data, not just block storage that's being presented out, uh, through file storage objects, and then moving onwards to being a data platform for ai. That was a couple of years ago, they decided that, uh, AI was gonna be good for them and that all of your AI data types could land on your vast storage platform.

Now, your AI applications themselves can land on the same operating system that's distributed ag agentic computing and AI analytics platform or operating system that they're offering. It's interesting that it's a, a company that we think of as entirely as a storage vendor who is now talking about an operating system for AI and delivering these AI applications easily with low, low-code, no-code kind of deployment. And there's some real perspectives around the move to age to a shift away from work being done by humans who are using tens to hundreds of applications and, and an, uh, enterprise to having potentially millions of little, tiny agents doing a large amount of the work, offloading the routine, predictable, repeatable things that we really would like AI to be doing.

And that's what the agent engine is all about. It's the ability to build very simply these agents that will take some task and autonomously complete that task. Going beyond simple automation of, I, I push a button and it does a thing.

Uh, actually having this event driven structure where something occurs in the world, maybe an order comes in, maybe a memo arrives from, uh, my boss and an agent then handles that. Maybe the agent reads the, the memo from my manager or his manager and, uh, identifies whether it should send me a text to tell me I should really read this right now, get up off, off my, uh, uh, off my chair next to the barbecue, uh, or whether I can wait until tomorrow. Maybe that's what the agents doing.

This shift towards lots and lots of very small agents doing a lot of work for us is definitely one of the things we're seeing in the AI industry. And vast as well placed to have all of the inputs and outputs actually residing on their platform. And then much more use of these agents as, as this platform, as this operating system for AI agents, uh, is built out.

I think the agent engine is gonna be interesting to see what it actually ends up being. What more infrastructure beyond your vast array you need in order to use this, particularly as it scales because your storage array, your vast array is using quite a lot of CPU power to do all of the smart things with storage. We're probably gonna need some more CPU power to run vast numbers of agents.

So it'll be interesting to see as this moves onwards as we see more personal assistance and more prebuilt ais, um, possibly even where we're using vibe coding to create new AI agents that having an AI write our AI agent for us. Uh, yeah, certainly a, a vast future with, uh, agents. Salesforce has reportedly close to an 8 billion deal to acquire the data management firm.

Informatica, according to the Wall Street Journal report, the potential acquisition could mark a significant investment by Salesforce into strengthening its data capabilities, particularly as it continues to integrate and expand its AI offerings. Buying Informatica as the biggest deal since the, it's nearly 28 billion, uh, dollar acquisition of Slack technologies in 2021. And what helps Salesforce expand its data management tools is it doubles down on AI powered products.

The deal would also allow Salesforce to tighten the control over how business data is managed and used an essential step as it races to embed generative AI deeper into its products. And this seems to be a general trend. I think Salesforce has done a couple of AI acquisitions recently.

Corey? Yeah, I mean, who isn't buying into AI right now? Uh, surely Salesforce is, is, uh, putting their money where the trends are and, and where the business is.

So for me, I, I think this is a really smart play by Salesforce, especially when we focus on the data management piece. For years, we've preached garbage and garbage out when it comes to enterprise data management. Um, but with generative AI and larger models, that principle is now staring us straight in the face and simply cannot be ignored.

Um, it can't be just a saying anymore. If an organization has poor data management practices, they're going to have a very steep uphill climb to leverage and innovative AI tools. Well at least leverage them.

Well, they may work, but they may not be, uh, giving you great answers. So you can have the most advanced ai, but if it's fed inconsistent, incomplete, or dirty data, its intelligence is just an illusion. Um, and it's gonna lead to flood insights, bad decisions, and poor user confidence, and an in, in turn, poor adoption.

So Informatica has been a longstanding leader in data management, cleansing, integrating, and governing data across complex enterprise landscapes. So for Salesforce, especially, while they're aggressively pushing the Einstein AI and Data Cloud, this isn't just about more data, it's about injecting trust into their customer data fabric. So their CRM may be a hub, but customer data lives everywhere.

Um, so unifying and purifying that data is going to be paramount to having quality AI solutions and having user confidence in the those answers and, um, and offerings that their AI solutions put out. So this is a massive investment that signals that Salesforce profoundly understands the critical need for robust data foundations. It's a strategic move to ensure their AI delivers genuine, actionable intelligence and not just automated misinformation.

Uh, this deal sets a new benchmark to truly to succeed with ai, you must first conquer your data quality and anything else is just very expensive garbage out. Red Hat believes Linux will play a key role in powering the next generation of software defined vehicles by offering open source platforms. Red Hat aims to help automakers build more flexible updateable and secure in-car systems.

The company is working with industry partners to create standardized cloud connection vehicle architecture. Alistair, what do you think? I think this is a sign of maturity.

One of the things that we see as new technologies get deployed out, they get deployed as a point solution and are very specialized. And then over time some things get standardized. Underneath the gating factor in here is a thing called the safety element outta context framework, which is part of the automotive safety integrity level B standard.

This is the certification that says you can put the system inside the the car. Now, historically, a car maker, a supplier to the car maker would make a custom solution for a specific use for a specific application and get that certified within the framework. What Red Hat's done now is released a new distribution of Linux called the Red Hat in-Vehicle Operating System.

And this is certified to the S-E-O-O-C or suit uh, framework. This allows the systems builders who are putting together a solution to go inside a vehicle to just trust that this lower layer layer is certified and only have to build the certification up above that. I think it's an awesome thing to enable much simpler creation of new and updated versions of application platforms that run these smart cars.

Uh, both electric vehicles and autonomous vehicles, both manually operated and autonomous electric vehicles require a whole lot of instrumentation. And we know that increasingly the systems inside our our cars are becoming more complex and more interrelated. And related to this is the vulnerabilities that we've seen in remote management, remote controller vehicles, and providing a high level of security at, at the lowest level of the operating system is going to be, uh, a great assistance to building more security further up.

We've seen remote start, remote stop of unknown vehicles over the internet through the, the remote management interfaces. Uh, getting better security around this is gonna be a good thing. And having more of a platform that includes that security includes that certainty, um, early on is, is gonna be very beneficial to the car makers and the supplies to those car makers like algae, electronics, and Texas instruments and auto wear.

Uh, there's a, a whole collection of people who will benefit from having a, a good solid foundation to build upon. Google Cloud is focusing more on building better AI technology by combining special hardware, fast networks, and smart storage into one powerful system. This AI hyper computer is designed to make developing and running AI easier and faster.

It uses important parts from Google's tus and video GPUs and the Unified storage system to help data, uh, manage data more simply. Uh, we saw this technology at AI infrastructure field day two, and Google is also growing its data centers and plans to spend $75 billion by 2020 to five to support the rising need for AI and cloud services. Once again, lots of money being spent on AI Co.

They're gonna get a return on this. What is this stuff? Yeah, uh, I like the tagline and infrastructure gets Google year, or should we say storage, but sexier.

Um, to me this is bringing, like you said, bringing the network, the GPUs, the CPUs, TPUs, how many use, uh, storage and everything all together to build a purpose-built solution. So automating and adding intelligence to those storage solutions, uh, this seems like a major strategic move and a good one amidst the AI arms race between major cloud providers. Uh, I won't name them, but you know who they are.

They're all trying to differentiate themselves right now. So from my standpoint, this isn't just about offering more compute with the $75 billion investment. It's a testament to Google's commitment to deeply integrated purpose-built systems designed to tackle the complexities of modern AI because, you know, everybody wants to do ai, but actually figuring out what does that mean for your enterprise or your business is a completely different ball game.

So the core of the hyper computer sounds like is simplicity, rather than providing piecemeal services, even if they're all provided under the Google umbrella, providing them with one name, uh, combining the hardware, the tus Nvidia, GPUs, high speed networks, unified storage, um, combining that all for the sake of simplicity. The unified storage system in particular, in particular is, I, I believe it'll help eliminate data bottlenecks, simplify data management for these petabytes scale dataset, and ensure that those extensive GPUs and TPUs aren't just sitting idle waiting for data. Um, we know how hard it is to get our hands on those Nvidia GPUs these days.

So once you have them, you want to be able to actually start running your workloads and doing what, what they were intended for, and not just sitting collecting dust. So I think this is directly going to address one of the biggest pain point pain points for those ML ops and large scale AI trainings. And that is the simplification.

And how do I get started? Um, on top of that, Google's commitment of 75 billion in 2025 to expanded data centers. I'll say it's, it's important, but honestly, um, it's just a requirement to be a leader in this space.

Uh, AI takes physical infrastructure. Uh, even if you're playing a paying a cloud provider, they have to have the physical infrastructure to support it. So that seems like a no brainer, a a good move by Google, but just something they've gotta do.

Um, for AI developers and researchers, I think this is gonna promise a more streamlined experience, uh, extract abstracting away some of that underlying infrastructure complexity is gonna mean faster model training, more rapid experimentation, more efficient inferencing at scale, more science, more cool stuff, uh, faster, simpler, easier Google Year say. Um, so in essence, sounds like Google's aiming to provide a turnkey supercomputing environment for ai, which is going to allow practitioner practitioners to focus on what matters. So a good move by Google, and in my book, Datadog is expanding its observability platform to cover more areas of modern IT environments, including user experience, security, and cost optimization.

The company aims to give organizations a more unified view across infrastructure, applications, and business metrics. This broader approach is designed to help teams detect issues faster, improve performance, and reduce operational complexity. And I think there's some really interesting pieces in what's in here.

Uh, in particular, I like the toto, uh, AI model that, uh, Datadog is, is brought in here. Uh, this model is trained on time series data, which is quite different from being trained on the general corpus of the, the great internet download. Uh, and so having the specialized AI that can very rapidly identify anomalies and, uh, capacity issues from that time series data that Datadog's been collecting, is this more than just three linear regressions in a trench coat as, uh, Justin Warren's inclined to say quite probably.

There seems to be a little more complexity to this, a little more awareness of, uh, longer term trends rather than just being stuck on the shorter tip trends. Uh, Michael Wetten, who is the vice president of Datadog, uh, product at Datadog, says that in general, they, they're widely used monitoring platforms and they're moving to expand the reach and scope of the things that they're building out. Uh, this includes having some more experimental products, uh, the product analysis suite that they, uh, acquired a little while ago and building out tools to support the, um, data, uh, management and data science teams.

Definitely some tools around looking at, uh, automated analysis of your data and getting good data quality to feed up. Of course, Corey said on the, the essential element of data quality being vital for getting your business data into your AI application and getting good outcomes from this. Uh, it's not the first acquisition.

There's a definitely a bunch of, uh, acquisitions along the way for, for Datadog, uh, others like Meta Plaine added in, adding more Ai, uh, models into, into the offering, um, and getting more complexity or getting some handling of the greater complexity that we're seeing in applications as, uh, they're being built with AI tools, uh, AI applications, using those sets of data, having very different demands on workload, on, on resources, on networking. Quite different from the applications that we've previously run the enterprises, uh, particularly as we start to see that move towards agent, where it's an agent talking to another agent nearly as often as it's a piece of data or a user talking to the agent. Uh, having observability across these increasingly complex environments is absolutely vital.

And that's where Datadog has been focusing as getting the best telemetry data out of your applications, out of your ai, out of your infrastructure, and getting a holistic view of what's going on, making sure that you are getting the best value for the resources you are using, and that your users are getting the best experience of accessing your application. Now it's time for a little bit of a closer look and open AI has announced its largest acquisition to date, agreeing to acquire io, an AI device startup founded by former Apple executive Johnny I of well known as the, uh, chief designer at Apple. Uh, it's an all equity deal worth approximately six and a half billion dollars.

And Sam and Johnny, uh, gotten together and talked about how wonderful everything is working together. As part of the merger. Johnny and I will assume a significant creative and design responsibilities across both Open AI and io despite this merger i's design firm love from will continue to operate independently.

And that leaves me wondering, hang on, we haven't got the design bit and OpenAI had a lot of ai. What did they just spend six and a half billion dollars on? Corey, Your guess is as good as mine ster.

Um, so Johnny Ives, uh, is the brain behind some of the most iconic products like iPhones, iPads, and, and MacBook errors for Apple. So, uh, IO is a physical ai, um, a physical AI company. What does that mean?

Alistair, I don't think I have the answer for you, but Altman said that their mission would be to figure out how to create a family of devices that would let people use AI to create all sorts of wonderful things, which sounds wonderful to me. Um, are we, maybe going back to like a handheld for ai, why wouldn't this just sit on our phones? Uh, what do you think ster?

Well, it does make me wonder whether Johnny, I've left Apple because of the disarray that's been going on with AI at Apple. Uh, apple seems not to be able to deliver the sort of AI services that customers are expecting, though the reports I've heard of Apple Intelligence is that it's been underdelivering and maybe Johnny Live saw this coming and actually made a move to leave in order to have better options. 'cause of course, you couldn't go straight from Apple to somewhere like, uh, open ai.

So maybe there was some vision from Johnny Ive to get out and that he had ambitions to build much better ais than he felt that Apple were gonna, maybe this is why now he and, uh, Sam Altman have gotten very, very chummy together. Maybe this is actually the biggest acquihire that we've ever seen of bringing Johnny into to work with Sam. My suspicion is that it's not so much of a hire as a partnership that's going on here.

And that's very much the way the, the Post from Sam and Johnny, uh, on the, the Open AI page talks about is this, this partnership to build solutions together. And so this seems to just be the, the way they could make, make, uh, suite technology together. Cory, do you think that hardware as, as you say, um, this may end up being hardware devices in our hands?

'cause that's what we know Johnny, I for, I don't know. I mean, when I think of effort, I, when I think of Johnny as legacy, I think of these handheld devices, but who knows? Um, we know it, it's a hardware focus.

I don't know where these devices are gonna live. I'll just a quick anecdote that if you haven't looked at the Open AI page and this photo of Sam and Johnny, uh, you have to look. I mean, I think a true partnership, it's gonna be, they look like they could be brothers in the photo.

Um, I think they're both very excited for this. Um, so that was just pretty heartwarming to see a heartfelt photo of the two together on Open Eyes page. But speaking of EQU hires, this is not the only hire that OpenAI has made.

So I think OpenAI is getting serious about this AI hardware because they also hired the former head of Meta's Orion augmented reality glasses in November to lead robotics and consumer hardware efforts. So, you know, I don't know what it's gonna look like, but open AI is, they're up to something here and I, I can't wait. I mean, I think about when we had the first iPod, we wouldn't be able to conceptualize that.

And are we on the brink of another event like the unveiling of the iPod? Who knows? And I think that's what the hope is, and I really do hope we get to see something truly innovative and, uh, industry changing come outta this partnership because it's a lot of money to be, uh, putting out on the table to bring a partner into your organization.

As we are looking at our wonderful week, this episode is bring coming to you on Wednesday, the 28th of May, which is also the first day of Security Field Day. Uh, Tom Hollingsworth, my Ural whole, uh, partner in this particular, uh, piece of snark, uh, Tom is, is out in California and has two great days with a, a whole collection of people, uh, both delegates and presenting companies for Security Field Day. Of course, next week is my turn Cloud Field Day returns to the, uh, San Francisco area.

We'll be, uh, live streaming as usual on LinkedIn and Techstrong on June the fourth and fifth. And again, I have some awesome people there. There's gonna be quite a lot of focus on storage as it turns out, and, and object storage and massive scale object storage.

Then the following week, it's Tom again. Uh, tech Field Day Extra at Cisco live US the week, uh, June 10th and 11th. Um, and then Tom is also on, he gets a little bit of a break because it'll be July 9th and 10th from Networking Field, day 38.

Thanks for watching The Tick Field Day rundown. You can catch new episodes every Wednesday as a YouTube video or on your favorite podcast application. The Rundown is also streamed on Techstrong TV as well.

You can often catch us on other Techstrong and Future group programs. We'll be back next Wednesday to talk about all the IT news of the week. That was then for myself, forgi, and for all of us here at Tech Field Day.

Wishing you and yours a great day.