Skip to content
Daily Brief News Daily Brief

5 AI stories from June 2-8, 2026: Anthropic Partner Hub with 40k applicants, SpaceX IPO pricing Thursday, GitHub Copilot code review adds MCP, EU AI Act 55 days out, and Gemini 3.5 Pro launch window

· by Pondero Newsdesk · 5 stories

AI news daily brief: 2026-06-08

Five stories today: one partner ecosystem launch from Anthropic, one landmark IPO pricing event, one developer tools update with affiliate relevance, one regulatory deadline now 55 days out, and one model launch still pending.

Anthropic formalises Claude Partner Network with a three-tier Services Track and public Partner Hub

Anthropic published the Services Track and Partner Hub details on June 3, 2026, per the Anthropic announcement. The Services Track is a three-tier structure measuring what firms have built and delivered with Claude in production, not just what they have certified on paper.

The tiers are Select, Preferred, and Global Premier. Select requires a minimum of 10 certified practitioners, 2 active production deployments, and 1 published customer endorsement. Preferred requires 100 certified practitioners, 15 production deployments, and 5 endorsements. Global Premier sets the bar at 1,000 certified practitioners, 100 production deployments, and 10 endorsements. Tier promotions happen on January 1 and July 1 each year. The first October 1 review gives newly joined firms an additional evaluation window in year one.

More than 40,000 firms applied to the Claude Partner Network after its March 2026 launch, and more than 10,000 individual consultants earned a Claude certification, per Anthropic. The program was backed by a $100 million Anthropic investment in partner training, technical support, and shared marketing. Named firms in the program include Accenture (30,000 professionals training on Claude), Cognizant (350,000 associates), Deloitte (470,000 staff), KPMG (276,000), PwC, and Infosys.

The Claude Partner Hub connects directly to Claude via a new MCP connector and updates daily so customers can search for qualified firms by tier and practice area.

Read more: Introducing the Services Track and Partner Hub of the Claude Partner Network.

SpaceX prices its Nasdaq IPO on June 11 at a reported $1.75T valuation

SpaceX is pricing its Nasdaq offering under the ticker SPCX on Thursday, June 11, 2026, with trading opening Friday, June 12. The 21-bank syndicate is led by Goldman Sachs. Per reports citing institutional investor roadshow materials, the company is targeting proceeds of approximately $75 billion at a valuation near $1.75 trillion, which would place it among the largest IPOs in recorded history. Reuters and Bloomberg both covered the pricing timeline; Pondero could not independently verify the specific valuation figure from a primary filing since the S-1 remains under SEC review as of June 8.

Thirty percent of the float is allocated to retail investors via Robinhood, Fidelity, and Charles Schwab, an unusually large retail share for a mega-cap listing. Anthropic confirmed in May 2026 that it signed an agreement to use all compute capacity at SpaceX's Colossus 1 data centre, covering more than 300 megawatts and over 220,000 NVIDIA GPUs, per Anthropic's announcement. The financial terms of that compute agreement were not disclosed by Anthropic. The xAI Colossus facility will appear as a material contract in the SpaceX S-1, linking the AI compute market directly to SpaceX's IPO story.

Read more: Anthropic higher limits and SpaceX compute agreement (Anthropic, May 6, 2026).

GitHub Copilot code review adds MCP server support and a medium-depth analysis tier in public preview

GitHub published a changelog update on June 2, 2026 shipping two public previews for Copilot code review, per the GitHub changelog.

The first preview introduces agent skills and MCP server connections. Teams can create a .github/skills/code-review/SKILL.md file to define a review persona, and they can add MCP server configurations under repository settings to pull context from Jira, Confluence, service catalogs, or incident tooling into every review. Any MCP configuration already set up for the Copilot cloud agent carries over to code review automatically.

The second preview introduces a Medium analysis tier alongside the existing Low tier. Admins set the tier per repository. Medium routes pull requests to a higher-reasoning model for complex logic, security-sensitive code, and cross-service changes. Low remains the default for straightforward work. Medium consumes more AI Credits than Low, with per-tier cost visibility under usage-based billing. The features are available to existing Copilot Pro, Pro+, Business, and Enterprise users. Direct Org Billing lets non-Copilot users access code review on the same terms.

Try GitHub Copilot at /go/copilot.

Read more: Shape Copilot code review around your team.

EU AI Act enforcement for high-risk systems begins August 2, with Annex III deferred to December 2027

August 2, 2026 is 55 days away. That date triggers enforcement of the EU AI Act's core obligations for providers of high-risk AI systems under Articles 9 through 17, and deployer obligations under Article 26, per the EU AI Act Service Desk timeline. General-purpose AI model transparency rules under Article 50 and the EU AI Office's enforcement powers also activate on August 2.

The law covers any company selling AI products or services to European customers, or using AI to make decisions affecting European users, across employment, healthcare, education, financial services, and critical infrastructure. Large foundation models trained above the 10^25 FLOPs threshold face additional systemic-risk evaluation requirements.

The scope was narrowed in May 2026. EU lawmakers reached a provisional agreement to defer Annex III high-risk AI obligations by 16 months, from August 2, 2026 to December 2, 2027. Annex III covers AI used in hiring, healthcare triage, credit decisions, and similar automated high-stakes applications. The provisional agreement still needs formal European Parliament and Council adoption before taking legal effect. Law firm analyses from Holland and Knight and Orrick, both published in 2025-2026, note that the deferral is widely expected to pass but is not yet final.

Read more: Timeline for the Implementation of the EU AI Act.

Gemini 3.5 Pro has not shipped yet; Google targets June with a 2M-token context window and Deep Think mode

At Google I/O on May 19, 2026, CEO Sundar Pichai said Gemini 3.5 Pro was "coming next month," per the Google I/O Gemini announcement. As of June 8, the model had not shipped. Only Gemini 3.5 Flash launched at I/O.

Expected specifications, per I/O sessions and TechTimes reporting from June 6: a 2-million-token context window (double Flash's 1M), a Deep Think extended reasoning mode for the hardest chain-of-thought tasks, and frontier multimodal capabilities. Pricing is expected to follow the Flash-to-Pro ratio of prior Gemini generations. TechTimes cited a figure of approximately $15 per million input tokens and $60 per million output tokens, based on that historical ratio. Google has not confirmed pricing.

The Gemini API changelog at ai.google.dev, which listed Flash on launch morning, shows Gemini 3.5 Pro as not yet released. Google plans to make Pro available first to its $20 Pro and $250 Ultra consumer subscribers, with Ultra also receiving the Deep Think mode. If Pro ships on the June timeline, it competes directly with Claude Opus 4.8 and GPT-5.5 at the frontier reasoning tier.

Read more: Gemini 3.5 models at Google I/O.

Sources