Samsung is turning back to an old playbook, reinstating a dual-CEO structure as it tries to steer through one of the most turbulent periods in the history of the chip and smartphone industries. Instead of concentrating power in a single chief executive, the Korean giant is elevating the leaders of its two most critical businesses – mobile devices and memory chips – to co-CEO status. 
The strategy is simple on paper: put the decision-makers who live closest to fast-moving markets directly at the top, and hope that sharper focus translates into quicker responses to AI demand shocks, supply bottlenecks and brutal pricing pressures in flagship phones.
The decision, reported by Korean business media as part of Samsung’s annual leadership reshuffle, revives a split hierarchy the company has used in previous eras of rapid change. This time, the stakes are different. Artificial intelligence infrastructure is devouring high-bandwidth memory (HBM), the smartphone market is slowly recovering but far from stable, and competitors in both the US and China are racing to lock in supply of advanced chips. In that context, Samsung’s board appears to have concluded that one centralised CEO is not enough to keep pace.
Why Samsung is reviving the co-CEO model
Under the new structure, President Roh Tae-moon – better known globally as TM Roh – becomes the official chief of the smartphone-focused Device eXperience (DX) division and one of Samsung’s two co-CEOs. Roh has effectively been running the mobile and consumer electronics businesses for months, overseeing everything from Galaxy S flagships to foldables and wearables. His promotion formalises that role and signals how central the Galaxy ecosystem remains to Samsung’s brand power and cash flow.
Sharing the top job is Vice Chair Jun Young-hyun, who leads the Device Solutions (DS) division. This is the unit that makes memory chips, storage and other components that power everything from data-center servers to smartphones. Memory has historically been Samsung’s profit engine, and in the coming AI era it is once again the battlefield where fortunes are won and lost. By giving the DS boss equal footing with the head of DX, Samsung is acknowledging that chips and consumer devices must be managed as two halves of a single strategic story rather than as separate empires.
A Harvard scientist takes over Samsung’s blue-sky R&D
The reshuffle doesn’t stop at the C-suite. Jun Young-hyun will hand over leadership of the Samsung Advanced Institute of Technology (SAIT), the company’s central research lab, to Park Hong-kun, a Harvard University professor specialising in nanoscience and quantum technologies. SAIT is not about next quarter’s results; it is where Samsung incubates the technologies that could define its business five or ten years from now.
Bringing in an external academic heavyweight to run SAIT is a bold move by Samsung’s standards, which have traditionally favoured promotion from within. Park’s appointment hints at how seriously the company is taking foundational research in areas like quantum devices, advanced materials and unconventional chip architectures. It also shows Samsung’s willingness to blend corporate engineering culture with a more exploratory, university-style approach to discovery.
Neuromorphic chips: imitating the brain to power future AI
One of Park’s headline missions will be to drive Samsung’s research in neuromorphic semiconductors. Unlike traditional chips, which separate computation and storage, neuromorphic designs aim to mimic the way the human brain works. They distribute tiny processing elements and memory cells side by side, allowing massive amounts of information to be processed in parallel while keeping data where it is needed rather than shuttling it back and forth.
This brain-like architecture can deliver dramatic efficiency gains. For AI workloads that need to run locally – on phones, wearables, cars or edge devices – neuromorphic chips promise lower power consumption, less heat and more responsive on-device intelligence. If Samsung can turn SAIT’s experiments into commercially viable products, it would gain a powerful differentiator at a time when every tech giant is trying to squeeze more AI into smaller, battery-powered devices.
AI demand, HBM hunger and the LPDDR5 price squeeze
The leadership overhaul is also a reaction to harsh realities in today’s memory market. Explosive demand for HBM, the stacked high-bandwidth memory used in AI accelerators and data-center GPUs, is soaking up manufacturing capacity that could otherwise serve conventional DRAM products. As a result, mobile-oriented memory such as LPDDR5 is becoming more expensive, just as smartphone makers are trying to keep device prices under control.
Samsung sits uncomfortably at the intersection of these forces. On one side, its DS division wants to prioritise high-margin HBM to feed the AI boom. On the other, its DX division needs affordable LPDDR5 to keep Galaxy devices competitive. The co-CEO model effectively forces both sides of the business to negotiate in real time at the highest level, rather than fighting over capacity from within separate silos.
Galaxy S26 Ultra and the cost of going ultra-premium
Layered on top of memory inflation is the rising cost of cutting-edge application processors. Qualcomm’s Snapdragon 8 Elite Gen 5 system-on-chip is expected to power the upcoming Galaxy S26 Ultra and, according to industry chatter, will do so on an exclusive basis. That gives Samsung access to one of the fastest Android chipsets on the market, but it also means swallowing what is widely described as a very expensive component in the bill of materials.
Between pricier LPDDR5, premium HBM upstream and a flagship Qualcomm platform, Samsung’s mobile unit faces intense margin pressure just as it tries to differentiate its top-end Galaxy phones with AI features and camera upgrades. Agility in this context is not a buzzword; it is the ability to juggle supplier contracts, product positioning and internal cost allocations fast enough to react before a generation of devices is locked in.
Can two CEOs really make Samsung nimbler?
Critics of co-CEO structures often warn about blurred accountability and slower decisions, not faster ones. Samsung’s version is an attempt to sidestep that trap by giving each leader a clearly defined empire – DX for TM Roh, DS for Jun Young-hyun – while still expecting them to collaborate on issues where mobile and memory are inseparable. If the model works, Samsung could better align chip roadmaps with device launches, deploy AI features that fully exploit its own silicon, and respond more quickly to shifts in demand for servers, PCs or smartphones.
If it fails, the company risks mixed signals to investors, internal turf wars and a muddled long-term vision. For now, though, the message from Seoul is that the era of one-size-fits-all leadership is over. With AI reshaping the economics of both data centers and pocket-sized devices, Samsung is betting that pairing a mobile veteran with a memory specialist – and backing them with a research chief obsessed with brain-like chips – is the best way to stay relevant in the next wave of computing.