State of the Data & Analytics Market – Q2-2026

00:00:00: Yeah, my new favorite term for agents is agentic spaghetti.

00:00:03: Because you can't really right now understand what they're all doing how their performing.

00:00:09: are they hitting the goals that they've been told to hit?

00:00:22: Hello from

00:00:23: hello Kirsten.

00:00:24: great to have you in The Data Culture podcast.

00:00:26: again we do our quarterly market update.

00:00:28: I'm looking forward.

00:00:29: there's

00:00:30: a lot of fun things talk about Right Now.

00:00:31: For Sure

00:00:32: Absolutely and We Are Right Now In Beautiful Colorado.

00:00:36: You just finished our annual retreat event for vendors, like a strategy workshop.

00:00:42: What was your take?

00:00:44: The topics were very highly engaged with the audience.

00:00:48: I think one of the best things about to retreat in general is that we don't talk at the room, we talked with the room.

00:00:56: So we don't get the luxury of presenting a slide deck without interruption.

00:01:01: their interruptions are the best part as all these really smart vendors in the room help expand on the topics and the insights that we've got.

00:01:10: so yeah it's uh... It's a really fantastic couple days

00:01:15: And we can actually share a few of the insights.

00:01:18: Obviously, AI is disrupting also this software industry in many ways and that where... We had lots of discussion topics around it.

00:01:26: Forgive me I'm bringing up my notes so little bit to read them?

00:01:30: Absolutely!

00:01:32: As usual.

00:01:33: i think they should structure this market update into three parts.

00:01:37: You got vendor side which now very timely with event just being finished.

00:01:42: Then obviously we need to take a look at the user side, what's going on there.

00:01:46: And then we can try out few of the latest investments and M&A activity that went on in the market.

00:01:50: Fantastic!

00:01:51: Now should start with the vendors?

00:01:53: We certainly count.

00:01:54: yeah What was your big takeaway from The Conversations this week?

00:01:59: I know we covered things like semantic layers... ...and i want talk about headless environments little bit.

00:02:04: where do you wanna start?

00:02:06: Yeah..I think obviously AI is everywhere and the effects of AI.

00:02:10: I thought that discussions around AI governance to be very interesting.

00:02:15: so there obviously is a need also in the user base, we should cover what's going on there.

00:02:22: as you said semantic layer now mandatory.

00:02:25: it seems increasingly like key architectural element making agents actually work.

00:02:34: I mean, we talk about semantic layer for quite a while and it's hot topic on the vendor side but that was very clearly positioned.

00:02:43: I can tell you because i go to a lot of events as you do as well... ...I have been to analytic vendors who normally would start their keynote speech with us as analysts with Let Us Show You Our Platform.

00:02:57: And now what they're showing us is, let's show you our new semantic layer or how we can help your agents come and speak to our platform.

00:03:07: Things are shifting dramatically.

00:03:09: one of our vendors made a funny comment about Claude when I had it up on the screen this week.

00:03:15: he said oh good to see that using every BI vendor's new UI And it's because they've been able to expose the data about, the metadata of the semantic layers and so on.

00:03:28: To make agents in AI work better.

00:03:32: I know In my notes that use the word this is what going power A.I go forward

00:03:38: Absolutely yeah!

00:03:39: This comment was a common theme.

00:03:43: The big question how does future look like?

00:03:46: In terms especially interface with user?

00:03:49: And then there seemed to be quite a common understanding that gen AI or agents, both of them will be the future interfaces and vendors have to adopt on it.

00:04:00: Yeah!

00:04:01: It's almost startling in a way but also very exciting.

00:04:05: you can see the power being able drive your own experience through these platforms like Claude or OpenAI are other tools we've seen now regularly.

00:04:18: We know that vendors are pivoting to be prepared for that.

00:04:22: Some of them, they're even putting some their most important feature and insight features sets in the MCP layer or as a skill For you just simply bring into your new user environment And then to leverage it.

00:04:37: In that type of work flow.

00:04:39: I think we'll see more going deeper into this year.

00:04:45: users will continue to abstract the UI, should their side of the table.

00:04:50: And I think that vendors understand and they're working really hard be able to accommodate

00:04:55: it.

00:04:55: Exactly!

00:04:56: Right now we can add value as a user interface while all skills are mostly still with applications but SAI or GenAI vendors get better at that?

00:05:11: That's an interesting question.

00:05:13: How yeah, how do we see the future there?

00:05:16: Yeah

00:05:16: We got to talk about revenue and money a little bit.

00:05:19: this costs them when we talked about the end.

00:05:21: users will Talk About A few Things There but I know that a lot of The Vendors Are Kind Of Pivoting Right Now because Suddenly as we were Just Talking People that are using their platform aren't people as much and more, they're agents.

00:05:35: And agents are having a much quicker or iterative conversation with the platform.

00:05:40: A lot of vendors is starting to talk about new pricing models for users To speak to the platform through agentic workflows.

00:05:50: I think it's interesting but its gonna add in you variable.

00:05:53: It has revenue streamed from the vendor.

00:05:55: But there will be cost streams.

00:05:58: What is your position on?

00:05:59: Well, cost is a big topic for the users.

00:06:01: For sure and the vendors are really trying to figure it out now how we handle this new world.

00:06:08: in a way We see a big trend towards capacity pricing that's I think one or consumption based pricing.

00:06:17: so both as happening customers really dislike consumption-based pricing, they really dislike the unpredictability of it.

00:06:33: That's a big concern.

00:06:35: so we had vendors in the room that I think there were both strategies being present as some said.

00:06:40: well of course consumption basic pricing is what we do right now especially to also handle the agents and their tokens.

00:06:48: But capacity-based pricing, some winners said it's better yet in the discussion with users.

00:06:54: It is something that they like

00:06:56: better.".

00:06:57: So the models are switching and changing?

00:07:00: I think my takeaway... We were here last year where a lot of people who attend our event kind of alumni come every year And we talked about this earlier.

00:07:10: but things like charging for agent work flows against platforms as brand new.

00:07:14: it's less than a year old.

00:07:16: It didn't come up last year, but as the utilization of agents is becoming more prolific they're not going to stand by the vendors and have their platforms kind of taxed a bit by that type or iterative workflow.

00:07:34: so yeah its gonna be interesting

00:07:37: absolutely.

00:07:39: Maybe.

00:07:39: to move the discussion over to the user side, what we also discuss is that.

00:07:43: We see us in our research data and at their users site as pretty much dividing anyway.

00:07:49: almost fifty-fifty In companies that are quite active really advancing AI do a lot of things but then there's like half.

00:07:59: they're Really slow not doing a lot Of Things.

00:08:02: And The vendor said the big question yet how To work with That?

00:08:06: why it was?

00:08:06: It's Not A Uniform prospect base in a way anymore.

00:08:10: But they have some that are really demanding and others where, yeah cannot sell AI

00:08:16: anyway.".

00:08:17: And it's interesting for users because depending on who you're in your AI journey whether very mature or maybe leader kind of the space which has spent time money effort building an AI foundation will let you do sophisticated work?

00:08:35: The companies that do that are very different than the other side in the coin, which is just kind of running around doing AI without a lot of governance and a lot guardrails.

00:08:44: Which to me it's scary.

00:08:47: on behalf our user customers I do believe this is a technology that requires governance.

00:08:55: And if you go too far down the AI road or agentic road, whichever one that your really focused on right now... If don't have the right guardrails in place and haven't built foundation?

00:09:07: I'm worried about people!

00:09:09: I gave talk at an event recently.

00:09:12: we asked couple hundred for show of hands.

00:09:14: who's doing Agenic AI?

00:09:17: all two hundred hands went up.

00:09:20: Everybody who's super comfortable with the foundation, the governance layer that you have raise your hand on maybe four people in their room.

00:09:27: And I looked at The Crowd and said,"The rest of you need to stop

00:09:30: it.".

00:09:30: It gets slow down!

00:09:42: have or we do.

00:09:43: enter a second wave, level now if you want.

00:09:46: First priority was really to get it out into the organizations so like does everyone has access?

00:09:54: What are use cases and how can they be used beneficially

00:09:57: etc.?

00:09:58: But now it's really, you know that is being used and the usage is proliferating across the organization.

00:10:05: Now it's starting to become important to look after.

00:10:09: especially autonomous agents are not creating big chaos.

00:10:15: We were joking earlier.

00:10:16: I said my new favorite term for agents is agentic spaghetti because really right now understand what they're all doing, how their performing.

00:10:26: Are they hitting the goals that they've been told to hit and so on?

00:10:30: And as we start to move from single agent workflows into multi-agent sophisticated work flows... That type of governance around it is super important!

00:10:41: At the same time, Governance Around Cost is becoming sort of a thing.

00:10:46: lots very interesting stories about big name brand companies Burn through, I guess is the new term.

00:10:55: Quadonomics Is The New Term.

00:10:58: That They Burn Through Their Tokens So Fast that they Find Themselves Two Or Three Quarters Into The Year and They Already Spent Their Money.

00:11:08: And Those Numbers Are Getting Pretty Scary.

00:11:11: It's Because Even Though Token Costs Has Dropped Like Ninety-Five To Ninety Nine Percent Which Is Great.

00:11:18: Adoption As You Just Mentioned is prolific as an understatement.

00:11:25: Met and end user company the other day, And the person I was speaking with said we're so excited We are bringing Claude into our environment for all thirteen thousand employees.

00:11:36: And I went, oh okay.

00:11:37: Yeah and i started to ask governance questions How are you gonna govern the budget?

00:11:41: What kind of guardrails do they have in place?

00:11:44: any answer was a little scary.

00:11:45: You know it was while our ceo told us we need to do this and were likely to add open AI chat tbt To our lineup for all thirteen thousand a little later in The year.

00:11:56: yeah

00:11:57: that sounds like A big bill coming at the end of the year lady.

00:12:02: Yeah, but I completely agree.

00:12:04: So we have these genetic workflows.

00:12:07: now agents.

00:12:09: We see sometimes it's they're not very efficient in into token usage.

00:12:15: then They get more autonomous so that you get more to normal since spending yes money in a way using.

00:12:20: so um?

00:12:21: We see that the more mature more advanced companies already have that as a concern and really can only give The advice that everyone else should be very aware of that.

00:12:30: that is will become a major topic.

00:12:33: And I think, the big topic here next year at our retreat on behalf of the end users but also kind in collaboration with what the vendors are looking there is innovations coming out where they're And then they're deciding which model and where to disperse the prompt, too.

00:12:50: The decision is based on its ability to give you the right answer but also the cost factor.

00:12:56: So should you send this prompt to big model or small domain model?

00:13:01: Or highly optimized model?

00:13:04: That's a good thing.

00:13:05: New memory strategies around these environments This new UI To decide how make your work cost less from a token standpoint.

00:13:16: So I think next year when we sit here at Devil's Thumb Ranch, will talk about some of those more sophisticated ideas on controlling costs for end users?

00:13:26: I'm pretty sure about that.

00:13:27: yeah another recurring topic was data-seventy.

00:13:30: We highlighted it last year as like an upcoming topic.

00:13:34: and now what do you have?

00:13:35: new research data also in there but can very clearly see is really sophisticated topics.

00:13:43: It didn't go away in any way, but on the contrary it's now part of a strategy for many.

00:13:48: But they are also starting to execute on that.

00:13:50: I think there is an interesting point where companies take this so seriously... ...that they're willing to change architectures and maybe even provide us with their technology providers.

00:14:02: We see major investments again into data governance in terms of understanding what my critical data actually looks like.

00:14:10: What can i do?

00:14:12: Do I maybe even find on-premises systems for that?

00:14:18: We see in the survey, we did increasing usage which is kind of interesting.

00:14:22: So everyone was a bit surprised and they were like oh my god yeah!

00:14:26: After so many years trying to do pushing everything into the cloud... ...we are now at least on a smarter scale obviously but it's coming back as a return.

00:14:41: And what else do we see?

00:14:44: We still have a bit of uncertainty on both sides, vendors and users.

00:14:49: How to actually become sovereign in terms of work?

00:14:53: Because there are so many things to address.

00:14:55: it starts with security privacy topics.

00:14:59: then the big question is also about data.

00:15:01: location should be certain.

00:15:03: geography Can my often than the hyperscaler or any other type of application provider, what can they actually guarantee?

00:15:14: That's sufficient for me.

00:15:17: So there is a multitude of different things to consider and that makes it I think super interesting right now because There Is A Regulatory Site To Many Things To Many Industries That Clearly Say Okay This Is What You Have To Do And Then It'S Pretty Clear What you have to do.

00:15:33: but For many other things, it's a real decision for the company.

00:15:37: They can say well okay from me if I'm at let us say regional in the region of the cloud provider and contractually its confirmed that my data is not leaving.

00:15:49: then companies come to you saying oh thats fine My data sovereignty requirements are fulfilled And others says no That's not an option for me.

00:15:58: There is the US Cloud Act that says, okay there can be access to that data and so from my data sovereignty requirements?

00:16:06: So then we see these companies have to figure it out themselves.

00:16:09: basically they find their own strategy And vendors are right now trying to figure out what they can do about it

00:16:15: Right where to meet end user requirements.

00:16:18: I have to say We see politics come into tech off and on kind of way.

00:16:24: I'd say sovereignty is really affected by, primarily right now it's being effected US politics but i think the room of vendors agreed that globally sovereignty around your digital footprint year data Is important everywhere?

00:16:38: And we're going to see a more hybrid approach in the vendors are rushing into That!

00:16:47: premium service.

00:16:49: And I think that may start that way originally, but I think over the next couple of years it's going to get commoditized pretty quickly and To have a cloud environment you're going to have to have really comprehensive sovereignty solution available

00:17:03: right now.

00:17:04: so we see rising majority and Very interesting also.

00:17:09: our survey showed its global phenomenon such as something in Europe or Asia somewhere And we even saw that the North American companies are investing more than anyone else, to basically invest in architecture and governance.

00:17:23: So it's definitely a topic that will continue be of interest to all sides in the market.

00:17:32: I agree.

00:17:33: I used a fun word a few minutes ago.

00:17:35: what do they say?

00:17:36: Claudeonomics which is kind of fun.

00:17:38: did you hear sasspocalypse this week?

00:17:41: We talked a lot about that.

00:17:43: Yeah, but this is the big topic I think emerged as term to describe a bit of let's say loss of confidence for investors in enterprise software companies and it was very clearly at the beginning of the year we saw a lot share prices really plummeting going down And they didn't really recover yet.

00:18:06: What we saw is that investors are questioning the business model of enterprise software companies, because obviously AI's disrupting a lot here and what can say it still pretty unclear.

00:18:23: so some also have vendors in their rooms said well definitely this will rebound yeah?

00:18:28: It'll be clear at some point what value Enterprise Software has and that it cannot be replaced by AI, everything.

00:18:36: We know about the interface yes but not the logic nor their valued brings And others were not so sure.

00:18:43: So this says to play out.

00:18:44: But we see says Pocalypse means... It's

00:18:48: a body word say?

00:18:49: Body

00:18:49: Word!

00:18:50: Yeah..it was definitely a shock.

00:18:52: too many organizations and yeah often share prices didn't return.

00:18:58: What that means also is we see less activity.

00:19:01: So both, less M&A meaning acquisitions and less investments.

00:19:05: but still there's some things going on.

00:19:08: But the definition of quality has changed in terms.

00:19:11: what are quality vendors that investors want to invest or they want to acquire?

00:19:17: Well I know for a fact that in Silicon Valley here and a lot of investment is being made there.

00:19:28: The SaaS companies are kind of pivoting from being the place, the target, the platform where you went to do your work to be less of the landing spot in more of the enabler-to-do AI workloads on agents.

00:19:42: so it will an interesting year.

00:19:44: watch that pivot occur see how successful some or unsuccessful others.

00:19:51: I think It's going to force that conversation around openness again, in a way that probably hasn't happened at the past.

00:19:59: And I think it is gonna change the landscape.

00:20:03: We're seeing a lot of investment in the market A lot of purchases.

00:20:07: What did you see?

00:20:08: You gave this really great in-depth M&A talk yesterday with what are couple high points.

00:20:15: Absolutely, yeah.

00:20:15: So on the acquisition side I think SAP really stood out over the last month.

00:20:20: so they're quite relative.

00:20:22: and MDM vendor there were Dremel a lake house engine if you want to say it like that then Pryolabs.

00:20:30: They have a tabular foundation model?

00:20:33: I just kind of asked

00:20:34: that whole section yesterday.

00:20:36: That was something i haven't looked at...so I found very interesting..I think he had three or four different companies that we are doing activities in that zone

00:20:44: Absolutely.

00:20:45: And now with the acquisition, it used to be more like a research topic and then what very scientific.

00:20:53: so there were... How

00:20:54: do I add that down right in a while?

00:20:55: Yeah

00:20:55: no!

00:20:56: Now it really came out of that.

00:20:59: we have another company called Fundamental.

00:21:02: they got more than two hundred fifty million investments with very little revenue also.

00:21:12: So obviously there's interest and trust of investors into this new idea.

00:21:17: off a tabular foundation model to handle, yeah?

00:21:20: To build something like an LLM just for structured data.

00:21:23: so we see how that turns out.

00:21:25: SAP has big plans.

00:21:26: they announced it today want to use it as their AI lab in the way or gas.

00:21:31: more than a billion dollars down to that which is significant.

00:21:35: There's a lot going on and SAP I think makes pretty clear that they want to become a series or remain, however you wanna see it.

00:21:42: A serious player in the data management market also Data and AI so really bolstering their platform.

00:21:50: The

00:21:50: incredible investment, I mean you just said more than a billion into this initiative.

00:21:56: We're seeing that physically too with data centers across my market here in the United States which is causing a lot of discussion around electricity consumption water consumption use of land.

00:22:08: I think we're going to see some interesting things happen in the AI space from an investor standpoint because there are so much of this happening and they starting be a little pushback, at least here on my market.

00:22:20: Do you see it as well when Europe?

00:22:22: Yeah say thing that... but with data center build out is also happening in Europe right now.

00:22:31: But the big consumption that's going on in terms of training and using the big foundation models, obviously.

00:22:38: That is more an North American topic.

00:22:41: so I would imagine you see a lot pushback because investments are just unbelievable... There

00:22:46: was one being proposed here at Western side which where we're today or the completed data center when it's hundred percent functional will consume more electricity than the entire state times three.

00:22:59: and the entire physical footprint of the data center is proposed to be about the size of Manhattan, which it's kind hard to wrap your head around.

00:23:11: The Western United States has a very big place but that would one of the biggest buildings I think ever created.

00:23:17: so it'll interesting.

00:23:19: what else

00:23:20: are you seeing in there?

00:23:23: Interesting to me is that there's a lot of investment going on into what I would call BI players, and who have thought about it?

00:23:31: Obviously they are... They call themselves AI-first analytics or similar.

00:23:36: So obviously they have AI in mind or an idea how to position them in their environment.

00:23:41: but still a lot goes on!

00:23:44: Both SIGMA and Omni got investments and quite significant investments.

00:23:48: Significant!

00:23:48: Eighty million, hundred twenty million both at unicorn valuations, so they're both valued more than a billion.

00:23:54: So that's really interesting but also new ones still coming up here.

00:23:58: one very recent example would be golden analytics

00:24:02: and we know that founder

00:24:04: exactly Francois Einstein many you know him in the industry there being very prominent figure at Tableau before.

00:24:11: And so interesting yeah?

00:24:14: There is money going into an segment that we would have considered maybe being one of the fundamental but pretty disrupted.

00:24:29: I found from the investment side, it was a very interesting part.

00:24:32: and another one obviously is Clickhouse where they attracted over a billion dollars in investments.

00:24:38: Seven hundred fifty million of them only in last eight months with two rounds.

00:24:43: So that's definitely also, Bender.

00:24:45: That attracts a lot of money and shows the importance of good data foundation for any AI initiative.

00:24:55: And I'll add to our research.

00:24:57: when we talk about Data as a Foundation For AI We want make point that the data quality issue is still there.

00:25:03: We know it was the biggest challenge for all you who are listening today in doing AI.

00:25:11: It's hard to stop what you're doing when being pressured from your CEO or the board.

00:25:16: To do AI, I can tell that we see it's growing and smart companies are going put on breaks just a little so they make sure their data is better before moving forward.

00:25:27: Yeah!

00:25:28: And even in our own company We are exposing very soon.

00:25:34: We will expose all content of research we have via an MCP server.

00:25:39: So if anyone wants to use that, they can.

00:25:42: But our biggest discussions are right now around governance.

00:25:47: How can it be done?

00:25:48: how can we secure the whole environment and make clear what cannot be done?

00:25:55: so it's affecting even companies of our size...

00:25:59: We're in the West!

00:26:00: And my phrase for this would be Right.

00:26:02: Now AI seems like The Wild Wild West Right?

00:26:06: It's a little crazy.

00:26:07: And as people that spend a lot of time doing this, I would again tell the end users who are listening today to show a little bit of conservative decision making before you run off and do wild-wild west sort things with AI.

00:26:23: Yeah, sure!

00:26:24: As usual it has been a pleasure.

00:26:25: we continue to help you tame this surprise view.

00:26:28: Good work with our research.

00:26:31: Thanks for the market roundup this quarter and see you soon

00:26:34: again!