Skip to main content
Fintech Byte
Esc

Type to search

4,337 words · 19 min read
Monthly Market Intelligence
AI in Trading Primer
May 2026 · M05

The AI-in-trading space has moved through three distinct structural layers simultaneously in May 2026: the retail democratization layer, the institutional tooling layer, and the payment-rail infrastructure layer, each at a different maturity point but all accelerating in a compressed window.

  • The AI-in-trading space has — The AI-in-trading space has moved through three distinct structural layers simultaneously in May 2026: the retail democratization layer, the institutional tooling layer, and the payment-rail infrastructure layer, each at a different maturity point but all accelerating in a compressed window. At the retail layer, Robinhood's launch of Agentic Trading for its 27 million customers marks the clearest boundary event yet between advisory AI and autonomous execution AI; a retail investor can now delegate stock execution — and soon options, crypto, and futures — to a third-party AI agent operating inside a sandboxed, spending-controlled account via the Robinhood MCP server.
  • At the institutional layer, — At the institutional layer, Rogo's $160M Series D — bringing cumulative fundraising to approximately $300M with 250-plus institutional clients including Rothschild and Lazard — represents the benchmark funding size for AI tooling that automates multi-step investment banking workflows: deal screening, diligence assembly, and research synthesis. Symphony AI Agent Studio's BlackRock and CME Group partnerships, and the AIMA survey finding that 86% of hedge fund managers now use generative AI across 157 managers representing $783 billion in assets under management, indicate that LLM-driven research and diligence tooling has crossed from pilot to embedded workflow in a meaningful fraction of the institutional market.
  • At the infrastructure layer, — At the infrastructure layer, the Model Context Protocol is consolidating as the de facto connectivity standard between AI agents and trading platforms: cTrader/Spotware (11 million traders, 300-plus brokers), Sygnum Bank, Coinbase Base, and Robinhood have all launched production MCP servers within a single month, alongside Gemini's AI Agentic Trading connecting Claude and ChatGPT directly to exchange accounts. This convergence speed is atypical for financial infrastructure and reflects the AI industry's prior standardization work rather than finance-specific coordination; the protocol's open architecture means no single party controls the standard, which accelerates adoption but also means security and systemic-risk frameworks must be developed cooperatively or not at all.

Structural read: The structural change confirmed across May 2026 is that the AI agent has become a first-class participant in retail financial markets.

At The Institutional Layer
$160M
At the institutional layer, Rogo's $160M Series D…
Series D
$300M
At the institutional layer, Rogo's $160M Series D…
The Algorithmic Trading Market
$24B
The broader market projection is unambiguous: the…
CAGR Toward
$44.55B
2% CAGR toward $44.55B by 2030; AI trading is…
Confirmed
What Launched & Shipped
Confirmed
  • Robinhood Agentic Trading and Agentic Credit Card — mass-market retail launch: Robinhood launched autonomous AI agent trading for its full 27 million customer base on May 27, 2026, with a companion Agentic Credit Card.
    • Third-party AI agents connect to dedicated Robinhood trading accounts via an MCP server; accounts are sandboxed with spending controls and real-time activity feeds visible to the account holder; the feature launched in beta on May 27
    • Supported asset classes at launch are equities; the platform has stated plans to extend to options, crypto, and futures, and the Agentic Credit Card broader rollout is planned for later in 2026
    • The launch establishes Robinhood as the first large-scale retail broker to give autonomous execution authority to third-party AI models, converting the broker from a UI product into an execution API consumed by AI; 27 million addressable accounts represents the largest single retail agentic deployment yet recorded in any market globally
  • Moomoo API Skills — no-code agentic investing: Moomoo launched API Skills on May 25, 2026, enabling agentic investing without requiring users to write code.
    • The product provides a no-code abstraction layer over the Moomoo broker API, allowing retail users to connect AI agents for portfolio monitoring, rebalancing triggers, and execution instructions
    • The simultaneous launch by Moomoo and Robinhood on consecutive days confirms that the shift to AI-agent-accessible brokerage accounts is a platform category movement, not a single-platform experiment
  • Liquid Co-Invest — trading embedded inside ChatGPT and Claude: Liquid launched Co-Invest on May 26, 2026, embedding trade execution directly inside the ChatGPT and Claude chat interfaces.
    • Users interact with their AI assistant via natural language; the assistant executes trades in the user's Liquid account without leaving the chat interface, representing a distinct architectural approach from MCP-connected dedicated accounts
    • Where MCP-connected accounts grant the AI agent access to a broker's execution infrastructure, Liquid's model surfaces execution capability inside the AI's own interface; both approaches converge on AI-mediated trade execution and are likely to coexist for different user segments
  • cTrader / Spotware MCP servers — FX/CFD platform opens to AI agents at scale: Spotware launched official MCP servers for the cTrader platform on May 14, 2026, with remote and local server variants and a further rollout published across the subsequent two weeks.
    • The cTrader platform serves 11 million traders across 300-plus brokers globally; AI Agent Connect integrates Claude Code, ChatGPT Codex, and Gemini CLI as first-party supported agent runtimes; help documentation published May 16 includes a specific use case for daily market briefings delivered via AI prompt
    • Spotware's documentation explicitly flags prompt injection and tool-permission vulnerabilities — the platform shipped with security caveats acknowledged in production documentation, an unusual transparency posture that also signals the absence of a completed security framework at the time of launch
    • cTrader's 300-plus broker distribution means this single MCP launch wired a substantial fraction of global FX/CFD retail infrastructure to AI agent connectivity in a single step; no comparable single-platform coverage event occurred in equities or crypto within the same period
  • Sygnum Bank AI agent MCP — first regulated bank live digital asset transactions via AI agent: Sygnum Bank, the first regulated Swiss digital asset bank, completed live AI agent digital asset transactions using a Claude-powered MCP server, reported May 18, 2026.
    • The implementation uses Anthropic's Claude as the agent model operating through an MCP connectivity layer against Sygnum's digital asset custody and trading infrastructure; Sygnum has stated that regulatory and compliance review must be completed before client rollout
    • A FINMA-regulated institution completing live AI agent transactions sets a regulatory precedent in Switzerland; other jurisdictions will cite this as an existence proof when evaluating AI agent transaction frameworks in their own regulatory consultations
  • Coinbase Base MCP — DeFi wallet and on-chain operations via ChatGPT, Claude, and Cursor: Coinbase's Base network launched an MCP server on May 26, 2026, enabling AI agents to manage crypto wallets and execute DeFi operations from within ChatGPT, Claude, and Cursor.
    • The tool covers wallet management, on-chain transactions, and DeFi protocol interactions; the agent runtime does not need to leave the primary AI interface to complete on-chain operations
    • The Base MCP extends the standardization pattern from TradFi/FX into on-chain DeFi, completing a month in which MCP was adopted simultaneously across crypto, FX, and institutional equity/fixed-income workflows — three distinct market infrastructure segments that rarely coordinate technology adoption simultaneously
  • Rogo — $160M Series D closes at ~$300M cumulative raise: Rogo, the AI agent platform for investment banking workflows including deal screening, diligence, and research synthesis, closed a $160M Series D in May 2026.
    • The round brings cumulative capital raised to approximately $300M; the company counts 250-plus institutional clients including Rothschild and Lazard, establishing it as the category leader in investment banking AI tooling by both capital raised and named client density
    • At this funding scale, the moat is no longer model capability — it is the proprietary training data generated by 250-plus live institutional workflows, which no new entrant can replicate without years of deployment
  • Symphony AI Agent Studio — custom LLM services and MCP integration for institutional trading: Symphony launched AI Agent Studio in May 2026, enabling custom LLM service creation and MCP integration for institutional trading workflows, with BlackRock and CME Group among named partners.
    • The product positions directly in the institutional AI tooling segment, competing with Rogo but targeting a broader workflow scope across trading operations rather than purely investment banking research
    • BlackRock and CME Group participation signals that the buy-side and exchange infrastructure tier have both moved to active vendor evaluation in this category
On The Horizon
What's Rumored
Speculative
  • Apple Siri integration with Claude and Gemini as third-party AI agent runtimes — June WWDC announcement: Apple's Siri is reported to be in testing for integration with Anthropic's Claude and Google's Gemini as third-party AI agent runtimes, with a June WWDC announcement cited as the expected disclosure event.
    • If confirmed, the integration places trading-capable AI agents inside the iOS consumer device layer: any retail investor whose broker has shipped MCP connectivity would have a voice-to-execution pathway from a single iOS device interaction, without opening a brokerage app
    • The trading distribution implications of a Siri integration are substantially larger than any single broker MCP launch; Siri's installed base in iOS devices dwarfs all retail brokerage active-user counts combined, and the agent-delegation model does not require users to actively manage the integration
  • UpsideOnly (Gruhn) hundreds-of-millions funding round — discussions ongoing: UpsideOnly, the AI trading platform launched by former FTX executive Gruhn and built around BayesShield AI trained on 22 billion trades with a profit-sharing crowdsourced signal model, is reported to be in discussions for a hundreds-of-millions funding raise.
    • The platform's profit-sharing model — where signal contributors share in trading profits — represents a distinct crowdsourcing architecture not currently offered by any of the publicly disclosed AI trading platforms; if the capital raise closes at the reported scale, it would be the largest single funding event in AI-native trading strategy execution for the period
    • The CFTC's AI suspicious-trading detection implementation, cited in the same coverage, is directly relevant to a platform training on 22 billion trades and deploying a crowdsourced signal model; regulatory risk is the primary variable in the raise's valuation
  • CFTC AI suspicious-trading detection — implementation scope unconfirmed: The CFTC is cited as implementing AI-based suspicious-trading detection, but programme parameters, timeline, and enforcement posture remain unattributed to named officials or formal rulemaking notice.
    • The directional signal — the primary US derivatives regulator deploying AI for market surveillance — is consistent with stated directions from multiple regulators and should be treated as reliable at the directional level even without confirmed parameters
    • Market participants deploying AI agents in CFTC-regulated derivatives markets should begin mapping agent behavior logs to anticipated regulatory reporting requirements without waiting for formal confirmation
Money & Movement
Capital & People
Capital
  • Rogo — $160M Series D, ~$300M cumulative, 250-plus institutional clients: Rogo closed a $160M Series D in May 2026, the largest single round in AI-driven investment banking tooling this period.
    • Named clients include Rothschild and Lazard across deal screening, diligence, and research workflows; at $300M cumulative, the company is operating at a scale implying either a near-term IPO trajectory or strategic acquisition interest from data and analytics incumbents (FactSet, Bloomberg, S&P Global) that have the most to lose from enterprise AI tooling commoditizing their research distribution model
  • SUI Group — $15M into Nof1 (Alpha Arena) and $3M into Recursive Superintelligence: SUI Group co-led a $15M funding round for Nof1, operator of the Alpha Arena AI trading competition, on May 21, 2026, alongside a $3M investment in Recursive Superintelligence.
    • The $18M combined investment reflects continued venture appetite for AI trading benchmarking infrastructure even as Alpha Arena's own competition results showed mixed-to-negative outcomes; the investor thesis is that proprietary live-market benchmark data is valuable regardless of current model performance, because it becomes training data for the next generation of models
    • Nof1 has announced Alpha Arena Season 2 with proprietary model development, signaling that the competition operator is moving from a neutral venue to an active model developer — a significant scope expansion that changes the competitive dynamics of the benchmark
  • Taurex — $40M Series C, Matthew Wright returns as Global CEO: Taurex closed a $40M Series C for AI-driven trading technology on May 19, 2026; founder Matthew Wright returned as Global CEO alongside the capital event.
    • The capital targets expansion across Taurex's broker-serving client base; the founder-return pattern with a capital event is consistent with a growth-phase pivot from operational management to strategic product acceleration
  • Longbridge — $150M-plus raise, PortAI financial GPT, Gavin Chia as Southeast Asia CEO: Longbridge raised more than $150M and appointed Gavin Chia, formerly of IG Group, as Southeast Asia CEO in May 2026.
    • PortAI is positioned as the industry's first financial GPT; the Southeast Asia leadership hire signals geographic expansion targeting retail AI trading adoption in a high-growth regional market where existing incumbents have not launched comparable AI-native products
    • Combined with the China retail AI-trading data (ByteDance ArkClaw, Tencent QClaw, 80 million under-30 retail traders as the target demographic), Longbridge's raise reflects a well-capitalized bet that Southeast and East Asian retail AI trading is a distinct and large market opportunity
  • CoinQuant — $3M seed, 15,000 users, automated strategy execution on HyperLiquid planned: CoinQuant raised a $3M seed round in May 2026, with a product already serving 15,000 users and a planned automated strategy execution layer on HyperLiquid.
    • The seed scale and existing user base position CoinQuant as an early-stage operator targeting the on-chain systematic trading segment, where mature tooling comparable to TradFi algo execution does not yet exist
    • The HyperLiquid integration plan places CoinQuant in competition with native Hyperliquid trading infrastructure as well as with the growing cohort of on-chain strategy platforms that emerged from the DeFi ecosystem
  • Clear Street — Ed Tilly retires June 1, AI trading app beta timed to leadership transition: Clear Street CEO Ed Tilly announced retirement effective June 1, 2026; the AI-driven active trading app beta is timed to coincide with the transition to incoming CEO Cohen.
    • Clear Street exceeded $1B in annual revenue under Tilly; the AI trading app beta represents the flagship product initiative under Cohen's tenure, signaling that the incoming leadership is prioritizing AI execution capability as the primary growth vector
    • The timing of a technology-pivot announcement alongside a CEO transition is a sequencing choice: it associates the new leadership with a forward-looking AI product rather than the established institutional-clearing business
Structural Signal
  • The structural change confirmed across May 2026 is that the AI agent has become a first-class participant in retail financial markets
  • The previously implicit assumption — that AI tools assisted human trading decisions — is now operationally false for a material and rapidly growing segment of retail accounts
  • Robinhood's 27 million customer base, Moomoo's API Skills, and Liquid's embedded execution collectively represent a new category of market participant that is not a high-frequency trading firm, not a quant fund, and not a robo-advisor: it is a retail human who has delegated discretionary authority to an AI agent operating inside a sandboxed brokerage account
Policy Watch
Regulatory & Legal
Regulatory
  • EU PFOF ban effective June 30, 2026 — structural forcing function for AI agent execution economics: The EU ban on payment for order flow takes effect June 30, 2026, with direct implications for AI-mediated retail execution models in European markets.
    • European brokers operating PFOF-dependent revenue models must complete their transition by the effective date; AI agent execution models that route on best-execution terms rather than PFOF arrangements become structurally advantaged in the post-ban environment
    • The argument that disciplined AI agents — executing on best-available price rather than routed for rebate — represent the structural disruptor to the exchange PFOF churn model was made explicitly in corpus analysis; the June 30 date converts this from a theoretical advantage to a regulatory requirement that creates a competitive differential between PFOF-dependent brokers and AI-first execution platforms
  • EU AI Act enforcement operational horizon — concurrent AI agent transaction rules: The EU AI Act's operational horizon for AI agent transaction rules is cited as approximately mid-2026, creating concurrent timing with the PFOF ban and MiCA enforcement.
    • Institutions deploying AI agents for financial transactions in the EU face simultaneous compliance obligations across three regulatory regimes in a compressed timeline; MiCA governs crypto asset service providers, the AI Act classifies and imposes obligations on AI systems, and MiFID II / PFOF rules govern best-execution and order routing
    • No single compliance programme covers all three regimes; firms need integrated AI-agent frameworks — covering agent classification, transaction authorization, audit trail, and model accountability — that span all three simultaneously, and those frameworks do not yet exist as off-the-shelf solutions
  • IOSCO regulatory consultations on AI in capital markets — forthcoming, no date confirmed: IOSCO has indicated forthcoming regulatory consultations on AI applications in capital markets; the substance and timing have not been disclosed.
    • IOSCO consultations carry outsized structural weight because their conclusions shape domestic rulemaking in all major member jurisdictions simultaneously; the primary global standard-setter for securities markets developing an AI-in-trading framework will constrain or enable AI agent deployment in cross-border capital markets more consequentially than any single-jurisdiction rule
    • Institutional participants with cross-jurisdictional trading operations should treat the forthcoming IOSCO consultation as the most structurally significant near-term regulatory input and begin pre-consultation engagement to ensure that live deployment experience informs the standard's design
  • Visa Agentic Ready — live in Canada with five major banks; BOCHK trial complete: Visa launched its Agentic Ready programme in Canada with BMO, CIBC, RBC, Scotiabank, and TD as participating issuers; separately, Visa and BOCHK completed a live AI agent payment trial in Hong Kong executing real-world purchases.
    • The Canadian programme establishes consumer-protection and authorization protocols for AI agents making payments on behalf of cardholders; the Hong Kong trial provides empirical evidence that the protocols function in live transaction environments
    • Visa's simultaneous presence in Canada and Hong Kong reflects a deliberate multi-jurisdiction pilot strategy designed to generate regulatory-engagement data across different supervisory regimes — the two jurisdictions represent both a G7 banking market with established card infrastructure and a high-volume Asia-Pacific market with a distinct regulatory posture
  • Google AP2 protocol — 120-plus partners including PayPal, 95% of merchants seeing AI agent traffic: Google's Agentic Payments Protocol (AP2) has enrolled 120-plus partners including PayPal, with a PayPal survey indicating 95% of merchants are already experiencing AI agent traffic.
    • The 95% merchant AI-agent-traffic figure is the most consequential data point in the payments-infrastructure thread: it establishes that AI agent commerce is already at scale from the merchant side, even as the explicit authorization and payment-protocol infrastructure is still being built; merchants are encountering AI agents without having implemented machine-readable catalogs or agent-specific authorization flows
    • The gap between AI agent traffic prevalence (95% of merchants) and merchant preparedness (only 20% with machine-readable catalogs) represents a structural mis-match that the AP2/x402/MCP payment-infrastructure race is attempting to close
  • CFTC AI suspicious-trading detection — scope and timeline unconfirmed: CFTC implementation of AI-based suspicious-trading detection is reported without named attribution to confirm programme parameters.
    • The directional signal is consistent with stated regulatory direction across multiple jurisdictions; market participants deploying AI agents in CFTC-regulated derivatives markets should treat it as reliable at the directional level and begin documentation and audit-trail practices accordingly
Monthly Delta
Month-over-Month Shifts
Delta
Net-new
  • Retail AI agent execution at mass-market scale: Robinhood Agentic Trading, Moomoo API Skills, and Liquid Co-Invest collectively establish that retail autonomous execution has crossed from niche product to mainstream feature offered by multiple platforms simultaneously to tens of millions of users. The speed of concurrent launch by competing platforms — without coordination — indicates that each independently concluded the market timing was correct in the same window.
  • MCP protocol as the AI-trading infrastructure standard: Five distinct platform launches in a single month — cTrader, Sygnum Bank, Coinbase Base, Robinhood, and Gemini AI Agentic Trading — confirm that MCP has crossed the adoption threshold required to function as a de facto standard. Protocol standardization events in financial infrastructure typically take years; the convergence on MCP across FX, crypto, digital assets, and retail equities in a single month reflects the AI industry's prior work rather than finance-specific coordination, and it means interoperability across the resulting ecosystem is immediately available rather than years away.
  • Agentic payment rails converging on USDC and x402: Amazon Bedrock AgentCore, Google AP2, Visa Agentic Ready, Stripe's 280-plus AI-native updates, and the Solana Foundation/Google Cloud agent payment gateway deployed simultaneously; the Keyrock data quantifying 98.6% USDC dominance across 176 million AI-agent-settled blockchain transactions provides the empirical confirmation of what was previously a directional expectation.
  • AI agent performance benchmarking as a distinct commercial information category: Nof1's Alpha Arena, the Hyperliquid six-model competition, and the Alpha Arena 1.5 tournament results create a new category of market intelligence — live AI trading performance data with model-level granularity — that did not previously exist as a commercially backed, regularly published product. The $18M invested into Nof1 and Recursive Superintelligence despite mixed performance results confirms that the benchmark infrastructure is valued independently of what the benchmarks currently show.
  • AI agent trust and identity as a distinct security subsector: Lyrie.ai's Agent Trust Protocol and Sumsub's AI agent verification launch mark the emergence of a security sub-sector specifically addressing AI agent identity and authorization, distinct from existing AML, KYC, and fraud detection categories. The 44% of Greater China consumers reporting negative AI agent experiences and the 77% wanting human approval for payments quantify the trust gap the sector is addressing.
What This Means For You
Engagement Implications
Actionable
regulated equity venue or exchange evaluating AI agent market structure impact:
  • the EU PFOF ban effective June 30, 2026 combined with the explicit argument that AI-agent best-execution routing structurally displaces PFOF revenue models warrants a formal market microstructure analysis of projected agent-driven order flow composition before the ban date; initiate that analysis now to establish a baseline, and stress-test the scenario in which 5-10% of retail order flow in EU markets migrates to AI agent execution within 12 months of the Robinhood/Moomoo launches reaching European users.
prop-trading client or systematic hedge fund evaluating AI model procurement:
  • the Alpha Arena and Hyperliquid competition data provide the clearest available live benchmark for current LLM trading performance at the asset-class level; stress-test the assumption that off-the-shelf LLM deployment generates alpha before the next board cycle, distinguish between general-purpose LLM performance (consistently loss-making in equity competitions) and model-specific optimized performance (Qwen3-Max +22.32% in crypto, Grok-4.20 the only net-positive equity model), and evaluate whether model-specific infrastructure justified by those differentials is worth the vendor concentration risk.
retail broker or trading platform assessing MCP integration strategy:
  • the simultaneous MCP adoption by cTrader (300-plus brokers), Robinhood (27 million customers), Coinbase Base, and Sygnum Bank in a single month means the protocol is now table-stakes connectivity; any platform not on a confirmed MCP integration roadmap before Q3 2026 is ceding the AI-agent-accessible broker segment to competitors that moved first, and the segment's growth trajectory — driven by every AI model developer shipping agent capabilities — will make catching up progressively more expensive.
compliance or regulatory affairs client at a financial institution deploying AI agents:
  • the concurrent arrival of EU AI Act enforcement (mid-2026), EU PFOF ban (June 30, 2026), and forthcoming IOSCO AI-in-capital-markets consultations represents a multi-regime compliance stack arriving in a compressed window; escalate to the compliance committee the requirement for an integrated AI agent compliance framework covering agent identity, transaction authorization, behavioral audit trails, and model accountability — and begin that escalation before Q3 planning cycles close, because the integration effort is a multi-quarter program, not a point-in-time review.
crypto-native fund or on-chain trading operation:
  • the 98.6% USDC concentration in AI-agent-settled transactions and the Gartner projection of $15T in AI-agent-intermediated purchases by 2028 confirm USDC as the de facto AI agent settlement currency; evaluate treasury, settlement infrastructure, and operational USDC capacity against projected agent-driven transaction volumes, and assess concentration risk — 98.6% in a single stablecoin issuer is a systemic dependency that has no precedent in traditional settlement infrastructure and warrants explicit risk management rather than passive acceptance.
market-maker or liquidity provider assessing AI agent counterparty risk:
  • the Alpha Arena data showing AI agent retail counterparties predominantly loss-making in equity markets is favorable for market-maker profitability in the short term, but model generations improve rapidly and the Qwen3-Max crypto-leg performance (+22.32%) demonstrates that profitable AI agent trading is not a theoretical future state — it is happening now in specific asset class and model combinations; model the 12-24 month inflection scenario where agent performance crosses the profitable threshold in equity markets, and assess how that changes your adverse-selection exposure and spread-capture assumptions.
investment banking or M&A advisory client evaluating AI tooling procurement:
  • Rogo's $300M cumulative raise and 250-plus institutional client count at Rothschild, Lazard, and comparable firms establishes it as the reference deployment in investment banking AI tooling; evaluate Rogo and its nearest competitors — Symphony AI Agent Studio (BlackRock, CME Group partnerships) and the emerging Fiserv agentOS marketplace — as integration targets before the client-density flywheel makes competitive alternatives commercially unviable; delay of more than two quarters risks being a late entrant in a category where first-mover data advantage is self-reinforcing.
Watch These Closely
Forward Signals & Dated Catalysts
Upcoming
Confirmed
  • EU PFOF ban effective: June 30, 2026, with structural implications for AI agent best-execution routing economics in EU markets and a direct competitive differential between PFOF-dependent brokers and AI-first execution platforms (
  • Clear Street AI-driven active trading app beta: June 2026, timed to incoming CEO Cohen taking over from Ed Tilly on June 1 (
  • Fiserv agentOS wide availability: August 2026, expanding the 13-agent marketplace to the full Fiserv core banking distribution network (
  • FIS and Anthropic Financial Crimes AI Agent wider deployment: H2 2026, expanding beyond BMO and Amalgamated Bank to FIS's broader network of approximately 12% of world banks (
  • Robinhood Agentic Trading expansion to options, crypto, and futures: planned, no specific timeline; equities launch May 27 is explicitly framed as phase one of a multi-asset rollout (
Rumored / Analyst Projections
  • Apple Siri integration with Claude and Gemini as third-party AI agent runtimes: June WWDC announcement expected; if confirmed, creates the consumer-device pathway for voice-initiated AI agent trading at iOS-installed-base scale (
AI in Trading Monthly Primer — 2026-05 fintechbyte.com