The Feature Flag Consolidation Map - Who Got Bought in 2024 to 2026

In under two years, four of the more interesting names in feature flags and experimentation stopped being what they were. Statsig got bought by the biggest lab in AI. Split got folded into a DevOps platform. DevCycle got absorbed by an observability giant. And Bucket quietly became Reflag. Three of those are real acquisitions; one is a rename. All four matter if you are choosing a tool this year, because ownership is no longer a footnote - it decides whether the product you adopt gets invested in, repriced, or slowly turned into someone else’s internal infrastructure.

So here is the map. The four changes, what each one actually means for a buyer, who is still independent, and how to choose so you do not get stranded. Every date, valuation and funding figure below is drawn from our individual tool reviews and links back to them.

Statsig to OpenAI - the strategic absorption

OpenAI acquired Statsig on 2 September 2025 for a reported ~$1.1 billion in an all-stock deal. Statsig founder and CEO Vijaye Raji became OpenAI’s CTO of Applications. Statsig was founded in 2021 in Bellevue, Washington, by ex-Facebook engineers, and it is one of the strongest tools in the category - flag and config checks are unlimited and free on every tier, and its experimentation engine (sequential testing, CUPED) is the best of the bunch.

What it means for buyers. Nothing has broken. Statsig still operates independently from its Bellevue office, the free tier is still generous, existing customers are still served. The risk is roadmap direction. When a product is absorbed by a company as large and fast-moving as OpenAI, the open questions are the ones you cannot answer from the outside: does the standalone product stay a priority, does pricing hold, does the independent offering keep getting invested in, or does it slowly become internal infrastructure? None of that is decided. Statsig is not open source and has no real self-host, so if the direction shifts, you have no code to fall back on - which raises the stakes on the ownership question compared with an open tool. It is not a dealbreaker. It is a multi-year bet on a product whose owner now has bigger priorities, and you should price that in.

Split to Harness - the standalone product is gone

Harness completed its acquisition of Split with the deal finished on 11 June 2024, and rebranded it Harness Feature Management and Experimentation (FME). Split was one of only two flag platforms that took experimentation seriously - real statistically rigorous experiment analysis, not a percentage slider with a chart. Harness is a large, well-funded DevOps platform (a $240M Series E in December 2025 at a $5.5B valuation, led by Goldman Sachs), and Split is now a module inside it.

What it means for buyers. This is a rebrand, not a sunset - the engine and SDKs are unchanged, and existing users needed no application code changes to migrate. But standalone Split no longer exists as a product you can buy on its own. The old split.io/pricing page now 301-redirects to harness.io/pricing. FME is sold as a module under the Harness Enterprise plan on a contact-sales, MAU-based basis, with no public per-tier pricing. Two concrete risks follow. First, repricing - as your standalone Split contract turns into a Harness platform contract at renewal, the number can move, and you cannot price the flags in isolation because your cost depends on which other Harness modules you license. Second, migration friction - SSO must be reconfigured as a new Harness connection, and several admin API endpoints changed. If you just wanted Split’s flags and experiments as a point tool, that path is closed; you are buying into a DevOps platform now, and you should evaluate it as a platform commitment, not a flag choice.

DevCycle to Dynatrace - independence traded for an observability suite

DevCycle is now part of Dynatrace, acquired in an announcement around January 2026. DevCycle’s own site now reads “DevCycle is Now Part of Dynatrace.” DevCycle was a developer-focused, OpenFeature-native flag platform with the best edge story in the category - Cloudflare Workers evaluation, config propagated globally in about a second, and published pricing, which is rare here. It grew out of Taplytics and launched as DevCycle in 2022 from Toronto.

What it means for buyers. The acquisition cuts both ways. On the positive side, it is validation and an exit - the tech was good enough for an observability giant to buy. On the risk side, standalone roadmap, availability and pricing are now open questions. As DevCycle folds into Dynatrace’s enterprise observability suite, the independent product you sign up for today could be repackaged, redirected, or repriced. DevCycle is OpenFeature-native, which is the one thing that softens the blow - because the SDK layer follows the open, vendor-neutral OpenFeature standard, you can swap providers with less rework than a fully proprietary tool would allow. That portability is a real hedge against the uncertainty. Still, the honest read is strong technology, real longevity uncertainty, and a smaller ecosystem than the incumbents to begin with.

Bucket to Reflag - a rebrand, not an acquisition

Include this one so nobody gets confused, because it is easy to misread as consolidation. Reflag is the tool that was called Bucket until 4 September 2025. bucket.co now 301-redirects to reflag.com. No acquirer is involved - this is the same company under a new name. Bucket started in 2021 as a feature-metrics and feedback product, pivoted into feature management in June 2024, then rebranded to Reflag.

What it means for buyers. Practically, nothing about ownership changed - it is a deliberately narrow tool for B2B SaaS on TypeScript, focused on per-account feature entitlements, and it is still small (around $3.7M raised). The only thing to carry forward is name recognition: if you see Bucket referenced in older comparisons or docs, that is Reflag. The reason it belongs on this map is precisely that a 301-redirect from an old domain to a new one looks identical whether the cause is a rebrand (Bucket) or an acquisition (Split to Harness). Knowing which is which keeps you from misreading the market.

Who is still independent

The good news for buyers worried about consolidation - most of the field has not moved. Here is where the still-independent players stand.

• LaunchDarkly - the enterprise standard, still independent, having raised $200M at a $3B valuation in 2021. Well-capitalized; the concern with LaunchDarkly is its MAU-based pricing and renewals, not its ownership.
• PostHog - founded 2020 out of Y Combinator’s S20 batch, independent, and a reported $1.4 billion valuation after a Series E in October 2025. Growing fast and funded to stay.
• Flagsmith - bootstrapped and independent, founded 2018 in London, roughly 12 people on around $1.5M revenue. Small, but no acquisition roadmap risk, and BSD-3 open source, so the code outlives the company either way.
• GrowthBook - founded 2020, a Y Combinator company, with a reported Series A of $22.6M from Khosla Ventures and Nexus Venture Partners. Independent, and MIT-licensed at its core.
• Unleash - built by Bricks Software AS, founded 2015 in Oslo, Norway. Independent and mature (an 11-year track record), and AGPL-3.0 open source. Note the honesty caveat: its reported venture funding is unconfirmed, so we do not state a figure here rather than repeat an unverified number.
• ConfigCat - bootstrapped and unfunded, founded 2018 in Budapest, roughly 14 people. No VC pressure, which is exactly why its flat, no-per-MAU pricing exists.
• Optimizely - independent in the sense that it is not part of a larger buyer, but it is not an independent startup. It is PE-owned by Insight Partners, which bought Episerver in 2018 at a $1.16B valuation and renamed the whole company Optimizely. So its ownership is settled, but it is a private-equity-owned digital-experience platform, not a focused, founder-led flag vendor - worth naming precisely.

What this means if you are choosing today

The lesson of the last two years is not “avoid tools that might get acquired” - almost any of these could be next, and an acquisition is not automatically bad. The point is subtler. What protects you is not the tool staying independent forever. It is whether the outcome, if it gets bought, leaves you stranded.

Two things decide that.

First, the license. A permissive open-source license means the existing product cannot be taken away from you, whatever happens to the company. Flagsmith (BSD-3), GrowthBook (MIT) and Unleash (AGPL-3.0) all give you that floor - if the vendor is acquired or repriced, you can keep running the code you already have. The closed-SaaS tools give you no such floor. If Statsig under OpenAI, or Harness FME, or a Dynatrace-owned DevCycle ever changed course, you would have no code to fall back on - only a migration project.

Second, settled ownership. A tool whose ownership question is resolved - either genuinely independent and well-funded, or acquired by a parent with aligned incentives - is a safer multi-year bet than one that looks like an acquisition target with a closed core. The clearest risk to design against is absorption and repricing: a standalone product gets folded into a larger suite and re-quoted as part of a bundle at renewal. That is precisely what happened when Split became a Harness module, and it is the open question hanging over DevCycle inside Dynatrace.

So, choosing today:

• If you want maximum protection against ownership churn, favor an open-source tool you can self-host - Flagsmith, GrowthBook or Unleash. A bad outcome still leaves you with a product you control.
• If you want a proprietary tool, favor one whose ownership is settled and incentives are aligned - an independent, well-funded vendor like LaunchDarkly or PostHog, or a bootstrapped one with no exit pressure like ConfigCat.
• If you are eyeing a recently acquired tool like Statsig or DevCycle, the tech may well be excellent - just go in knowing you are betting on a roadmap that now answers to a much larger owner, and weigh whether OpenFeature portability (DevCycle) or a clean exit plan (Statsig) softens that bet enough for you.

The real failure mode is not acquisition itself. It is being locked into a closed tool that gets absorbed and repriced under you, with no code to fall back on and a migration you did not budget for. Favor settled ownership or a protective license, and that failure mode mostly disappears. Every date, valuation and funding figure on this page is drawn from our individual tool reviews and was verified against each vendor’s materials on 23 July 2026. This category is moving fast - we re-verify regularly and will update this map as the next deals land.