The story begins in Singapore, where the language of innovation carries unusual force. By 2019, the city-state had become a laboratory for wealth-tech, a place where cryptocurrency, automation, and aspirational finance could be packaged together and sold in the polished grammar of modernity. The public record around Torque Trading is thinner than the marketing material the company circulated, but the contrast itself is revealing: in the press and on social channels, the platform presented itself as a disciplined AI trading operation; in the aftermath, investigators and journalists described a structure that resembled a classic deposit-fueled fraud dressed in machine-learning clothing.
That contrast mattered because Singapore was not some anonymous offshore backdrop. It was a jurisdiction that marketed itself on order, compliance, and financial seriousness, a place where a polished office address and a professional presentation could carry persuasive weight. In 2019, that mattered even more because cryptocurrency remained unfamiliar to many would-be investors. Terms like “custody,” “execution,” and “proof of trading” were still opaque to the average retail customer, and that ambiguity created room for a platform to sound technical without having to prove much of anything. The aura of legitimacy was doing a great deal of the work.
Bernard Ong emerged as the face most closely associated with that pitch. Available reporting identifies him as a central figure in Torque Trading’s operations, though the precise division of labor among organizers, recruiters, and back-office handlers is not always fully laid out in public filings. That gap matters. In modern financial frauds, the architecture is often flattened for outsiders: one man speaks, another signs, a third moves the money. The result is a company that looks coherent from the outside and improvisational from the inside. The public sees a brand. The investigators later see the seams.
The conditions for the scheme were unusually favorable. Crypto markets were still young enough that many investors lacked a mental model for how a real trading operation should look. A serious platform would leave records: account statements, trade confirmations, exchange logs, custody arrangements, and a trail that could be checked against independent data. But to many ordinary investors, those mechanisms were invisible. Torque Trading could therefore present a polished façade without being immediately forced to supply hard proof. When a company says it uses advanced systems to generate returns, most people do not know what evidence to demand first.
The first crossing of the line likely came, as it so often does, not with a grand theft but with a test: a small pool of deposits, a modest set of claimed returns, a few early statements that could be adjusted by hand if necessary. In frauds of this sort, the initial temptation is rarely the hardest part; the hardest part is deciding that the temporary fix will become the business model. Once the operator learns that a fabricated result can be paid out with incoming funds, the enterprise changes character. It stops being a bad investment and becomes a machine for convincing strangers to subsidize its fiction.
A key structural advantage for Torque Trading was the aura of algorithmic competence. AI carries a peculiar authority in finance because it sounds like expertise without requiring explanation. A bot can be said to be learning, optimizing, arbitraging, adapting to volatility; each term is vague enough to comfort the nontechnical and opaque enough to frustrate the skeptical. In a market that rewards speed and punishes doubt, that opacity can function as a sales tool. It can also blunt the questions that might otherwise be asked early: Which exchange is the bot using? What are the wallet addresses? What performance records can be independently verified? Who audited the system? Those are the kinds of questions that separate software from theater.
One surprising feature of the case is how ordinary the setup appears once the gloss is stripped away. The public-facing promise was not a new economic theory or a novel asset class. It was a familiar one: hand over your money, let the system work, receive steady returns. The sophistication was in the packaging, not necessarily in the mechanics. According to later investigative descriptions, the “AI” component was far less active than advertised, while the business depended on a stream of new customer deposits to keep the illusion of performance intact. That is the tell in schemes like this: money does not vanish at the point of collection. It is kept moving just long enough to sustain belief.
Concrete scenes matter here. In Singapore offices fitted with glass walls and neat reception areas, employees and promoters could point to screens, dashboards, and branded materials that made the operation appear data-driven. Elsewhere, in the everyday spaces where prospects made decisions — café meetings, messaging apps, small investor seminars — the sales pitch traveled in short, confident phrases that sounded modern precisely because they were hard to verify. No single tactic was remarkable. The accumulation was the trick. A clean office. A technical term. A promise of steadiness. A presentation that looked more like a fintech startup than a high-risk pitch.
The early money matters too. Fraud rarely scales from zero; it begins with a small circle willing to trust, or to suspend disbelief, long enough for the first funds to enter the system. Those initial deposits are the oxygen that lets the operation breathe. Once withdrawals are delayed, explanations can be crafted. Once explanations are accepted, the next solicitation becomes easier. That is the dangerous arithmetic of early success: it creates evidence where there is only momentum. A successful first payout can be more persuasive than an entire brochure. It tells the investor that the machine is real, even if the machine is being fueled by the next person’s money.
There is still a tension in the record about exactly how much of Torque Trading’s internal machinery was designed from the start and how much was assembled under pressure. Public documents do not fully resolve whether the founders intended a pure Ponzi from day one or whether the fraud hardened after the trading claims proved unsustainable. What is clear is that the enterprise was operational long before most victims understood the difference between a trading platform and a transfer mechanism. By the time those distinctions became relevant, the structure had already been built to favor speed over scrutiny.
This is where the stakes sharpen. A true trading operation can survive questions because it can produce records. A fabricated one has to rely on trust, delay, and momentum. If someone had demanded independent proof early enough — ledger records, exchange custody details, bank trail documentation, a clear explanation of where the supposed AI was actually deployed — the story might have become harder to sell. But frauds do not need universal belief; they need enough belief for long enough.
By the time the first money began circulating through Torque Trading’s accounts, the central lie had already taken shape: the company was not merely trading crypto; it was performing competence. That distinction would matter later, when the pressure of real withdrawals forced the performance to answer to mathematics. For now, the screens stayed lit, the returns were credited, and the machine was ready to meet the people who wanted to believe it.