Note

Malaysia's SMEs Have RM442 Billion in Financing on Tap. Only About RM203 Million of It Is Earmarked for AI.

Outstanding SME financing in Malaysia hit RM442 billion as of May 2026 — growing 5.3% a year, with close to 80% of applications approved, according to Bank Negara Malaysia (BNM), the country's central bank. Money, broadly, is flowing. Set against that, this year's two flagship government schemes aimed specifically at AI and digital adoption — the SME Digitalisation Matching Grant and the Malaysia Digital Acceleration Grant — together total around RM203 million, given out first-come-first-served until it runs out. That's roughly 0.05% of everything currently on loan to Malaysian SMEs, earmarked for the one capability almost everyone agrees is coming next.

That contrast looks like the obvious problem. It probably isn't the real one.

What SMEs actually say they're short of

A Xero-commissioned survey of Malaysian micro, small and medium enterprises this year found 81% have already implemented AI in some form, and 48% plan to expand their use of it within the next year. That's a lot of adoption happening well ahead of any grant cheque clearing. Asked what they need to use AI properly, though, 82% said more education, and only 56% said they actually understand the different ways AI could apply to their business. Nearly a third of those already using it have no governance policy in place at all.

Here's the detail worth sitting with: when the same survey asked SMEs to rank what kind of support they most wanted, financial support came in last — behind training and education (61%) and behind advisory and consulting help (50%). Malaysian SMEs aren't mainly saying they can't afford AI. They're saying they don't yet know what to do with it.

Why this is a different shape of gap than 1993

That's a meaningfully different problem than the one Petronas's Vendor Development Programme was built to solve thirty years ago. Back then, the barrier really was capital: a small vendor couldn't afford the certification, the lab, the equipment needed to qualify for oil and gas work, and financing structured against a real contract was what unlocked it. AI tools today are cheap, often free at the entry level, and SMEs are clearly willing to try them without waiting for anyone's permission. The scarce resource has shifted from money to judgment — knowing which use case is worth the effort, how to govern it, and how to tell a genuinely useful deployment from a novelty one.

BNM's own governor made a version of this point in August, explaining why viable SMEs still get turned down: not because they're bad businesses, but because "innovation-driven activities are underfinanced because they're less understood," his words, describing young or asset-light businesses that don't fit a traditional credit-assessment lens. The bank is now shifting from direct SME lending toward guarantee schemes — RM10 billion of it, through the Credit Guarantee Corporation — and toward assessing businesses on cash flow and transaction data rather than collateral alone. That's a bank admitting the old scorecard doesn't capture the new kind of business. It's a different admission than "there isn't enough money."

What the earlier story actually teaches here

The lesson from the VDP era was never simply "give SMEs money" — it was that financing only converted into real capability when it was matched to a contract that told the vendor exactly what to build toward, and to training built around the real technical requirement rather than something generic. Money without that matching function sat idle or got spent on the wrong thing.

The AI version of that matching function looks less like a bigger grant pot and more like large companies doing what Petronas did in 1993: telling their smaller suppliers specifically what AI-enabled capability they'll actually pay for, and helping them build toward it. A RM203 million government grant pool spread across thousands of SMEs, each guessing independently at what "AI adoption" should even mean for their business, is a much weaker lever than one anchor company telling a hundred vendors precisely which capability earns the next contract.

The question this leaves

If Malaysian SMEs are already adopting AI on their own, and money isn't what they say they're short of, the RM203 million question isn't whether it's big enough. It's whether it's aimed at the right leg of the problem at all — or whether the more useful use of that same effort is getting large anchor companies to do what financing alone can't: tell their smaller partners exactly what capability is worth building, and stand behind it with real demand.

Sources: Malaysia SME financing figures and BNM schemes, Fintech News Malaysia; BNM Governor on SMEs falling through financing cracks, New Straits Times; BNM Annual Report 2025, SME financing chapter; Xero survey on Malaysian MSME AI adoption, Malaysia SME; Budget 2026 SME digitalisation and AI grant figures, GreatRise IT.

← All notes