Why BBK killed OxygenOS: The Brutal Maths of Running Three Android Skins in the AI Era

In February 2022, at MWC, OnePlus told its community that it had listened. The plan announced five months earlier, to merge OxygenOS and ColorOS into a single operating system, was off. Gary Chen, who ran OS products at the time, put it in a way that sounded like a concession: the two would continue to develop on the same code paths, but they would remain separate brands after feedback from customers. The forums read that as a win. The company had proposed something, the community had pushed back hard, and the company had blinked.

Read that sentence again four years later and it says something quite different. The merger was never actually up for negotiation. The code paths were already shared, and OnePlus was confirming they would stay shared. The only thing on the table was the name on the boot animation, and that was the part OnePlus gave back. What looked like a reversal was a deferral, and on 16 July 2026, when the company confirmed to Android Authority and Android Central that it is winding down its North America and Europe operations and moving its phones to ColorOS, the deferral simply expired. OnePlus India's marketing director Ishita Grover told Gizbot that the Indian business is unaffected, and the company has not spelled out what the India roadmap looks like in software terms. But globally, the position is now official, and it is worth saying plainly that nothing about it is a surprise to anyone who has been reading the release notes since 2021.

The interesting question is not why OnePlus gave up. It is why BBK kept paying for OxygenOS as long as it did.

The Model that Made Three Skins Affordable

To understand the decision you have to understand the architecture behind it. BBK Electronics does not run a smartphone brand. It runs a portfolio, and the portfolio is engineered so that its brands do not eat each other. Oppo took the mid-market and the offline retail depth. Vivo took the same territory with a different distribution personality and, per Canalys, now part of Omdia, for Q2 2025, ended up leading India outright at around 21 percent share on roughly 8.1 million units, excluding iQOO. Realme took volume and youth, iQOO took gaming, OnePlus took premium, and it took the piece of premium that Samsung and Apple could not easily reach, which was the buyer who cared about what the software felt like after six months rather than the first.

The three BBK brands Canalys reports separately accounted for 43 percent of shipments between them in Q2 2025, before counting OnePlus and iQOO. That is not an accident of five companies competing well. It is a design, and in that design OxygenOS was not a cost centre. It was the entire reason OnePlus existed as a separate line item. A OnePlus phone and an Oppo phone could come off adjacent production lines and share half a bill of materials, and the thing that made a customer pay a premium for one of them was that the software was clean, fast, close to stock, and free of the stuff the other one shipped with. The skin was the product and the hardware was, to be slightly rude about it, the delivery mechanism.

That logic holds only as long as the brands stay genuinely apart. They did not and the Nord series pushed OnePlus down into price bands where its own siblings were already fighting, and Oppo and Vivo pushed up into premium from the other direction. By the time you reach 2024 and 2025, the segmentation that justified five brands had eroded into five brands selling similar phones at overlapping prices with converging software. India's numbers made the erosion impossible to ignore. OnePlus was the country's top premium brand in 2019 with roughly a third of the segment, per Counterpoint. Canalys had it at about 6 percent of premium by 2024. IDC recorded a 38.8 percent shipment decline in 2025, the sharpest fall of any major brand in the market. A great deal of that has nothing to do with software, and the 2024 retailer boycott and the green line display crisis did more immediate damage than any release of OxygenOS ever did. But the effect on the internal maths is the same. The premium position that OxygenOS was built to defend was no longer being held.

What A Skin Costs Now

Here is the part that does not get argued often enough, and it is the reason 2026 rather than 2023.

A custom Android skin used to be a launcher, a theme layer, an icon pack, a settings menu with some extra toggles, and a modest team to keep it from breaking. That is roughly what Motoblur was, what HTC Sense was, what LG's UX was. It was cheap enough that a company could maintain one purely as a marketing asset, which is more or less what most of them were.

A skin in 2026 is not that. It is an on-device model, it is a relationship with Google over Gemini integration and where the assistant boundary sits. It is a computational photography pipeline, a translation stack, a summarisation layer, a set of writing tools, and a QA matrix that has to be run across every SKU, every chipset variant and every carrier configuration you ship. Oppo has spent heavily on all of that for ColorOS, which the company now says serves more than 740 million users globally. The cost of building it once is enormous and the cost of building it twice, for two skins that share a codebase and now differ mainly in wallpapers and fonts, is indefensible in any budget meeting where somebody is asked to justify the second line.

So the fixed cost of maintaining a distinct Android roughly tripled at exactly the moment the differentiation it purchased fell to zero. Nobody has published a credible figure for what it actually costs to run a skin team at this scale, and that remains one of the genuinely under-reported numbers in this industry. But you do not need the figure to see the direction. When the thing you are paying for stops being visible to the customer, the payment stops.

Notice, too, that OnePlus has already been telling us where its priorities sit. When Samsung and Google moved to seven years of updates, OnePlus president Kinder Liu argued that the seven-year commitment misses the point, and pointed out that a phone's battery may not survive that long anyway. OnePlus offers four OS upgrades and six years of security patches. You can agree or disagree with the reasoning, but as a signal of internal resource allocation it was fairly clear even then.

The Pattern is Older Than OnePlus

None of this is novel. Motorola killed Motoblur in 2011 and went near stock, on the reasoning that Android itself had made heavy custom skins largely redundant. HTC let Sense fade away. LG's UX died with LG's phone business in 2021. Samsung's TouchWiz did not survive so much as get renamed and rebuilt into One UI once the company decided it needed to stop apologising for it.

The pattern in every case is the same, and it is not the one enthusiasts tell themselves. Skins do not die because users stop liking them. They die because the parent company needs efficiency and the differentiation has stopped paying for itself. OxygenOS is not being killed as punishment for anything. It is being closed as a line of business, in the same unremarkable way any other unprofitable line gets closed.

What OnePlus has Left to Sell

If the software is no longer the argument, the argument has to be hardware, performance, battery, camera tuning and price. That is a perfectly respectable list. It is also, word for word, the same list Oppo, Vivo, Realme and iQOO are working from, which returns us to the original problem. BBK built OnePlus to be the brand that was different. The company has now removed the thing that made it different, and it is fair to ask what the answer is meant to be when a buyer asks why this phone and not the one from the sibling brand on the next shelf.

The honest loss here is not OxygenOS, and the eulogies have that slightly wrong. It is that the variance is going. What made 2014 to 2018 an interesting time to review phones was that Android genuinely felt different depending on whose logo was on the back, and that variance was a subsidy paid by companies that could afford to be inefficient about software. The AI era has made inefficiency expensive. Almost nobody can afford it now, and the ones who still can are not doing it for our benefit either.

That is the actual story of July 2026. Which basically proves that business leaves no room for sentiment, and it's a call based on the balance sheet.