Opinion
Data vendors are vagueposting about MCP prices
Exchanges and data providers are charging for AI access to their data. Nobody will say how much.
Compute is leaving land. Recoverability isn’t following.
The limits of land may push datacenters to the ocean or even outer space, writes Vishal Gupta. But these options introduce new challenges.
Who’s counting the agent army?
The Waters Wrap: Agents are growing faster than the controls to manage them. Wei-Shen says that as firms struggle to measure their returns on investment, it may come down to who foots the bill.
Private credit has outgrown its spreadsheets
Private credit severely lacks operating leverage, writes Daniel Hellwig, but standardizing how terms are recorded and applying AI could inch it toward maturity.
The SEC’s innovation exemption is facing its first test
Waters Wrap: With more than 20 issuer companies submitting objections to OKX and ICE’s new joint venture for tokenizing stocks, Nyela wonders if some cracks are starting to show in the regulator's innovation exemption.
Waters Wavelength Ep. 360: Innovation, regulation, and exposure
This week, Tony and Shen talk about the balance between innovation and regulation.
AI assistants for human overseers? Or human checks and balances for AI actors?
The Waters Wrap: Speakers at a recent Bloomberg event emphasized the importance of keeping a human in the AI loop—but, Max asks, for how long?
The Big Bang—a complex 40-year data journey yet to reach its destination
In this guest column, Finbourne’s Gus Sekhon reflects on the deregulation and privatization of the London Stock Exchange and what it meant for data.
AI regulation: Who’s responsible for eliminating ‘gray areas’?
The Waters Wrap: In light of the Hugging Face hack and similar incidents involving rogue AI agents, Anthony wonders whether industry regulators will take the reins.
Waters Wavelength Ep. 359: GPU compute market and Sibos
This week, Shen and Nyela talk about the GPU compute market and recent technology conferences.
When trading systems disagree: a deterministic architecture for event sequencing
Modern trading platforms execute millions of events at extraordinary speed, but speed alone does not ensure that every downstream system shares the same view of reality, says Gaurav Tiwari.
How long will AI tokenomics matter?
The Waters Wrap: Banks and asset managers are rethinking the economics of AI. Wei-Shen examines how firms will manage token costs.
If everyone can be an expert in everything, is anyone an expert in anything?
The Waters Wrap: AI has promised to democratize knowledge and skillsets. Reb wonders whether that’s an entirely good thing.
Mind the stack: the foundation markets run on
Exchange data is an underappreciated component of the AI era, writes Jeff Kimsey.
Trading firms rethink infrastructure proximity as Asia’s FX markets grow
Firms that treat local infrastructure as core to how they serve clients in Asia stand to benefit from Singapore’s growing role as an FX hub, writes Judy Goh.
The impact of market data and alt data converging
The Waters Wrap: As firms combine traditional data sources with proprietary research and alt-data signals, Max has some questions about where the data market is headed.
Agents are invading. Will it stick?
The Waters Wrap: Agentic features are becoming commonplace in the workflow tools of capital markets. Nyela wonders if the trend is sustainable.
Waters Wavelength Ep. 358: Tradeweb’s Chris Bruner
This week, Tradeweb’s chief product officer joins the podcast to discuss fixed income, prediction markets, agentic AI, and overnight trading.
Can AI beat exceptions out of the back office?
The Waters Wrap: Agentic AI can help operations teams tackle exceptions. But first, they need to get their house in order, writes Wei-Shen.
Photonics: time for trading tech to see the light
The Waters Wrap: While the sector is dominated by Big Tech, photonic-based solutions could one day help trading firms take more control over their AI ambitions, Anthony says.
Banks should not count humans out in adoption of agentic AI
It will take time to strike a balance between humans and AI agents, and it may start in post-trade, Diederik Geeraerts writes.