RobinsonLabs/ news

SpaceX's Colossus data center racks up $80B in AI compute deals as Reflection AI signs $6.3B lease

SpaceX has become the world's largest commercial AI compute provider, with its Colossus infrastructure in Memphis now hosting clients including Google, Anthropic, and new open-source lab Reflection AI on contracts worth more than $80 billion through 2029.

SpaceX has signed a computing power agreement with Reflection AI, an open-source AI startup valued at $25 billion. Under the deal, Reflection will pay $150 million per month beginning July 1, 2026, for immediate access to Nvidia's latest GB300 chips at the Colossus 2 facility. The contract, worth up to $6.3 billion through 2029, can be terminated by either party with 90 days' notice after the first three months. The deal follows similar massive contracts with Google ($920 million per month) and Anthropic ($1.25 billion per month), marking a strategic pivot for SpaceX from using Colossus solely for its own Grok AI chatbot to becoming a major revenue-generating compute infrastructure provider. The shift has drawn mixed investor reaction, with SpaceX stock dropping 16.4% as markets weigh the compute rental model against the company's ambitions to build its own AI products.

OpenAI releases GPT-5.5-Cyber and "Patch the Planet" initiative to tackle open-source vulnerability debt

OpenAI has released an upgraded GPT-5.5-Cyber model to vetted cybersecurity organizations under its Daybreak initiative, describing it as the strongest model yet for finding and helping patch software vulnerabilities at scale.

The model scored a record CyberGym score of 85.6 percent, outperforming the standard GPT-5.5 model's 81.8 percent, with an ExploitGym score of 39.5 percent. It can navigate large codebases, trace attack paths, validate exploitability, and generate patches in a single workflow. Alongside the model, OpenAI launched "Patch the Planet" in partnership with Trail of Bits, targeting critical open-source projects including cURL, Go, Python, and Python.org. In its initial five-day sprint, the program scanned over 30 million commits and identified findings across the Linux kernel, OpenBSD, FreeBSD, dnsmasq, and major browsers including Chrome, Safari, and Firefox. The initiative's emphasis on mandatory human review before submissions reflects OpenAI's awareness that flooding under-resourced maintainers with unvetted AI-generated reports could worsen rather than solve the open-source security crisis.

Nobel laureate John Jumper joins Anthropic in a double blow to Google DeepMind's talent base

Google DeepMind has lost two of its most senior AI researchers in quick succession, with Nobel laureate John Jumper departing for Anthropic and Gemini co-lead Noam Shazeer leaving for OpenAI, triggering a 7.2 percent intraday drop in Alphabet stock.

Jumper, who co-created AlphaFold and shared the 2024 Nobel Prize in Chemistry with DeepMind CEO Demis Hassabis, announced his departure on X after nearly nine years at Google. AlphaFold has predicted over 200 million protein structures, fundamentally transforming biological and medical research. Jumper was also a key member of Google's team developing AI coding tools. His move came just days after Shazeer, a vice president of engineering and co-author of the Transformer architecture that underpins modern AI, announced he was joining the IPO-bound OpenAI. Bloomberg described the simultaneous departures as a "one-two punch" of foundational AI talent leaving Google at an unprecedented pace, underscoring the intensifying talent war between Big Tech and frontier AI startups.

White House issues executive order promoting AI innovation while tightening cybersecurity oversight

President Trump signed Executive Order 14409 on June 2, 2026, directing federal agencies to accelerate AI cybersecurity capabilities while explicitly rejecting mandatory government licensing or pre-clearance requirements for AI model development and deployment.

The order, titled "Promoting Advanced Artificial Intelligence Innovation and Security," establishes a voluntary framework for "secure frontier model deployment" that would give the government up to 30 days of early access to designated models before they are released to trusted partners — a compromise from earlier drafts proposing a 90-day window. Key directives include forming an AI cybersecurity clearinghouse within 30 days, developing a classified benchmarking process led by the NSA to assess models for advanced cyber capabilities, and directing DHS to ensure AI-enabled cybersecurity tools are available to federal agencies, state and local authorities, and critical infrastructure operators including rural hospitals and community banks. The order also directs the Attorney General to prioritize enforcement against individuals using AI to illegally access computer systems or steal data, leveraging existing statutes including the Computer Fraud and Abuse Act. All industry collaboration under the order is voluntary, reflecting the administration's innovation-first philosophy of partnering with developers rather than regulating them.

ServiceNow cuts thousands of jobs in sweeping June restructuring

The company laid off hundreds of employees across sales, engineering, marketing, training, and product teams, citing AI-driven efficiencies and acquisition-related headcount offsets.

On June 10, 2026, ServiceNow executed an abrupt, large-scale workforce reduction at its Santa Clara headquarters and San Diego office, with California WARN filings confirming significant cuts. Sources confirmed a three-figure number of employees were let go, though some reports estimate the total may reach up to 2,500 when acquisition offsets from Moveworks (~$3B, 600+ staff), Armis (1,300+ staff), and Veza are factored in. The cuts spanned every major function — sales, solution consulting, marketing, engineering, training, product, CEG, and ServiceNow University — and affected both long-tenured employees and recent hires, with no clear selection logic.

ServiceNow told Salesforce Ben that the company is "managing headcount with discipline" and that its platform is "generating real AI efficiencies inside our own business." Employees described a highly disruptive process: HR joined routine calls to announce eliminations, and laptops were locked within minutes. The layoffs have raised concerns about cultural erosion and the impact of recent executive transitions, though ServiceNow experience remains in high demand in the partner ecosystem.

Inspira Enterprise becomes full-spectrum ServiceNow delivery partner with AICT deployment

The cybersecurity and AI services firm has expanded its partnership to cover the entire ServiceNow platform, leveraging AI Control Tower, Veza identity governance, and Armis asset intelligence.

Announced on June 22, 2026, the strategic partnership positions Inspira Enterprise as a trusted full-spectrum delivery partner for enterprises worldwide. The company deployed ServiceNow's AI Control Tower (AICT) as its enterprise-wide governance layer, onboarding more than 50 AI agents across identity and access management, threat and vulnerability management, cyber defense, risk management, and internal operations. The deployment delivered a 40% increase in AI adoption and a 35% improvement in operational productivity.

Inspira's ICNOW Center of Excellence offers 200+ ready-to-deploy use cases across six industries, including pre-built cybersecurity accelerators that reduce implementation effort by 60–70%. The partnership combines Veza (identity governance) and Armis (asset intelligence) capabilities with AICT to provide unified IT, security, and AI governance outcomes. "Inspira's unique position—as both an AICT customer and a full-spectrum ServiceNow delivery partner—means enterprises now have a trusted partner to help them govern, implement, and scale their AI investments," said Adrian Johnston, president of APAC at ServiceNow.

AI Control Tower June 2026 release delivers automated discovery, governance, and regulatory readiness

The ServiceNow platform update generally available with the Australia release advances the Discover, Govern, and Secure pillars with new integrations and pre-built compliance content.

The June 2026 release of AI Control Tower brings automated rules for managing assets at scale, expanded Service Graph Connector integrations with Databricks, Snowflake, and Hugging Face for AI asset discovery, and simplified credential management for AWS and Azure. A new Microsoft Agent 365 integration allows managed ServiceNow agents to be published directly to the Agent 365 directory from the AI Control Tower asset record. MCP (Model Context Protocol) servers are now a governed asset type with lifecycle tracking and approval workflows enforced in AI Agent Studio.

On the governance side, the release introduces pre-built compliance content packs for the California AI Act, Colorado AI Act, and EU AI Act, with controls that map across multiple frameworks so organizations can test once and demonstrate compliance across jurisdictions. The AI Gateway now offers automated PII detection with a single toggle to scan and block payloads containing sensitive data. Asset management enhancements include formal offboarding workflows, a structured Use and Purpose field for accurate risk classification, and updated APIs for external integrations.

Wall Street consensus holds ServiceNow as a Strong Buy, though Zacks flags a Hold rank

82% of 45 brokerages rate the stock as Strong Buy, but earnings estimate stability keeps the Zacks Rank at Hold.

As of June 23, 2026, ServiceNow commands a 1.33 average brokerage recommendation on a 1-to-5 scale (Strong Buy to Strong Sell) based on 45 firms — 37 of which (82.2%) rate it Strong Buy and 3 rate it Buy. Despite the bullish analyst sentiment, the Zacks Rank sits at #3 (Hold) because the consensus earnings estimate for the current year has remained unchanged at $4.13 over the past month, indicating analysts see the stock performing in line with the broader market rather than outperforming in the near term. The Zacks model weights earnings estimate revisions over static opinions, suggesting limited near-term upside momentum despite the overwhelmingly positive brokerage consensus.