Malaysia Sets 2030 for Its First Arm-Based Chips to Reach Production

On 21 August, Economy Minister Akmal Nasrullah Mohd Nasir said Malaysia aims to have chips designed locally under its Arm partnership in production by 2030. He also named two more companies with access to Arm technology: Oppstar Technology and Alphaswift Industries. The New Straits Times reported that eight technology access tokens have now been awarded. Four are Compute Subsystem (CSS) tokens and four are under Arm's Flexible Access programme, supporting designs for uses that include automotive automation and servers.

The minister set out the timeline. The first token was granted at the end of 2025. Design takes about three years, so the first designs should be finished in 2028. Production and commercialisation are targeted for 2030, after testing and market validation.

What the programme has delivered so far

The deal dates from March 2025, when Prime Minister Anwar Ibrahim launched a collaboration under which Malaysia would pay Arm US$250m over ten years for IP licensing, technology transfer and training, and Arm would open its first ASEAN office in Kuala Lumpur. MIDA's announcement at the time set an aspiration of about ten Malaysian chip companies with combined annual revenue of US$20bn.

Eighteen months later, The Vibes reports that five Malaysian companies hold the eight tokens. Around 1,530 people have completed Arm training, against a target of 10,000 over four years, and an agreement with 12 universities will put semiconductor training into their curricula. Token recipients must hire highly skilled Malaysian engineers as a condition of access.

The training numbers are the most honest measure of progress, and they are behind. At about 1,500 trained against a target of 10,000, the pipeline needs to speed up considerably. The minister appears to recognise this: he framed the challenge as moving from reacting to talent shortages to building a structured pipeline.

The hard part is not the licence

Access to Arm IP lowers the barrier to starting a design. It does not solve the three things that decide whether a chip company survives.

Customers. A server or automotive system-on-chip needs a buyer who will commit before tape-out. Neither announcement named one. Automotive qualification cycles are long, and the server market is dominated by hyperscalers that increasingly design their own Arm silicon. A Malaysian design has to be either cheaper, more specialised or politically preferred to win a slot.

Manufacturing. The announcement speaks of production, but leading-edge wafers for a CSS-class design would almost certainly come from a foreign foundry. Neither report says which foundry, at what node, or who pays for the masks. Those costs, not the licence fee, dominate a first chip's budget.

Time. A 2030 production date means first revenue arrives five years after the deal was signed and four years into a ten-year payment schedule. Royalties to Arm only start once chips sell.

The governance overhang

There is also unfinished business. In February, the Malaysian Anti-Corruption Commission confirmed it was reviewing complaints from three NGOs over alleged mismanagement linked to the RM1.11bn commitment. At the time, the minister said the government would cooperate fully and that the programme would continue. Neither August report gave an update on that review. A clean result would help the programme's credibility. An unresolved one hangs over every new token award.

What I would like to see published: which of the five companies is furthest along, whether any has taped out a test chip, and a named launch customer for at least one design. Until then, 2030 is a target, not a roadmap.

---

Sources

• New Straits Times: Malaysia targets Arm chip production by 2030, Oppstar & Alphaswift get access to tech
• The Vibes: Two local firms get access to Arm technology in push for homegrown chip design
• MIDA: Malaysia to invest US$250m over next decade via strategic collaboration with Arm Holdings
• The Sun: Malaysia seeks to continue Arm Holdings deal despite MACC probe