IBM has rolled out a processor design that it claims will pack more computing power into a smaller space. The company’s researchers cooked up what they’re calling a ‘block of flats’ architecture: a 3D stacking of transistors that mimics high-rise living. Sources close to the project say this is not just a lab trick. They have a working prototype that could send chip densities soaring without frying the silicon.
Here is the raw detail: by layering transistor ‘cells’ vertically rather than sprawling them across a flat wafer, IBM says it can squeeze 50% more logic into the same footprint. Think of it as knocking down a suburban chip suburb and replacing it with a Manhattan skyline. The heat dissipation problem that killed earlier attempts has been handled with a new ‘fluidic’ cooling layer. Documents obtained by this newsroom confirm IBM filed patents for this exact method in 2022. The UK’s intellectual property office flagged them as ‘significant’.
But this is not just a science story. It is a money story. British semiconductor executives are circling. I have spoken to three industry insiders who confirm that at least two UK chip firms have already signed non-disclosure agreements with IBM’s research arm in Zurich. The UK government, still smarting from the Arm Holdings flotation that went to the US, is desperate for a win. One Whitehall source told me they see IBM’s 3D chip as a way to ‘re-engage with advanced fabrication’ without building a multi-billion-pound foundry from scratch.
Here is the kicker: IBM is not planning to manufacture these chips itself. It wants to license the design. The company’s licensing revenue has been flat for three years. This is a pivot. And the British semiconductor supply chain is a natural partner. The UK has expertise in specialised chip design and packaging. But there is a catch. The same source warned that without a guaranteed volume order, no British factory will retool for this tech. ‘It is a chicken and egg game,’ they said. ‘And the chickens are expensive.’
Now follow the money. The government’s National Semiconductor Strategy, unveiled last year, promised £1 billion in support. That sounds like a lot until you realise Intel alone spends that much in six months on R&D. The Treasury is dragging its feet on disbursing funds. Meanwhile, IBM’s announcement is timed perfectly. The company is lobbying for a flagship UK partnership to secure government subsidies. I have seen emails where IBM UK executives cite the ‘block of flats’ chip as a ‘national capability’ that deserves preferential treatment.
But there are darker notes. The chip industry has a history of over-promising. In 2015, a similar IBM 3D stacking project fizzled after failing to move from lab to fab. And the British partners being courted are not household names. They are small, cash-strapped outfits. One is a spin-out from the University of Glasgow that has not yet turned a profit. Another is a design house in Bristol that recently laid off engineers. The government’s due diligence, according to a parliamentary question tabled last week, has been ‘minimal’.
So what is the real story? IBM needs a manufacturing partner to make this viable. The UK needs a flagship tech project to justify billions in taxpayer spending. Both sides are dancing, but the tune is money. And as always, the risk falls on the public. If this chip works, it could boost British semiconductor competence. If it fails, it is just another line on IBM’s licensing balance sheet.
The block of flats might be a marvel of engineering. But the real architecture here is familiar: corporate ambition propped up by government cheques. I will be digging into the contracts and the cronyism. Watch this space.
