Loading story…
Reporting focused on AI infrastructure scale-up and compute constraints: OpenAI CFO Sarah Friar outlined an infrastructure roadmap that ties revenue growth directly to increased compute capacity, describing available compute rising from ~0.2 GW (2023) to ~0.6 GW (2024) and projecting ~1.9 GW by 2026 alongside projected ARR exceeding $20B. Separately, Microsoft job postings indicate continued datacenter expansion in APAC, including hiring “Senior Energy Program Managers” to secure power interconnections and energy supply contracts for regional datacenters—underscoring the operational reality that AI datacenters are increasingly power-limited.
Other items in the set are adjacent but not the same story: one report covers Elon Musk’s lawsuit against OpenAI/Microsoft and a California Attorney General cease-and-desist aimed at xAI’s Grok over alleged generation of non-consensual deepfake sexual imagery, while another describes Musk’s claimed roadmap for Tesla AI processors on a nine-month design cadence (AI5/AI6 and beyond) and the automotive safety-certification constraints that can slow deployment. A separate market-research piece describes China’s lead in humanoid robot deployments (nearly 13,000 of ~16,000 installed globally in 2025), which is AI-adjacent but not directly tied to the infrastructure/power and compute-roadmap thread.
Mallory correlates global threat intelligence with your attack surface — know if you’re exposed before adversaries strike.
4 events from the most recent confirmed update back to the earliest known activity.
Microsoft began recruiting senior energy and engineering roles in Australia and Singapore to secure power and infrastructure for its expanding Asia-Pacific datacenter footprint. The hiring reflects growing electricity and capacity demands tied to AI datacenters.
Australia’s eSafety Commissioner said regulated technology companies had removed access to 4.7 million social media accounts belonging to users under 16 under the country’s youth social media ban. The regulator said it was still too early to assess the policy’s full impact.
Within 12 days of deployment, the University of Dhaka said its honeynet recorded more than 63,000 attacks from thousands of unique IP addresses. Many of the attempts used default or commonly used credentials, indicating broad opportunistic scanning and intrusion activity.
Bangladesh’s University of Dhaka deployed a new honeynet sensor to observe malicious activity targeting exposed systems. The sensor began attracting hostile traffic almost immediately after going live.
Vulnerabilities, threat actors, malware, products, organizations, and breaches Mallory has linked to this story.
2 references tracked. Mallory keeps watching after this page renders.
Map indicators from this story to your assets and identify affected systems in minutes.
Every observed campaign, victim, and pivot linked to actors named in this story.
Malware, exploits, and IOCs connected to the activity described here.
YARA, Sigma, and Snort rules deployed to your SIEM as soon as they’re published.
Get matching new stories delivered to your team as they break — not the next morning.
Ask questions about this story and take action on the answers.
Other stories Mallory is tracking that overlap on entities, sources, or topic labels.
Reporting and commentary highlighted intensifying **U.S.–China competition in AI** driven less by capital and more by access to advanced compute and the ability to shape global AI governance. In China, a wave of Hong Kong IPOs raising **more than $1B** for domestic AI firms was framed as a confidence signal, but industry leaders warned that funding alone cannot close the gap with leading Western labs; Alibaba *Qwen* leadership reportedly assessed China’s odds of “leapfrogging” **OpenAI** and **Anthropic** via fundamental breakthroughs as **below 20%**, citing structural constraints such as compute availability and ecosystem maturity. Separately, policy analysis argued China is expanding international influence through **AI capacity-building diplomacy**, including a **UN General Assembly resolution** on AI capacity-building (co-sponsored by 140+ countries) and initiatives like training workshops, governance action plans, and infrastructure support aimed at the Global South—while warning the U.S. risks ceding agenda-setting power if it cannot sustain consistent engagement. A third piece captured **Nvidia CEO Jensen Huang** publicly pushing back on “doomer” narratives and the idea of imminent “god AI,” emphasizing current systems’ limits; while not a cybersecurity incident, it reinforces the broader theme that near-term AI outcomes are constrained by practical factors (capability limits and compute), not hype alone.
Jun 29, 2026
Recent coverage focused on **AI adoption, governance, and societal impacts** rather than a discrete cybersecurity incident. OpenAI CEO **Sam Altman** argued that comparing AI energy use to human cognition is “unfair,” claiming the energy cost of “training a human” (years of living and food consumption plus evolutionary history) should be considered when judging AI efficiency, and separately warned that some companies are engaging in **“AI washing”**—attributing layoffs to AI as a pretext for workforce reductions—while also acknowledging real job displacement is likely to become more noticeable in the next few years. Enterprises and public-sector organizations highlighted practical AI rollouts and associated risk considerations. **Intel** introduced *Ask Intel*, a support assistant built on **Microsoft Copilot Studio**, alongside a shift away from public phone support toward web-based case handling, while noting response accuracy “cannot be guaranteed.” **Microsoft** removed a blog post that had described training LLMs using a Kaggle dataset derived from **pirated Harry Potter ebooks**, amid ongoing legal uncertainty around fair use and potential contributory infringement exposure. Separately, U.S. federal officials emphasized **targeted AI adoption** and expectation management (with the VA reporting hundreds of AI use cases), while other items included a hobbyist AI dashboard project shared on GitHub and a generic startup article on AI-accelerated MVP development—neither of which provided substantive security-relevant disclosures.
Jun 29, 2026
Reports said **Anthropic** was finalizing a roughly **$20B funding round** that would value the company at about **$350B**, with **Microsoft and NVIDIA** cited as major sources of capital alongside a broader syndicate of large investors. The coverage framed Microsoft’s participation as a strategic hedge given its longstanding alignment with **OpenAI**, and tied the valuation narrative to Anthropic’s recent product momentum (notably *coding agents* and specialized domain models) amid escalating compute costs. Separately, analysis of Big Tech financial guidance and filings described a sharp step-up in **AI infrastructure CapEx** for 2026, led by **Amazon** disclosing **$200B** in planned capital expenditures and warning via an SEC filing that it may need to raise equity and debt to fund build-out. The same piece cited **Alphabet ($175–$185B)**, **Meta ($115–$135B)**, and **Microsoft (~$145B run rate)**, putting the “Big Four” at roughly **$635–$665B** and arguing the industry is restructuring internet-scale infrastructure around AI workloads, with capital intensity reaching historically high levels (e.g., Oracle and Microsoft cited as elevated vs. revenue).
Jun 29, 2026