Your Vendor Just Bought Your Other Vendor

Somewhere in your estate there are two contracts you signed on different days, for different reasons, with two companies that had nothing to do with each other at the time. Look up who owns each of them today. After this year's run of deals, check rather than assume.

On 1 October, Dynatrace completed its purchase of Arize, the AI evaluation and tracing company, in a deal announced in August at $915 million. A team that picked Arize last year picked an independent specialist. Its contract now sits with one of the big platform vendors, and the team had no say in it.

That is one deal. The more useful question for a leader is what the run of them adds up to, and what it does to your next renewal.

I have been on the wrong end of this. At a tier-1 bank in Asia, the network monitoring tool we depended on, one of the most integral we ran, was bought by a much larger company. Not a word from anyone at the new owner, salespeople included. Eighteen months later the renewal arrived, up 500 per cent. We fought it, and fought it, and fought it, and in the end it came down to the clauses in a contract that was still in the original company's name. We ended up paying more than 250 per cent of the old price, then spent a fortune getting off the tool within a year.

Read this year's deals by what each buyer took home rather than by the headline price, and a pattern shows up.

Palo Alto Networks closed its purchase of Chronosphere in January, a deal it had announced at $3.35 billion, then bought Embrace, a mobile and user monitoring company, over the summer. Snowflake bought Observe, and its own filing puts the price at about $596 million. ClickHouse raised $400 million and, the same day, bought Langfuse, an open source tool for tracing language model applications. Cisco agreed in April to buy Galileo, an AI evaluation company, for its Splunk business. Elastic completed its purchase of Deductive AI, which builds agents that investigate incidents. And Dynatrace took Bindplane, a telemetry pipeline company, in April, four months before it agreed the Arize deal.

By my count that is at least a dozen observability deals announced or closed this year, and most of the buyers named above did not start out as observability companies. A data platform, a security vendor and a database company have each decided that watching production belongs inside their own product.

What they bought is the interesting part. A telemetry pipeline. Evaluation tools for AI. Agents that read an incident and propose a cause. A platform built to bring down the cost of storing telemetry. None of them is mainly a dashboard company. The buyers are paying for what sits between your raw data and a decision, which tells you where they expect the next contract to be won.

If that were the whole story, the advice would be short and gloomy. Your suppliers are getting bigger, so your hand is getting weaker.

There is a second half to it. Gartner's latest Magic Quadrant for observability, as reported by Network World in July, says three things at once. Consolidation continues to favour platform vendors that combine full-stack observability with AI. Telemetry cost management remains one of buyers' top concerns, and 5 per cent of Gartner's clients now spend more than $10 million a year with a single observability provider. And OpenTelemetry has lowered the barriers to switching, so vendors are trying to compete on analytics, automation and AI rather than on owning how your data is collected.

Meanwhile the independents raising money are selling against the bill. groundcover's $100 million Series C in July came with a line that sums up the whole pitch: "your vendor makes money when you send more". Coralogix raised $200 million in June on the problem of "runaway ingestion costs". Dash0 raised $110 million in March, aimed squarely at proprietary data formats and "runaway pricing".

So the market is being restructured from both ends. The big vendors are buying what sits between your data and a decision, and the challengers are attacking how you are billed. You sit in the middle with a contract priced under the old model, and the renewal is where you find out which model you are now on.

Each of them turns up on an invoice long after the announcement.

Independence becomes a promise. Dynatrace says Arize "will continue supporting both Phoenix, its open-source platform, and AX, its enterprise platform", and that its capabilities will be folded into Dynatrace over time. That is a reasonable thing to say, and I would not read it cynically. It is still a roadmap statement, and a roadmap statement does not bind anyone the way a contract term does. When Helicone, an open source tool for monitoring language model apps, joined Mintlify in March, its service went into maintenance mode, with security updates, bug fixes and new models still shipping, no end date given, and an offer to help every customer migrate to another platform. That is about as decent as these announcements get. It is also a different thing from the product its customers chose.

Pricing logic moves. Snowflake now lets customers put Snowflake credits towards Observe. Palo Alto launched Cortex XCOR on 1 October, written up by Chronosphere's co-founder Martin Mao, with "expanded consumption-based pricing models" across it. Splunk, under Cisco, has put activity-based pricing that weights search as well as ingest into controlled availability, and TechTarget reports that users stand to cut costs significantly. As Mike Leone of Moor Insights & Strategy put it to TechTarget, "A lot of Splunk customers have spent years architecting around ingest-based licensing". None of that is bad for you by default. It does mean a tool you bought by the gigabyte becomes a line inside a platform commitment, and a bundle is much harder to hold up against a rival quote. Confusion at the negotiating table rarely works in the buyer's favour.

Your advisers have an interest. Resellers, partners and consultants all gain something from where you land. That includes me, so put to me the question you would put to anyone else: what do you gain if I choose this?

It is tempting to read all of this as a market in revolt, with customers streaming away from the big vendors. The numbers do not say that.

Datadog's second quarter, reported on 6 August, brought revenue of $1.12 billion, up 36 per cent on the year. On the earnings call, its chief financial officer put net revenue retention in the low 120s and gross retention in the mid-to-high 90s. Those are the numbers of a vendor whose customers are buying more. The cost pressure is real and well documented, and it is why renewal conversations have changed. A migration wave is not documented anywhere I can find, and I would be wary of anyone who tells you otherwise.

That matters for how you walk into the room. Assume your supplier is frightened of losing you and you will be corrected quickly. Your bargaining power does not come from threatening to leave. It comes from being able to leave, and the supplier knowing you can.

None of this needs a programme. It needs an afternoon and the patience to reread a contract you signed some time ago.

Here is what I would say to a CIO whose monitoring vendor has just been bought. In my experience you get a year or two, at most, while the buyer integrates it and takes the best bits into its own product. After that you can probably stay on the old tool with no support, while they try to sell you the new products it is now integrated with, and it becomes a mess. The same salespeople and customer success people generally stay with the product for at least a year, I suspect so as not to worry the customers. You are never going to get the right answer from the vendor. You have some big choices to make, and your strategy has to be lined up before you make them.

Cisco is a prime example of a buyer settling on one platform. It bought AppDynamics years before it bought Splunk in March 2024, and AppDynamics is now sold as Splunk AppDynamics, with its development teams moved into Splunk. Splunk's own site says AppDynamics is not being sunset and that no migration is needed. That is a public commitment, and the place to make it binding is your own contract.

Being able to move your pipelines to a different tool with OpenTelemetry is the early indicator to watch. None of this is a reason to panic, or to wait for an announcement. The time to find out how portable your estate really is comes before the press release about your own supplier.

Make the map this month, while it is still just paperwork.