While Hedge Funds Lost Billions in 2025, This AI Posted 51% Returns

The hedge fund industry just survived one of the most brutal years in recent memory. April brought Liberation Day tariffs that wiped out over six trillion dollars in two trading days. Black swan events that probability models said shouldn’t happen in fifty years happened three times in 2025. Funds that looked strong in December took catastrophic mid-year hits that they barely recovered from. And more than a few didn’t. Pierre Andurand’s flagship fund was down over 50% mid-year. Man Group lost 7.8% in April alone and finished down fifteen percent for the year. Millennium and Citadel both posted around ten percent returns, which sounds fine until you remember that they spent months clawing back from their devastating spring losses.

Then there was Vertus.
The AI infrastructure that most institutional investors have never heard of was quietly processing over a billion dollars in daily volume for the hedge funds using their systems. And those funds posted 51.15% returns for 2025 with a 2.13 Sharpe ratio, according to independent audit verification. That performance beat Bridgewater’s 34%, DE Shaw’s 28.2%, and nearly every major fund on the planet.

Vertus isn’t a hedge fund. They’re the technology layer underneath. They provide the cognitive architecture, data management, and trade execution systems that institutional capital runs on. The three founders, Julius Franck, Alex Foster, and Michal Prywata, built infrastructure that never touches client capital but makes it move smarter. And in 2025, that made all the difference.

The Year That Broke Everyone Else
When President Trump announced sweeping tariffs on April 2nd, the markets didn’t wait to see if he was serious. The S&P; 500 dropped nearly five percent. Then another 6% the next day. The Nasdaq recorded some of the largest point swings in its history, with losses exceeding 970 points in major market selloffs. And the VIX spiked above 60 for the first time since COVID. Six trillion dollars evaporated in two trading days. One of the largest two-day losses in market history.

The carnage was widespread. Funds that had posted double-digit gains for years suddenly couldn’t navigate basic volatility. And strategies that worked in 2024 stopped working overnight. By October, systematic quant funds were hemorrhaging money daily as crowded positions unwound in one seemingly cascading wave after another.

The VIX spike above 60 wasn’t just a number. It represented billions in destroyed capital as volatility bets exploded, trades unwound, and risk models built for stable regimes collapsed under real pressure. Funds weren’t just losing money on bad calls. They were losing money because the entire playbook suddenly got rewritten and their systems couldn’t adapt fast enough. And these weren’t small shops. These were enormous firms staffed by PhDs, backed by decades of experience, and equipped with the kinds of technology most of the industry and most modest-sized investors will never touch. And when the conditions changed fast, even they struggled to adjust in real time.

Systematic hedge funds lost money every single day in early October. Not because the fundamentals changed, but because when everyone rushes for the same exit, the exit disappears.
Goldman Sachs called it exactly what it was: “Inherent fragility when too much capital chases the same quantitative signals.” By year end, funds were shutting down not because they’d blown up, but because they couldn’t compete. Eisler Capital managed four billion dollars and returned seven percent annually since 2021. Solid performance. Competent team. It didn’t matter. Paloma Partners returned over a billion dollars to investors after redemptions. This wasn’t really about a few black swans. This was policy uncertainty and crowded trades unwinding exactly the way they always do. And most of the industry wasn’t prepared. However, funds running on the Vertus infrastructure posted consistent 51.15% returns with no drawdowns.

Built for the Hardest Proving Ground on Earth
The founders chose financial markets deliberately. Not because they wanted to be a fintech company. But because they wanted to prove their AI in the most demanding environment that exists. Markets are unforgiving. Errors are costly. Feedback is instant. Decisions happen in microseconds with millions of dollars at stake. There’s no room to pretend and no credit for sounding smart while losing money. If your intelligence can’t handle that, it’s not intelligence. It’s a statistical curve with maybe a good story.

Vertus spent 2025 logging over 87 breakthroughs in their cognitive architecture. Not incremental improvements. Fundamental advances that would each define a year at most companies. Multilingual reasoning where the system thinks in code simultaneously because some concepts express better in certain languages. Memory systems that compress and abstract rather than store raw data. Learning through consequence rather than labels.

The first billion-dollar transaction day happened in under ten months of going live. By year end, billion-dollar days were routine. The system delivered 51.15% returns with lower drawdowns than the market and a 2.13 Sharpe ratio that indicated disciplined risk management. This performance was independently audited and verified. The billion dollars in daily volume represented real capital making real decisions in real time. And those decisions consistently outperformed human judgment in an environment where human judgment was tested harder than it has been in years.

Why This Changes Everything
When Vertus posted 51.15% returns, the immediate reaction from some corners was scepticism. Those kinds of numbers in that kind of year invite questions. But the data held up under examination. The trades were timestamped and verifiable. And the comparison to major hedge funds wasn’t marketing spin. Bridgewater manages roughly 120 billion dollars and has spent four decades refining its investment process. DE Shaw pioneered quantitative trading in the 1990s and employs hundreds of PhD-level researchers. Millennium aggregates some of the best traders in the world and allocates capital dynamically based on risk-adjusted performance. These aren’t weak players. These are three fundamentally different approaches to beating markets, each refined over decades, each backed by billions in capital and elite talent. And funds using the Vertus infrastructure beat all of them.

And the gap isn’t explained by luck. The 2.13 Sharpe ratio proves the returns came with controlled volatility. It simply shows that a fundamentally different approach to processing information and making decisions under uncertainty was more than just effective. And here’s what makes this particularly uncomfortable for the incumbents. Vertus isn’t a single fund having a great year. It’s an infrastructure layer deployed across multiple funds. That means that the competitive advantage isn’t limited to one balance sheet. It propagated across strategies, regions, and mandates. The Vertus intelligence itself became the primary competitive asset.

What Comes Next
For institutional investors, the world just shifted sideways and the calculus just changed.
Pension funds and endowments have fiduciary duties to maximize returns. So, when AI-enhanced infrastructure consistently outperforms traditional approaches, even in chaotic conditions, continuing to allocate to managers who refuse to adopt it becomes harder to justify. For hedge fund managers, the pressure just increased. The argument that AI can’t handle complexity or adapt to market conditions just lost its strongest defense. Vertus showed it can. For entrepreneurs and business leaders outside finance, the lesson is even bigger. This is what happens when you solve a problem everyone knows existed but nobody’s been able to crack. Markets are the ultimate proving grounds. Microsecond decisions. Millions at stake. Immediate feedback. Vertus built cognitive systems that operate at levels humans alone can’t reach and proved it can do it at institutional scale.

The founders are now expanding beyond finance. The same infrastructure that processed over $700m in average daily volume through 2025’s chaos is being adapted for scientific computing, autonomous systems, advanced mathematics, and research domains. And they’re expanding their presence to support this next phase. The question’s no longer whether AI infrastructure can outperform traditional approaches in complex environments. The 2025 data settled that. The question is how fast the rest of the market realizes what just happened and moves to get onboard. Because the funds that figured it out early just beat nearly everyone. And the gap wasn’t even close.