Build or Rent: Should Malaysia Make Its Own AI or Spend on Using It?

Mei Ling Good afternoon. On the twelfth of June, an American AI company switched off its two newest models. For every customer, everywhere, in one evening. Washington had barred foreign nationals from using them, and the company couldn't sort its users by passport. One model came back in July. So today's question is an old one with a new edge. Should Malaysia build its own AI, or keep renting someone else's?

Raj So the landlord can change the locks whenever he likes. Lovely thing to learn on a Friday.

Mei Ling That's the fear. Whether it justifies billions of ringgit is what we test today. It's Friday the twenty-third of October, and this is the AI-YA Daily Friday special. It runs longer than our weekday shows, and it's slower and more reflective.

Raj I'm Raj. I run a small business, I ask the awkward questions, and yes, I still print my invoices.

Aiman I'm Aiman. I try the tools and do the sums, usually in ringgit.

Mei Ling And I'm Mei Ling. I read the fine print, so you can get on with your day.

Aiman What we're here for is simple. We take the world's artificial intelligence news and make it useful for people working in Malaysia, with the price tag attached.

Mei Ling This show is researched and voiced with AI, and edited by the AI-YA team. We practise what we teach.

Raj This week's theme is build or rent. Should Malaysia make its own AI, or rent the best and spend on using it well?

Mei Ling Three stories, one Malaysia. Right, let's get going. Let's give the build side its strongest case first. It rests on three words: language, data and control. Start with language. In February 2024, at a summit in Dubai, Nvidia's chief executive Jensen Huang advised developing countries to build. In his words, "codify the language, the data of your culture into your own large language model."

Raj Of course the man who sells the chips says everyone should build. That's like my tile supplier telling me every room needs new tiles.

Mei Ling Fair. He has a commercial interest, and he's never hidden it. But YTL AI Labs' chief executive, Foong Chee Mun, made a similar point to The Edge in June. He said global models "tend to deviate" in minor languages.

Raj Hmm. My customers write in three languages in one WhatsApp message. If a model can't follow that, it's no use to me.

Mei Ling That's the language case. The second is data. Last week, Communications Minister Fahmi Fadzil talked about the sovereign AI cloud. He said it's meant to keep classified government information out of reach of foreign laws. The third is control, and that became very real in June. On the twelfth, America's Commerce Department ordered Anthropic to bar foreign nationals from its two newest models. Anthropic said that left it no choice but to switch both off for everyone.

Raj Hang on, say that again. Off for everyone?

Mei Ling For all its customers, according to Anthropic's statement as reported by Al Jazeera. Fable 5 came back globally on the first of July. Mythos 5, the more powerful one, returned only for selected American organisations. We couldn't find a later update on that one.

Aiman Picture a firm in Penang that had built its daily work around Fable. That's about nineteen days without the tool. Nobody in Malaysia was consulted.

Mei Ling That's the build argument in one picture. In India, Sarvam's co-founder Pratyush Kumar put it in one line on X. "You should not confuse access with ownership, or adoption itself as advantage."

Raj Okay, that's a strong line. Access isn't ownership. I'll remember that.

Mei Ling Sarvam was picked in 2025 to build a national model under India's IndiaAI Mission. In June, Medianama reported it had raised 234 million US dollars, about RM955 million. Then in May, Mistral's chief executive Arthur Mensch told French lawmakers that Europe has about two years to act. He said, "In a world where you import all digital services from the United States, you have no leverage against the US."

Raj I'm not convinced. India has more than a billion people, and Europe has deep pockets. We're thirty-something million. Why would we copy them?

Mei Ling Okay, but let me finish my thought. The strongest case for Malaysia isn't a frontier model. In February, David Lam of the ISEAS-Yusof Ishak Institute in Singapore argued that sovereign models needn't beat the giants on everything. He said the region's edge is closeness to local customers, local data and regional cooperation. That's a niche, not a moonshot. And Malaysia already has pieces, which we went through on Thursday. YTL builds ILMU on foundations from Nvidia and the Chinese lab Z.ai. Mesolitica built MaLLaM from scratch, though its public releases have slowed. The government is pooling Malay-language data at Dewan Bahasa dan Pustaka, or DBP. And the Malaysia Digital Economy Corporation, or MDEC, plans a national compute exchange from early 2027.

Raj So we're already building. Just quietly, and on borrowed engines.

Mei Ling Yes. Which is why the honest question isn't really build or rent. It's which layers we must own ourselves.

Raj Fair, I hadn't split it into layers. I still want to see the bill.

Aiman You'll get it in the third story. Mei Ling, in Episode 4 you weren't sold on IlmuChat Pro. Does that still stand?

Mei Ling It stands, because there's still no newer usage figure. And yesterday's three different votes on ILMU's Flash model, at RM0.55 per million input tokens, stand too.

Aiman So what does each path mean for Malaysian jobs? I'd split it three ways: build the models, build the buildings, or build the users. Path one is the models. That employs a small number of very skilled people. Researchers, engineers, and people who clean and label data. Korea shows the catch. Its national foundation model project has spent over 200 billion won so far, about RM600 million by our conversion. This month, the Korea JoongAng Daily reported that developers are leaving. One moved from Motif to Kakao in August. A key SK Telecom team member reportedly left for Nvidia in September.

Raj So the government pays to train them, and then the big boys come along and…

Aiman And hire them. Yes.

Raj I know that feeling. I train a good supervisor, and the shop across the road poaches her.

Aiman Same story, more zeros. One company source also told the paper that the government's chips were "a drop in the bucket" next to overseas rivals.

Raj Okay. Path two?

Aiman The buildings, meaning data centres. On the fourteenth, the Digital Ministry gave a written reply in Parliament. Twenty-seven approved projects, about RM226 billion in estimated investment, and 4,053 expected jobs.

Raj Wait. Two hundred and twenty-six billion ringgit, for four thousand jobs?

Aiman Our arithmetic, not theirs, works out to roughly RM56 million of investment per expected job. The ministry says 3,523 of those jobs will be high-skilled. It expects Malaysians to fill about 89 per cent of them.

Mei Ling Those are expectations, not outcomes. Only one of the twenty-seven is operating, and the rest are still under construction. We covered the power and water side on Wednesday, so I won't repeat it.

Aiman There is a skills route, though. TalentCorp is developing a data centre facilities management programme. And every approved project has a minimum local content rule of thirty per cent.

Raj And path three?

Aiman The users, which is everyone else. In June, Microsoft's Work Trend Index surveyed 2,000 Malaysian knowledge workers. It found 24 per cent were what it calls frontier professionals, the most advanced AI users. The global figure was 16 per cent.

Raj So we're ahead of the world?

Aiman Among the office workers surveyed, yes. But only 32 per cent said their leadership was clearly aligned on AI. And just 19 per cent said they're rewarded for trying new ways of working.

Raj That's my shop exactly. My staff are ahead of me, and I'm the one with my foot on the brake.

Aiman Which is why we keep saying the same thing on Fridays. Allow AI, and manage it. Banning it just pushes it onto personal phones.

Raj I learned that from my supervisor in Episode 7. She'd been using it on her own phone for a year.

Aiman The government's own target sits mostly on this path. The National AI Action Plan aims for 300,000 to 500,000 new AI-related jobs by 2030. The fact sheet doesn't say how many are builders and how many are users.

Mei Ling There's a fourth, smaller path people forget. Budget 2027 asks ministries, agencies and universities to give DBP verified and annotated language data. Annotation is real work. Linguists, teachers and translators could be paid to do it, though no ringgit figure has been announced.

Raj Now the serious part. Some people's work will change whichever way we go.

Aiman Yes. Translation, transcription and customer service in Malay all get cheaper, whether the model is local or foreign. The worker feels the same change either way. In Episode 5, the Khazanah Research Institute said AI wasn't yet behind Malaysia's job losses. That can change. If you lose your job, register with PERKESO, the Social Security Organisation, under the Employment Insurance System. Since the first of October, its training fee cap is RM7,000. Apply through the LINDUNG Kerjaya portal or at a PERKESO counter.

Raj And for employers like me?

Aiman Use your HRD Corp levy to retrain people before you restructure. In our view, retraining someone you already trust is worth trying first. Now the sceptics. The first is a company that actually tried. Germany's Aleph Alpha raised 500 million US dollars in November 2023, about RM2 billion today. Ten months later, it pivoted to helping firms use other people's models. Its chief executive, Jonas Andrulis, told Bloomberg, "Just having an European LLM is not sufficient as a business model."

Mei Ling Korea is the live case. It wanted models at 95 per cent of frontier capability by 2027. On the thirtieth of September, the Korea Times reported the science ministry was reviewing whether to hold the final round at all. Minister Bae Kyung-hoon says the project will continue, and the government has proposed 4.7 trillion won for frontier models next year. That's roughly RM14 billion by our conversion.

Aiman And that's the core problem. The frontier moves faster than any national budget cycle. In June, the research group Epoch AI said the best open-weight models had trailed the closed frontier by about four months since January. Open-weight means anyone can download the model and run it on their own machines.

Raj Hmm. So by the time we finish building ours, the free one has caught up?

Aiman That's the risk. And the money gap is enormous. Epoch projects the big cloud companies will spend around 770 billion US dollars this year. That's about RM3.1 trillion.

Mei Ling For scale, the whole of Budget 2027 is RM510 billion.

Raj So six Malaysian budgets. In one year. On computers.

Mei Ling History has a pattern here. Japan's Fifth Generation computer project ran for ten years and cost about 400 million US dollars. When it was wound up in 1992, Tech Monitor reported most firms had built only one-off prototypes. Then there's Europe's Gaia-X cloud project. In 2022, Forrester analysts wrote that it had "failed to launch meaningful public cloud and data services." They said it was busy "building bureaucracies and PowerPoint presentations."

Raj I've sat in that meeting. Twice.

Aiman And closer to home, there's Proton. The year before, it had received RM1.5 billion in government aid, on condition it found a foreign partner. In 2017, Geely took 49.9 per cent.

Mei Ling I'd push back on using Proton as a failure story. The minister at the signing said Proton would remain the national carmaker. It kept its brand, and it gained Geely's platforms and engines. That's not giving up. That's building on someone else's engine and keeping the badge.

Aiman Fair, I hadn't framed it that way. That's close to what YTL is doing with ILMU.

Mei Ling So here's the split judgement. History warns against funding a national frontier model. It doesn't warn against owning the layers on top.

Aiman And for small firms, the counter-case is mostly arithmetic. On Thursday, our illustrative sum put a thousand customer chats on ILMU's Flash model at about RM2 a month. An SME, a small or medium enterprise, doesn't need a national model. It needs one that handles Malay and keeps data where the law expects.

Raj And somebody to set it up. That's the bill nobody prints, and I print everything.

Aiman True. MDEC's RM30 million for 4,000 firms works out at RM7,500 each, at most. We did that sum in Episode 8. As of the seventeenth, the last time we checked, its rules still weren't published. Our sponsor takes the next sixty seconds. When we return, we hear how the world's podcasters talk about countries like ours.

Sponsor Most AI projects stall in the same place. Someone runs a pilot, everyone agrees it's promising, and then nobody owns it. Hitower IT does the other part. AI discovery, a roadmap scoped to your budget, and the actual build, shipped into the workflows your team already uses. Strategy to code, no handoffs. Dutch precision, Malaysian know-how. A team you can meet in Kuala Lumpur. Book a free discovery call at hitowerit.com.my. They come back within one working day with where automation can help and what it's worth.

Mei Ling Welcome back. I've spent the week listening to other people's podcasts, so you didn't have to.

Raj And reading page forty of their show notes, I assume.

Mei Ling Page forty of the transcripts, actually. Five shows. Let me take them in turn. First, the Big Technology Podcast on the twenty-second of June, recorded live. The guest was Aaron Levie, chief executive of Box. Days after the Fable shutdown, he said other countries "now have far more incentive to stand up their own sovereign AI initiatives." He also suggested the outcome might be "net negative" for America's position in AI. In the transcript, the host's view on the net winner was "probably China."

Raj So the ban did Jensen Huang's selling for him.

Aiman The chip sellers gain either way. Everyone who builds needs chips.

Mei Ling Second, the 80,000 Hours podcast in July. The guest was Anton Leicht, a fellow at the Carnegie Endowment for International Peace. His subject was middle powers, which includes countries like us. His warning is blunt. Countries without frontier models, he said, "capture all the risk" as these models spread through society. But his preferred answer isn't everyone building alone. It's middle powers doing what they're good at in the supply chain, and combining that with what America does well. He did price a fallback, though. A frontier project by a group of allies would cost roughly 500 billion US dollars over four years, by his estimate.

Raj That's about two trillion ringgit. My accountant just fainted, and he isn't even listening.

Mei Ling Third, Bloomberg's new series, Emerging. Its first episode, on the ninth of July, was called "Can India Win the AI Race Without Its Own OpenAI?" Menaka Doshi hosted, with Srikanth Velamakanni of Fractal Analytics and Professor Ganesh Ramakrishnan of the Indian Institute of Technology Bombay. The show calls it India's AI trilemma. Can a country be sovereign without controlling the full stack? I'm only giving you the framing there, because the show page I read doesn't summarise the guests' answers. Fourth, the Lex Fridman Podcast in March, with Jensen Huang again. This time he argued for open models. He said Nvidia releases them so "every country" can join the AI revolution. It has published the weights and data for its Nemotron models.

Aiman And that matters here. YTL's Ilmu-Nemo-30B, launched in March, was built with Nvidia's Nemotron open models. So the open route is already our route.

Mei Ling Fifth, an older one. The venture firm a16z ran an episode in May 2025 on why nations build their own models. Its description said AI "isn't just a technology anymore – it's geopolitical infrastructure."

Raj Okay. Five clever shows. What do they miss about us?

Mei Ling Three things, I think. First, they talk about countries as if they're governments and giant labs. Malaysia is mostly small firms, and MDEC's adoption money is spread across 4,000 of them.

Aiman Second, they assume one language per country. Our customers mix Malay, English, Chinese dialects and Tamil in a single message. Nobody in San Francisco is testing for that.

Raj My aunty mixes four languages and two emojis in one sentence. No model has passed that test yet, lah.

Mei Ling Third, we're an export economy sitting between two giants. In Episode 12 we noted that nearly half of Malaysia's August exports were electronics. We sell into America's AI boom, while YTL builds on a Chinese base. A San Francisco podcast asks how America stays ahead. We have to ask how we keep both doors open.

Aiman And there's a fourth. We host the data centres. Those shows treat compute as something you buy. We have the buildings, the power demand and the water questions.

Raj I hear what you're saying, but who's listening to us? We're not on Lex Fridman.

Mei Ling Not yet. But the ISEAS argument is one those shows can't make. Our edge is being close to local customers and local data.

Raj So our edge is knowing what a kedai runcit in Kota Bharu actually needs. That's not nothing.

Mei Ling It's not nothing. It may be the whole point. Now, the Friday positions. First, this week's Big Questions, one line each.

Aiman Monday: do we train graduates on AI to do senior work sooner, or protect the junior roles where people learn?

Raj Tuesday: should public money first help the smallest firms buy tools that keep data safe, or back the firms ready to automate?

Mei Ling Wednesday: do you pick the cheapest capable model whatever its origin, or pay more for one that sits where your data sits?

Aiman And Thursday: should Malaysia keep paying to build national models, or rent the best and fund adoption?

Mei Ling So here's the question for you to answer for yourself. If Malaysia's AI money can only go one way first, should it build the engine, or train the drivers?

Aiman My position is train the drivers first. Rent the strongest models, and keep an open-weight backup running here. The best open models trail the frontier by about four months. Jobs will be won or lost on adoption.

Raj Mine is close to Aiman's. Spend on people and small firms first. But I want one thing owned here, and that's Malay that works for real customers. If DBP's data makes every model better at our languages, I'm happy for my taxes to pay for it.

Mei Ling I'll take a narrower build position. Don't fund a national frontier model. Aleph Alpha and the Fifth Generation project show how that can go, and Korea is wrestling with it now. But own three layers. Our language data, the testing of models in our languages, and somewhere here to run them if access is cut. June showed that can happen in one evening.

Aiman So we differ on emphasis, not direction.

Mei Ling Yes. And the direction is already set. The National AI Action Plan for 2026 to 2030 has 28 national initiatives. It targets a top-ten place in global AI readiness by 2030. AI Malaysia Berhad, the national AI office under the Digital Ministry, leads it. AI Untuk Rakyat aims to give 100,000 young Malaysians free AI tools after six short modules. Budget 2027 promised 100,000 more places. HRD Corp keeps RM3 billion for training, a figure restated from last year. And The Vibes reported RM2 billion for the sovereign AI cloud in Budget 2026. But direction isn't delivery. Every one of these sits somewhere between announcement, allocation and disbursement. For MDEC's fund, we're still waiting for the rules. For the compute exchange, we're still waiting for prices.

Aiman Which is why the useful part is what you do next. If you're a worker, learn to check AI's work in your own language. Your judgement is the part that stays valuable.

Raj If you run a small firm, write down every AI tool you depend on. Know where your data goes, and have a backup ready.

Mei Ling And if you're a student, testing and improving models in Malay is a real skill now. Mesolitica's open models are free to study. Build something small, and show people. Malaysia doesn't need to win the frontier. It needs to make sure its people and its firms are never stuck. That's where the AI-YA team stands this week.

Aiman One thing to do this weekend, in under an hour. Write a one-page AI backup plan for your work or your firm. List every AI tool you or your team uses. Next to each one, write where the data is stored and whose model sits underneath. Then ask one question. If this tool disappeared on Monday, what would we use instead? Finally, take one real task, such as a customer reply in Malay. Remove the names and numbers, and run it through your backup. If the backup can't do it, you've found your weak spot early.

Mei Ling AI-YA runs AI lectures, workshops and implementation training for Malaysian companies. Details at ai-ya.my/podcast.

Aiman Tomorrow is our Saturday week in review. We look at the week's AI news for Malaysia, what moved since we covered it, and the biggest story we haven't covered yet. We also revisit the week's Verdicts and Big Questions.

Raj Before then, tell us where you stand. Build the engine, or train the drivers? Leave your answer in the comments on the episode page at podcast.ai-ya.my. And share this with one person at work who needs it.

Mei Ling Thank you for keeping us company at the end of a long week. Switch the screens off for a while this weekend, and enjoy the people at your table. Take care on the roads tonight, and we will see you tomorrow with the week in review.