MCP Weekly: OpenAI’s $250M BCI Bet, Azure MCP GA, Agentforce MCP

Welcome to the latest installment of the MCP Weekly Digest for 2026, covering major developments from January 14th to January 21st. As Agentic AI becomes more mainstream, we’re continuing to see an increase in partnerships and security developments.

TL;DR

OpenAI invested $252 million in a seed round for Merge Labs, a Brain Computer Interface (BCI) startup. ChatGPT Go was introduced, powered by the GPT 5.2 model and priced at $8 a month. OpenAI partners with [Cerebras] to deploy 750 MW wafer-scale systems. 

Github Security Lab announced the open source release of the Taskflow Agent. Salesforce architecturing with a partnership with informatica, releasing the Moirai Agent along with Agentforce MCP support.

Major Updates of the Week

Open AI & Its Developments

• Merge Labs: Investment in merge labs, their aim of bridging biology and artificial intelligence. Brain computer interfaces (BCI)  are the new frontier, they create a new way to interact with AI which is why OpenAI is interested in the seed round of the labs.
• ChatGPT Go: With the latest model GPT-5.2 Instant which has 10x more messages inclusive of file uploads and image creation comparative to its free tier and longer memory is now global where ChatGPT is available.
• Cerebras: Open AI signed a multi year deal to deploy 750 megawatts (MW) wafer-scale systems. The rollout begins in early 2026 and will proceed in multiple tranches through 2028.

Azure Functions GA for Model Context Protocol

Microsoft, announced the General Availability (GA) of MCP as support for Azure Functions.

Core features include:

• Streamable HTTP transport for both streamable-HTTP and Server-Sent Events endpoints.
• Built-in authorization to implement MCP requirements and free developers from writing custom security code.
• On-behalf-of (OBO) support for managed identities, simplifying API calls using the user’s identity.

The offering also provides a Self-Hosted option for deploying existing MCP SDK servers to Azure Functions, enabling scaling without platform-managed authentication.

Salesforce Architecturing with the Context

• MoiraiAgent: It is an agentic framework for time-series forecasting, unlike the traditional models, it picks up a broad range of information including historical values and contextual signals to give more reliable predictions.
• Agentforce MCP Support: Integration of MCP in the Salesforce ecosystem, providing an interface that allows Agentforce agents to connect public and private third-party servers with trusted gateways, able to control what servers are connected to the registry and what tools are available.

Further Spread & Partnerships

AWS announced Amazon Bedrock AgentCore, to build and deploy AI agents in a serverless hosting environment and by GitHub Actions workflow it automates the deployments of AI Agents on AgentCore Runtime.

Recently, Google Cloud introduced the usage of gRPC for MCP. Previously, organisations had to translate between JSON-RPC and gRPC services, now making gRPC their native support.The functiongemma model can now be fine-tuned. This capability allows for more precise selection of AI agent tools or enables the model to effectively handle highly specific, niche tasks.

Security Advancements

• GitHub Security Lab: The open source release of the Taskflow Agent , experimental agentic framework to scale security research stating MCP can be used to build on existing tools, for example CodeQL. It has been used by the internal team, now it has been shared with the participants of GitHub Secure Open Source Fund.
• Zscaler Integration MCP Server: This MCP server is an open source bridge connects the Zscaler Zero Trust Exchange with the AI Assistants like Claude or ChatGPT and provides access to a bunch of tools across Zscaler’s portfolio, rather going through consoles and scripts, it go through the preferred chatbot.

My Thoughts: The Trends for MCP Adoption in 2026

There has been a lot to take in this week regarding the ongoing advancements in Agentic AI. Reflecting on the past, we’ve seen tech giants undergo changes rapidly; remember how quickly cloud computing reshaped the enterprise landscape. Security has been a priority as more trust is being placed in AI. We can expect more automation, seamless AI-driven workflows, and perhaps even new regulatory frameworks emerging as AI becomes a trusted decision partner.

‍