These top five AI developments illustrate how artificial intelligence is embedding itself across the financial technology landscape in 2026, be it hardware, banking infrastructure, or credit card payments.
Nvidia Enters the Personal Computer Market With AI-Focused Chips
Nvidia, the world’s most valuable company and chipmaker whose data-centre processors have defined the current AI buildout, announced on 1 June a significant expansion into consumer computing. At Taiwan’s Computex conference, chief executive Jensen Huang unveiled the RTX Spark superchip — an Arm-based processor that combines central and graphics processing capabilities designed to power a new generation of Windows laptops and desktops.
The first devices running the chip are expected to reach the market later this year, manufactured by Microsoft, Dell, HP, ASUS, Lenovo and MSI. Huang described the development as one that would “reinvent the PC,” positioning the chip as a means of running AI workloads locally rather than relying on remote data centres. Initial models will be targeted at creators, developers and gamers, with a premium price point and chassis thickness of around 14 millimetres.
The announcement marks Nvidia’s first serious foray into the processor market traditionally held by Intel, AMD, Qualcomm and Apple. Huang also confirmed at the event that the company’s Vera CPU is in full production, with early deployments by OpenAI, Anthropic and SpaceX.
Anthropic Files Confidentially for an IPO
Anthropic, the AI laboratory behind the Claude family of models, confirmed on 1 June that it has submitted a confidential draft registration statement on Form S-1 to the US Securities and Exchange Commission, initiating the process for a potential initial public offering.
The company stated that the filing “gives us the option to go public after the SEC completes its review,” while noting that neither the number of shares nor the offering price has been determined. Any listing would remain subject to market conditions.
The move comes days after Anthropic closed a $65 billion Series H funding round that placed its post-money valuation at $965 billion, and follows a reported annualised revenue run rate of $47 billion — up from $10 billion in annual revenue at the end of 2025. Anthropic’s filing positions it ahead of rival OpenAI, which has also been preparing its own confidential submission for a market debut expected later in 2026. The AI company is now targeting an annualized revenue run-rate of about $20 billion for 2026, following a rapid climb in commercial uptake of its Claude models and enterprise products.
OpenAI Offers UK Banks Access to Cybersecurity Model
Nine major UK banks have been granted access to OpenAI’s GPT-5.5 Cyber model, a system designed to identify concealed software vulnerabilities and security weaknesses across complex digital infrastructure. The institutions gaining access include Lloyds Banking Group, HSBC and Nationwide. NatWest and Santander had already secured access through prior agreements. OpenAI has also extended the model’s availability to financial institutions in Canada, Japan and several European markets.
The development is set against a notable constraint: Anthropic has not yet made its own cybersecurity model, Claude Mythos Preview, available to UK financial institutions. Mythos, released in early April, attracted widespread attention for its capabilities in vulnerability detection and was initially made accessible to approximately 42 organisations, the majority of which were US-based technology companies. Bank of England Governor Andrew Bailey noted last week that British banks remained unable to use Mythos to test their own systems.
The UK’s AI Security Institute assessed both models and found them to deliver comparable performance across cybersecurity testing tasks. Former UK Chancellor George Osborne, who now works with OpenAI, stated publicly that the company’s objective is to keep such tools in the hands of institutions responsible for protecting critical infrastructure, rather than withhold them. The European Central Bank separately summoned lenders to a meeting last week, urging them to accelerate efforts to address IT vulnerabilities that advanced models have begun to expose.
Gradient Labs Doubles Its Series A to $26 Million
Gradient Labs, a London-based startup founded in 2023 by former Monzo employees, has raised an additional $13 million to bring its Series A funding to $26 million. The extension was led by Octopus Ventures and CommerzVentures, with continued backing from Redpoint Ventures and Exceptional Capital.
The company builds AI agents intended to automate regulated workflows within financial services, including customer onboarding, KYC verification, dispute resolution, lending processes and voice-based customer support. Its clients include Wise, Monzo, Zego, Current, Stash and Rho, reaching a combined base of 32 million end users. The company reported revenue growth of 900% over the past year.
Gradient Labs says its agents operate within compliance frameworks aligned with the FCA’s Consumer Duty requirements and the EU AI Act, and that customer satisfaction scores across deployments have reached as high as 98%. Co-founder and chief executive Dimitri Masin described the company’s goal as building “the agent layer that financial services need to run their customer operations autonomously.” The new capital will be used to support expansion in the United States and further product development.
Robinhood Extends Its Gold Card to AI Agents
On 27 May, Robinhood introduced an Agentic Credit Card feature that allows users to assign AI agents the ability to make purchases on their behalf, linked to a virtual card derived from the Robinhood Gold Card account. The feature is available to existing Gold Card holders, of whom there are approximately 700,000.
Rather than granting agents direct access to the underlying card number, the system generates an isolated virtual card for each agent. Users configure a monthly spending cap and can optionally require manual approval for transactions above a chosen threshold. The virtual card can be revoked at any time. Purchases made through the agentic card earn the same 3% cash back as those made directly by the cardholder. The agentic payments system connects to third-party AI agents, including Claude, Cursor and OpenAI’s Codex, via the Model Context Protocol, an emerging standard for AI-platform integrations.
Robinhood described example applications, including automated purchasing of limited-release goods when prices fall below user-defined thresholds, or restaurant reservation booking as soon as availability appears. The company introduced the feature alongside an Agentic Trading product, which allows AI agents to execute equity trades within a separate, ringfenced account. Chief executive Vlad Tenev characterised the launch as an extension of the company’s original mandate to broaden access to financial tools.
