The timing was no coincidence. Just six days earlier, on June 8, the UK published its AI Hardware Plan, backed by over £1.1 billion in public and private investment, with a core mission to build globally competitive AI hardware companies and maximise the adoption of AI semiconductors . The Rapidus MoU provides the critical manufacturing piece for that plan, offering UK-based fabless chip designers—from AI startups to established firms—a tangible path to produce their designs on a leading-edge process that is not controlled by TSMC or Samsung .

A £18 Billion Backdrop: The Broader UK-Japan Partnership

The semiconductor agreement was one plank in a far larger diplomatic and economic framework. On the same day, Prime Ministers Starmer and Takaichi sealed a comprehensive partnership expected to generate over £18 billion ($24 billion) in investment across AI, nuclear technology, defence, renewable energy, and financial services . The Rapidus-UKSC MoU sits within a bilateral frontier technology cooperation agreement, reinforcing a January 2026 commitment by both leaders to deepen technology collaboration .

The deal also serves geopolitical economic security goals. A parallel joint declaration focused on protecting stable energy and mineral supplies, directly expressing concerns over China's export restrictions on critical minerals and broader supply-chain instability . For the UK, securing a friendly-nation source of advanced chips reduces reliance on a concentrated set of East Asian suppliers.

How the MoU Activates the UK's £1.1 Billion AI Hardware Plan

The UK's AI Hardware Plan is structured around four pillars: innovation, skills, procurement, and investment . Its headline investments include £750 million for a new national AI supercomputer and up to £150 million from the British Business Bank for an early-stage hardware fund led by Playground Global . But a recurring weakness in UK semiconductor policy has been the "design-to-fab" gap. Britain excels at chip architecture—ARM's origins are the most famous example—but lacks any domestic facility capable of fabricating chips at sub-7nm process nodes.

The MoU with Rapidus directly addresses this. UKSC officials characterised the collaboration as a mechanism to allow UK startups and established enterprises to "directly tap into Rapidus production as it comes online" . The deal aligns with the AI Hardware Plan's stated goal of "maximising the adoption of AI semiconductors" by ensuring British-designed chips can actually be manufactured at a cutting-edge node . It also anchors international investment in the UK semiconductor ecosystem, fulfilling another key objective of the plan .

The government has also committed £400 million to purchase next-generation AI chips, with £150 million ring-fenced as an advance commitment to buy chips from innovative British startups . Access to a 2nm foundry partner gives those procurement commitments far more credibility, because the UK firms winning those contracts now have a production path that doesn't depend entirely on TSMC or Samsung.

Rapidus' 2nm Manufacturing: The Road to H2 2027

Rapidus is moving at an extraordinary pace for a company founded only in 2022 by a consortium including Toyota, Sony, NTT, SoftBank, and other Japanese industrial giants . Its timeline to date:

• April 2025 – Pilot production line (IIM-1 facility) began operation in Chitose, Hokkaido, with over 200 main manufacturing devices and 300 processes installed .
• July 2025 – Unveiled first 2nm GAA transistor test wafers, confirming expected electrical characteristics . This milestone was reached barely three months after air-freighting its first EUV lithography system from the Netherlands in December 2024 .
• August 2025 – Completed tape-out of 2nm GAA test chips .
• Q1 2026 – Released first 2nm Process Design Kits (PDK) to early-access customers, enabling fabless companies to begin designing chips tailored to the process .
• H2 2027 – Mass production target (second half of fiscal year 2027), starting at approximately 6,000 wafers per month and aiming to ramp to 25,000 wafers per month within roughly one year .

Rapidus is already planning a second Hokkaido fab targeting 1.4nm production by 2029, signalling that this is not a one-generation bet but a sustained push to become a permanent third force in advanced logic manufacturing .

The Global Customer Hunt: Why the UK Deal Matters for Rapidus

For all its technical momentum, Rapidus needs customers. The foundry has pursued a deliberate multi-continent strategy to build a client base before mass production begins:

• IBM remains the foundational partner. Since signing a strategic partnership in December 2022, IBM has licensed its 2nm technology to Rapidus and provides ongoing manufacturing expertise. IBM is planning a dedicated office in downtown Chitose to support the fab's operations .
• Tenstorrent, the AI chip startup led by legendary chip architect Jim Keller, has been named alongside IBM as a lead prospective customer .
• Cadence is collaborating on AI-driven reference design flows and an IP portfolio for the 2nm GAA process, making the platform more accessible to fabless designers .
• Silicon Valley subsidiary – Rapidus operates a US subsidiary specifically to recruit fabless clients among big tech firms and AI startups .
• European push – Alongside the UK MoU, Rapidus executives briefed PM Takaichi on parallel R&D cooperation agreements with Italian institutions, signalling a broader European strategy .

The UKSC MoU gives Rapidus a government-backed entry point into a specific national ecosystem of AI chip designers who are being actively funded and incentivised to produce hardware. For a foundry trying to break into a market dominated by TSMC and Samsung, these sovereign-government partnerships are a critical customer acquisition channel.

The UK Semiconductor Centre itself was created precisely for this moment: to act as a bridge between Britain's design talent and global manufacturing capacity. The MoU is the first major test of whether that model can produce real commercial outcomes for British chip firms. With Rapidus targeting mass production in approximately 18 months, the runway for UK designers to prepare 2nm tape-outs starts now.

The timing was no coincidence. Just six days earlier, on June 8, the UK published its AI Hardware Plan, backed by over £1.1 billion in public and private investment, with a core mission to build globally competitive AI hardware companies and maximise the adoption of AI semiconductors . The Rapidus MoU provides the critical manufacturing piece for that plan, offering UK-based fabless chip designers—from AI startups to established firms—a tangible path to produce their designs on a leading-edge process that is not controlled by TSMC or Samsung .

A £18 Billion Backdrop: The Broader UK-Japan Partnership

The semiconductor agreement was one plank in a far larger diplomatic and economic framework. On the same day, Prime Ministers Starmer and Takaichi sealed a comprehensive partnership expected to generate over £18 billion ($24 billion) in investment across AI, nuclear technology, defence, renewable energy, and financial services . The Rapidus-UKSC MoU sits within a bilateral frontier technology cooperation agreement, reinforcing a January 2026 commitment by both leaders to deepen technology collaboration .

The deal also serves geopolitical economic security goals. A parallel joint declaration focused on protecting stable energy and mineral supplies, directly expressing concerns over China's export restrictions on critical minerals and broader supply-chain instability . For the UK, securing a friendly-nation source of advanced chips reduces reliance on a concentrated set of East Asian suppliers.

How the MoU Activates the UK's £1.1 Billion AI Hardware Plan

The UK's AI Hardware Plan is structured around four pillars: innovation, skills, procurement, and investment . Its headline investments include £750 million for a new national AI supercomputer and up to £150 million from the British Business Bank for an early-stage hardware fund led by Playground Global . But a recurring weakness in UK semiconductor policy has been the "design-to-fab" gap. Britain excels at chip architecture—ARM's origins are the most famous example—but lacks any domestic facility capable of fabricating chips at sub-7nm process nodes.

The MoU with Rapidus directly addresses this. UKSC officials characterised the collaboration as a mechanism to allow UK startups and established enterprises to "directly tap into Rapidus production as it comes online" . The deal aligns with the AI Hardware Plan's stated goal of "maximising the adoption of AI semiconductors" by ensuring British-designed chips can actually be manufactured at a cutting-edge node . It also anchors international investment in the UK semiconductor ecosystem, fulfilling another key objective of the plan .

The government has also committed £400 million to purchase next-generation AI chips, with £150 million ring-fenced as an advance commitment to buy chips from innovative British startups . Access to a 2nm foundry partner gives those procurement commitments far more credibility, because the UK firms winning those contracts now have a production path that doesn't depend entirely on TSMC or Samsung.

Rapidus' 2nm Manufacturing: The Road to H2 2027

Rapidus is moving at an extraordinary pace for a company founded only in 2022 by a consortium including Toyota, Sony, NTT, SoftBank, and other Japanese industrial giants . Its timeline to date:

• April 2025 – Pilot production line (IIM-1 facility) began operation in Chitose, Hokkaido, with over 200 main manufacturing devices and 300 processes installed .
• July 2025 – Unveiled first 2nm GAA transistor test wafers, confirming expected electrical characteristics . This milestone was reached barely three months after air-freighting its first EUV lithography system from the Netherlands in December 2024 .
• August 2025 – Completed tape-out of 2nm GAA test chips .
• Q1 2026 – Released first 2nm Process Design Kits (PDK) to early-access customers, enabling fabless companies to begin designing chips tailored to the process .
• H2 2027 – Mass production target (second half of fiscal year 2027), starting at approximately 6,000 wafers per month and aiming to ramp to 25,000 wafers per month within roughly one year .

Rapidus is already planning a second Hokkaido fab targeting 1.4nm production by 2029, signalling that this is not a one-generation bet but a sustained push to become a permanent third force in advanced logic manufacturing .

The Global Customer Hunt: Why the UK Deal Matters for Rapidus

For all its technical momentum, Rapidus needs customers. The foundry has pursued a deliberate multi-continent strategy to build a client base before mass production begins:

• IBM remains the foundational partner. Since signing a strategic partnership in December 2022, IBM has licensed its 2nm technology to Rapidus and provides ongoing manufacturing expertise. IBM is planning a dedicated office in downtown Chitose to support the fab's operations .
• Tenstorrent, the AI chip startup led by legendary chip architect Jim Keller, has been named alongside IBM as a lead prospective customer .
• Cadence is collaborating on AI-driven reference design flows and an IP portfolio for the 2nm GAA process, making the platform more accessible to fabless designers .
• Silicon Valley subsidiary – Rapidus operates a US subsidiary specifically to recruit fabless clients among big tech firms and AI startups .
• European push – Alongside the UK MoU, Rapidus executives briefed PM Takaichi on parallel R&D cooperation agreements with Italian institutions, signalling a broader European strategy .

The UKSC MoU gives Rapidus a government-backed entry point into a specific national ecosystem of AI chip designers who are being actively funded and incentivised to produce hardware. For a foundry trying to break into a market dominated by TSMC and Samsung, these sovereign-government partnerships are a critical customer acquisition channel.

The UK Semiconductor Centre itself was created precisely for this moment: to act as a bridge between Britain's design talent and global manufacturing capacity. The MoU is the first major test of whether that model can produce real commercial outcomes for British chip firms. With Rapidus targeting mass production in approximately 18 months, the runway for UK designers to prepare 2nm tape-outs starts now.