A growing market, with limits on its use
Indonesia's Financial Services Authority, OJK, reported 23.21 million registered crypto-asset customer accounts in August 2026, a 1.23% increase from July's 22.93 million. The July figure was itself up from 20.19 million in December 2025.
Trading also increased in August. Crypto-asset transactions totaled IDR 23.16 trillion, up 12.87% from IDR 20.52 trillion in July. Transactions in digital financial asset derivatives reached IDR 4.56 trillion, a monthly increase of 33.76%.
Those monthly gains followed a weaker full year. OJK recorded IDR 482.23 trillion in crypto-asset transactions in 2025, compared with IDR 640.61 trillion in 2024, attributing the decline to global factors and the crypto market cycle. The regulator also reported that Indonesia placed seventh in the 2025 Global Crypto Adoption Index. Crypto-asset tax revenue amounted to IDR 796.72 billion in 2025 and IDR 1.96 trillion in February 2026.
OJK assumed responsibility for regulating and supervising digital financial assets, including crypto assets, from the Commodity Futures Trading Regulatory Agency, Bappebti, on January 10, 2025. Mokalu and Jamhari argue that a larger customer base must be matched by stronger operational capacity and consumer protection, rather than leaving risks to accumulate.
They also distinguish permission to trade an asset from permission to use it for payments. Bank Indonesia Governor Perry Warjiyo affirmed in 2021 that Bitcoin is not legal tender and that only the rupiah is valid as legal tender in Indonesia.
What the anniversary brings to the debate
The authors draw their central analogy from the closing of the Second Youth Congress in Batavia on October 28, 1928. Wage Rudolf Supratman performed "Indonesia Raya" on his violin without lyrics to avoid attracting the colonial government's suspicion. In their account, today's digital economy has technological momentum but still needs a shared framework of rules, institutions and trust.
They illustrate the stakes with the January 2024 fraud at the Hong Kong office of London-based Arup. A finance employee initially questioned an email about a confidential transaction. A subsequent video meeting with apparent colleagues, including his finance director, dispelled his doubts. He made 15 transfers totaling HK$200 million, about US$25.6 million.
Arup later confirmed that the meeting had used AI-generated fake video and audio. For Mokalu and Jamhari, the case shows why enthusiasm for AI must be accompanied by verification. The same technology can support useful work, including helping small retailers track sales, plan inventory and cut waste.
Productivity and the rights behind a token
The authors propose judging AI by practical results rather than the number of features a product offers. They cite a 2025 study by Brynjolfsson and colleagues in the Quarterly Journal of Economics involving 5,172 customer service agents. Generative AI assistance increased average productivity by 15%, with the largest gains going to less experienced workers.
They see potential applications for micro, small and medium enterprises facing disorganized accounts, excess stock or overdue invoices. The relevant measures, they argue, are time saved, fewer mistakes and better decisions.
In finance, tokenization represents assets or rights through tokens recorded digitally. The Bank for International Settlements, or BIS, sees potential for lower costs and greater efficiency. But its 2025 annual report cautioned that stablecoins were not yet sound money, and its 2026 report emphasized preserving trust in money as tokenization develops.
A gold-backed token illustrates the distinction between a record and the asset behind it. The authors explain that a digital entry alone cannot establish that gold exists. Inspection, secure custody and clear redemption terms remain necessary, while the legal relationship between the token and the gold determines whether its holder can enforce a claim. Smaller ownership units can reduce the minimum purchase amount without necessarily reducing investment risk.
Indonesia is testing several such models. By September 2026, seven participants had completed OJK's regulatory sandbox testing and received a "Passed" designation. Their models included gold tokenization, securities tokenization through a Fund Management Contract (KPD) scheme, tokenized property ownership benefits, a rupiah stablecoin issuer, custody of non-trading digital financial assets and crypto fund management. OJK had received 39 sandbox applications and 372 consultation requests since POJK No. 3 of 2024.
The authors also caution against treating an AI label as evidence of economic value. A 2023 paper by Saggu and Ante in Finance Research Letters reported average returns of 35.5% to 41.3% for AI-themed crypto assets in the two months following ChatGPT's launch, despite those assets having no direct connection to ChatGPT. They urge prospective investors to examine who pays for a service, why its token is needed and how token supply changes.
Data centers need power as well as capital
Computing depends on physical infrastructure. The International Energy Agency's April 2026 outlook projected global data-center electricity consumption of 950 terawatt-hours in 2030, up from 485 TWh in 2025. Consumption at AI-focused data centers was projected to roughly triple over that period. Separately, Alex de Vries estimated in Joule in 2023 that AI-related electricity use could increase by 85.4 to 134 TWh annually through 2027.
Indonesia has attracted substantial commitments. Microsoft announced in April 2024 that it would invest US$1.7 billion over four years in Indonesian cloud and AI infrastructure, alongside AI training for 840,000 Indonesians. The Jakarta Post reported IDR 114 trillion, or US$6.5 billion, in realized investment during the first half of 2026 in a services category that the investment minister said was dominated by data centers. Coordinating Minister for Economic Affairs Airlangga Hartarto has also said Nvidia is considering a US$4.5 billion investment in Batam, Riau Islands.
At the Artificial Intelligence Impact Summit on October 21, 2026, Airlangga said state utility PLN's 2025-2034 power supply plan calls for 70 gigawatts of additional capacity. He also described a push for a 100 gigawatt-peak solar program and put existing national data-center capacity at about 0.45-0.58 GW.
Airlangga projected Indonesia's digital economy at US$100-130 billion in 2026 and said AI's contribution was expected to exceed US$360-400 billion by 2030. He also outlined a government goal of developing 9 million people with digital skills by 2030.
Expansion beyond established locations is still being planned. Digital Infrastructure Director General Wayan Toni Supriyanto said on October 21 that a roadmap would seek private data-center investment in eastern Indonesia, with Sulawesi and Kupang among the possible locations. Development would depend on demand and infrastructure readiness. Draft presidential regulations covering data centers and subsea cables were entering public consultation, with the government seeking to reduce investment concentration in Java and Batam.
Mokalu and Jamhari argue that investment should be assessed both by the money arriving and by the capabilities it builds. Their questions include whether Indonesian suppliers obtain higher-value work, universities gain research opportunities and local professionals develop expertise. They also call for scrutiny of whether electricity demand fits grid capacity and surrounding communities' interests.
They cite Beata Javorcik's 2004 study in the American Economic Review, based on Lithuanian company data. It found productivity spillovers from foreign direct investment through local upstream suppliers, with those effects appearing in joint ventures involving domestic partners rather than wholly foreign-owned investments.
Controls for AI, asset custody and verification
For businesses adopting AI, the authors recommend specifying which decisions a system may advise on, which actions it may carry out and which transactions require human approval. They treat access restrictions, transaction caps, activity logs and emergency shutdown mechanisms as core operating controls.
Digital asset services also need clear arrangements for private keys, which authorize transactions. In February 2025, Bybit lost about US$1.5 billion in virtual assets after attackers took control of one of its ether wallets. The FBI subsequently identified North Korea as responsible.
The Arup fraud raises a related issue: a recognizable face or voice is not sufficient verification for a high-risk transfer. Kietzmann and colleagues warned in Business Horizons in 2020 that deepfakes exploit people's tendency to believe what they see and hear.
The US Financial Industry Regulatory Authority, FINRA, recommends checking requests involving money, account access or sensitive information through a separate trusted channel. Mokalu and Jamhari argue that such procedures require an organizational culture in which employees can question instructions, including requests presented as coming from senior leaders.
Preparing for uncertain quantum risks
The authors also call for longer-term cryptographic planning. Sufficiently powerful quantum computers could weaken some methods in use today, although the timing remains uncertain.
The US National Institute of Standards and Technology, NIST, finalized three post-quantum cryptography standards in August 2024. A BIS quantum-readiness roadmap published in July 2025 stressed governance and migration planning.
Mokalu and Jamhari recommend beginning with an inventory of cryptographic use, identifying dependencies on suppliers and testing changes gradually. They say claims that a product is "quantum-resistant" need technical explanations and verifiable testing.
Building skills across regions
The authors identify work for AI developers, data analysts, smart contract and security auditors, technology legal specialists and digital asset risk managers. Their education agenda combines technical training with an understanding of economics, including cash flow, consumer protection and conflicts of interest.
They propose practical university-industry projects such as AI-assisted bookkeeping for small businesses, digital asset service sandboxes and transaction-security research. Internships, they argue, should give students real problems, mentoring and opportunities to assess their results.
Participation also depends on access to education, reliable internet and understandable learning materials outside established technology centers. The authors highlight Indonesian, a shared language rooted in the Youth Pledge, as a means of spreading knowledge about AI, investing and digital security. They call for children and novice users to have learning environments protected from fraud, manipulation and misleading promises of profit.
