google launches amazon while anthropic expands its platform

Key Takeaways

• Bybit launched the Model Context Protocol (MCP) to standardize multi-agent trading infrastructure, enabling agents like Claude and ChatGPT to access market data and execute trades without custom API work.
• SAP and Google Cloud partnered to integrate customer experience data with AI agents, shifting AI from an assistance tool to an execution layer for real-time marketing.
• Anthropic released the Mythos AI model, which has sparked global security concerns due to its ability to find hidden flaws in critical infrastructure software like banks and power grids.
• Scale AI acquired ICG Solutions to strengthen its national security AI stack, integrating real-time data stream capabilities to support the Department of War's AI-first strategy.
• Cyberhaven expanded its Unified AI & Data Security Platform to AWS, Azure, and Google Cloud Marketplaces, allowing enterprises to purchase security tools through existing cloud commitments.
• MIT researchers developed the RLCR technique, which teaches AI models to express uncertainty and reduces calibration error by up to 90 percent.
• WealthReach introduced the Multiply AI referral engine, which integrates over 50 years of referral coaching into automated workflows for registered investment advisors.
• Acronis launched GenAI Protection to give Managed Service Providers visibility into AI usage and protection against prompt injection attacks.
• The Commonwealth Secretariat hosted an AI election training event in Trinidad and Tobago to discuss opportunities and risks for electoral integrity in the region.
• Business leaders are recognizing that data quality and context are the biggest obstacles to successful AI implementation, driving a shift toward data fabric architectures.

Bybit launches official MCP for multi-agent trading

Bybit released its official Model Context Protocol (MCP) to standardize infrastructure for multi-agent trading systems. This tool allows developers to build automated trading desks using AI agents like Claude, ChatGPT, and Cursor without custom API integration. The protocol provides four core modules including real-time market data, trading capabilities, account visibility, and WebSocket data streams. Bybit ensures security by keeping API keys on developer infrastructure and not storing them. The release is available immediately with documentation on GitHub for immediate integration.

Bybit expands AI infrastructure with new MCP protocol

Bybit announced an official Model Context Protocol (MCP) to support multi-agent trading architectures on April 22, 2026. This infrastructure allows traders to access live market data and execute trades using natural language commands through compatible AI applications. The system scales from single-agent to multi-agent setups while removing the need for custom API configuration. Four functional modules are available including market data, trading, account management, and real-time data streams. Victor Wu, Head of AI Agent Architecture at Bybit, stated this creates a standardized foundation for enterprise-scale AI trading.

Bybit introduces MCP for automated trading systems

Bybit launched its official Model Context Protocol (MCP) to enable standardized multi-agent trading infrastructure. The protocol allows AI agents such as Claude, ChatGPT, and Cursor to integrate with Bybit services without custom code. It exposes four modules covering market data, trading functions, account assets, and real-time data streams via WebSocket. This shift moves exchange infrastructure toward AI-native workflows designed for continuous operation. The release reduces integration friction and accelerates the deployment of automated trading strategies.

Bybit AI launches MCP for multi-agent trading

Bybit released an official Model Context Protocol (MCP) to standardize infrastructure for multi-agent trading systems. The protocol enables AI agents like Claude, ChatGPT, and Cursor to access Bybit services without custom API work. It exposes four core modules including market data, trading capabilities, account visibility, and real-time data streams. The system supports scalable single-agent and multi-agent architectures for coordinated strategies. This release reflects a broader industry move toward AI-native infrastructure for financial workflows.

Strong data fabric essential for effective AI

Business leaders are discovering that data quality and context are the biggest obstacles to successful AI implementation. Without understanding business context, AI systems can generate fast answers but make wrong decisions. Irfan Khan, president of SAP Data & Analytics, explains that context is essential for good judgment and return on investment. Organizations are rethinking their data architecture to connect information across applications while preserving business semantics. This shift toward data fabric helps ensure automation reflects real business priorities rather than isolated decisions.

SAP and Google Cloud partner for AI marketing

SAP and Google Cloud announced a partnership to integrate customer experience data with AI agents for real-time marketing execution. The collaboration connects SAP CX and Engagement Cloud with Google's Gemini Enterprise Agent Platform. This allows marketers to access unified customer data and take real-time actions across tasks. The integration shifts AI from assistance to an execution layer that drives end-to-end marketing activity. Marketers can now define business goals and let agents coordinate execution across systems automatically.

WealthReach deploys AI referral engine for advisors

WealthReach introduced a new AI-driven referral tool called Multiply to help advisory firms generate consistent client growth. The platform integrates decades of referral coaching into automated workflows for registered investment advisors. It includes over 50 training videos, interactive modules, and a conversational AI interface based on the Feedback Marketing framework. The tool helps advisors refine referral conversations and supports systematic follow-up. WealthReach acquired the underlying intellectual property to make these proven strategies accessible to all firms on the platform.

Commonwealth Secretariat hosts AI election training

The Commonwealth Secretariat convened an AI election training event in Trinidad and Tobago on April 22, 2026. Election professionals from across the region met to discuss AI opportunities and risks for electoral integrity. The training examined fundamentals of AI, safeguarding information integrity, ethical issues, and legal considerations for AI in elections. The event was organized under the theme Artificial Intelligence: Opportunities & Risks for Electoral Integrity in the Commonwealth Caribbean & Americas. Participants will also receive strategy presentations and evaluations.

Cyberhaven available on major cloud marketplaces

Cyberhaven expanded its availability to AWS, Azure, and Google Cloud Marketplaces as of April 2026. This allows enterprises to purchase its Unified AI & Data Security Platform through existing cloud commitments without new budgets. The platform combines Data Security Posture Management, Data Loss Prevention, Insider Risk Management, and agentic AI capabilities. Iulia Stefoi-Silver, VP Alliances and Strategy at Cyberhaven, stated this makes securing AI and data easier for customers. The listings also support channel partners and resellers in their sales efforts.

Acronis GenAI Protection empowers MSPs

Acronis launched GenAI Protection to give Managed Service Providers control over AI usage and data risks. The solution allows MSPs to monitor AI usage across customer environments through a centralized console. Key capabilities include shadow AI visibility, sensitive data protection for AI interactions, and prevention of prompt injection attacks. Over half of SMBs now use AI tools, primarily in marketing and sales. Acronis plans to introduce additional AI-powered capabilities within its broader Cyber Workspace offering.

Vogue Business hosts AI in luxury webinar

Vogue Business will host a live webinar on May 7 featuring editors Amy O'Brien and Maliha Shoaib. The discussion explores survey findings on how consumers feel about AI applications in fashion and luxury brands. The team surveyed almost 300 consumers to understand their preferences for AI innovations in retail. The webinar will also cover predictions from brands, analysts, and futurists about the future of AI in luxury over the next decade. Topics include in-store retail transformation and whether AI will change consumer taste.

MIT researchers teach AI models uncertainty

MIT CSAIL researchers developed a technique called RLCR to teach AI models to express uncertainty in their answers. This method reduces calibration error by up to 90 percent while maintaining accuracy. Standard training methods often make models overconfident regardless of their actual certainty. RLCR adds a reward term that penalizes the gap between stated confidence and actual accuracy. The approach was tested on a 7-billion-parameter model across various benchmarks and outperformed standard training in calibration.

Anthropic's Mythos AI model sparks global security concerns

Anthropic released a new AI model called Mythos that has raised global security alarms. The model is capable of finding hidden flaws in software used by banks, power grids, and governments. World leaders are scrambling to understand the scale of security risks and how to address them. Russian outlets have called the model worse than a nuclear bomb. Experts warn that access to powerful AI models is becoming a geopolitical issue. Governments are urged to treat this as a policy wake-up call.

Scale AI acquires ICG for national security AI

Scale AI acquired ICG Solutions to strengthen its national security AI stack. ICG's LUX platform ingests and correlates real-time data streams for defense and intelligence environments. The acquisition aims to integrate streaming data capability directly into Scale AI's broader AI platform. This move supports the Department of War's January 2026 AI-first strategy memo and GenAI.mil rollout. Scale AI positions itself as a full-stack AI infrastructure company for national security. ICG will operate as a wholly owned subsidiary to maintain contract continuity.

Sources

• Bybit AI Expands to Infrastructure Layer with Official MCP Release for Multi-Agent Trading
• Bybit AI Expands to Infrastructure Layer with Official MCP Release for Multi-Agent Trading
• Bybit AI Expands to Infrastructure Layer with Official MCP Release for Multi-Agent Trading
• Bybit AI Launches MCP for Multi-Agent Trading Infrastructure
• AI needs a strong data fabric to deliver business value
• SAP and Google Cloud Partner to Bring Agent-Based AI Execution to Marketing
• WealthReach rolls out AI referral engine to systematize advisor growth strategies
• Commonwealth Secretariat convenes AI election training in Trinidad and Tobago
• Cyberhaven Now Available on AWS, Azure, and Google Cloud Marketplaces, Giving Enterprises Flexible Procurement for AI-Native Data Security
• Acronis GenAI Protection gives MSPs control over AI usage and data risks
• Register Now | The Future of AI, Live Webinar
• Teaching AI models to say “I’m not sure”
• Anthropic’s New Mythos A.I. Model Sets Off Global Alarms
• Scale AI Moves to Strengthen National Security AI Stack with ICG Acquisition