Thinking Machines partners with Bridgewater to build AI model that cuts errors by nearly 30%. Mira Murati's startup and the world's largest hedge fund built a custom model that outperforms GPT, Claude, and Gemini on financial document tasks while slashing inference costs by 13.8x 7 hours ago When the world's largest hedge fund decides its analysts are spending too much time on document busywork, it doesn't just buy a ChatGPT subscription. Bridgewater Associates teamed up with Thinking Machines Lab, the AI startup founded by former OpenAI CTO Mira Murati, to build a custom fine-tuned model that reduces errors by 29.8% compared to the best available frontier models. The results, published June 30 by Bridgewater's AIA Labs and Thinking Machines Lab, show the specialized model hitting 84.7% average accuracy across six information-filtering tasks. Leading models like GPT, Claude, and Gemini variants, even when juiced with expert prompt engineering, were stuck in the mid-70s. How they built it. The model was constructed on the Qwen3-235B base and trained using Thinking Machines' proprietary Tinker platform. Two training techniques did most of the heavy lifting: interleaved batching delivered a 12.1% accuracy boost, while on-policy distillation added another 3.1%. The Tinker API handled the infrastructure side, letting the team iterate rapidly without managing GPU clusters directly. Inference costs dropped by a factor of 13.8x per task compared to frontier models. What the model actually does. The six tasks the model handles fall into the category of document triage. Relevancy classification, truncation, and labeling are the core functions - the information-filtering steps that happen before the actual analysis begins. The key ingredient that made this work wasn't just clever engineering. It was Bridgewater's proprietary expert-labeled data. General-purpose models are trained on internet text. When you fine-tune on thousands of examples labeled by domain experts who understand exactly what matters in financial documents, the model learns distinctions that no amount of prompt engineering can teach a generic system. What this means for investors. Bridgewater manages roughly $100B in assets. For the broader AI industry, this partnership validates Thinking Machines Lab's approach. Murati left OpenAI in late 2024 and launched the startup with a thesis that the next wave of AI value creation would come from customization, not just scaling. Disclosure: This article was edited by Editorial Team. For more information on how Crypto Briefing create and review content, see its Editorial Policy.

Oliver Simon recognized as a 2026 Institutional Investor Hedge Fund Rising star. June 2, 2026 Oliver Simon, Head of AI & ML Investment Strategy at Bridgewater Associates, was named to Institutional Investor's 2026 Hedge Fund Rising Stars - a list recognizing the next generation of leaders reshaping the investment industry. His selection reflects a decade at Bridgewater gravitating toward the hardest, most important problems and building AI systems that push the limits of what investing can be. Oliver grew up in New York City, the son of documentary filmmakers, and came to markets early - through poker and a growing conviction that financial systems were among the most complex and consequential worth understanding. That curiosity brought him to Bridgewater, where problems worth solving are the hardest ones. He made his mark early in his tenure. In partnership with Co-CIO Greg Jensen, Oliver developed Bridgewater's single-name equity trading strategies - something the firm had not attempted to scale in years. Oliver drove the effort end-to-end, combining research, systemization, and execution to bring the strategy to life in a way that was entirely new. His work remains a significant part of our single-name equity trading today. In 2023, Oliver was one of only six people Greg selected to launch AIA Labs with a mission to build a fully artificial investor capable of outcompeting humans. Oliver brought a rare perspective to the work: a deep fluency across investment, science, and technology - and the ability to unite investors, scientists, and engineers. In his current role, he leads the investment system at AIA's core, building what he describes as a "brain" - an integrated, autonomous investor that combines structured and unstructured intelligence into a coherent whole. Past performance is not indicative of future results. There can be no guarantee that any expected performance can or will be achieved and expected performance should not be solely relied upon in making any investment decision. Investment strategies that utilize AI are subject to certain additional types of risk, described here. Please see important disclosures and other information.

Data centers at sea: Oregon's Panthalassa nets $140M led by Peter Thiel for wave-powered AI. Wave energy had largely been bobbing around in the background of the U.S. clean energy sector - until now. On Monday, Oregon-based Panthalassa announced a $140 million round led by Peter Thiel. The new funding from the PayPal co-founder and others will allow the startup to finish building its pilot manufacturing facility near Portland. Panthalassa is developing technology that pairs wave power generated by massive floating orbs with onsite AI computing. The systems transmit data via low-Earth-orbit satellites. "We've built a technology platform that operates in the planet's most energy-dense wave regions, far from shore, and turns that resource into reliable clean power," said Garth Sheldon-Coulson, Panthalassa's co-founder and CEO, in a statement. "We're now ready to build factories, deploy fleets, and provide a sustainable new source of energy for humanity." The planet is scrambling to find new energy sources to meet demand from data centers and electrified transportation, building heating and cooling, and industrial applications. One of the biggest challenges historically with wave power is the need to build costly infrastructure to move energy from the ocean to where it's needed. Panthalassa's approach sidesteps that problem by using power onsite to run already-trained AI models, while tapping cold ocean water to cool the hardware - solving two problems at once. The strategy shares parallels with surging interest in space-based data centers that harness solar energy. In March, Starcloud, a Redmond, Wash.-based startup, announced $170 million in new funding, vaulting it to unicorn status with a $1.1 billion valuation. "The future demands more compute than we can imagine," said Peter Thiel. "Extra-terrestrial solutions are no longer science fiction. Panthalassa has opened the ocean frontier." Founded in 2016 as a public benefit corporation, Panthalassa has spent nearly a decade developing power generation, propulsion, autonomous operations and computing technology. That work has included prototypes - Ocean-1, Ocean-2 and Wavehopper - deployed in sea trials in 2021 and 2024. The company is now preparing to deploy its Ocean-3 pilot series this year, with commercial systems planned for 2027. Sheldon-Coulson previously served as a senior investment associate at Bridgewater Associates. Chief Innovation Officer and co-founder Brian Moffat developed a novel wave energy system for Spindrift Energy before launching Panthalassa. The company has 120 employees. Other Pacific Northwest companies pursuing wave energy include Seattle's Oscilla Power and Oregon State University spinout C-Power. Wave energy startups that have exceeded $100 million in investment globally include Sweden's CorPower Ocean and the United Kingdom's Marine Power Systems. The Series B round included participation from new investors John Doerr, Marc Benioff's TIME Ventures, Max Levchin's SciFi Ventures, Susquehanna Sustainable Investments, Hanwha Group, Anthony Pratt, Fortescue Ventures, Future Positive, WTI, Nimble Partners, Super Micro Computer, Sozo Ventures, Dylan Field, Planetary VC, Leblon Capital, Resilience Reserve, Portland Seed Fund, and the Intrepid Oregon Fund. Returning investors include Founders Fund, Gigascale Capital, Lowercarbon Capital, Unless and WovenEarth. Panthalassa has raised a total of $210 million, a spokesperson confirmed. Editor's note: Headcount and total funding updated with new information from Panthalassa.