Tech Intelligence Copy RSS link Link copied to clipboard!
Meta shares fall 10% as investors question AI capital expenditure strategy
Meta Platforms shares dropped 10% following first quarter results despite healthy revenue growth of 33%. The decline was driven by significant capital expenditure concerns, with CFO Susan Li admitting the company underestimated computing power needs for AI and facing rising data center component costs. While competitors like Alphabet, Microsoft, and Amazon saw share gains due to clear cloud backlogs and immediate revenue from AI, Meta lacks a booming cloud division. CEO Mark Zuckerberg provided no precise financial roadmap for AI returns, citing a long-term monetisation strategy similar to the failed Metaverse initiative, leading to investor skepticism regarding the timing and clarity of AI payoffs.
Five CEOs lead convergence of AI demand and critical mineral supply in 2026
As of April 2026, Kathleen Quirk (Freeport-McMoRan), Tim Gitzel (Cameco), Robert Friedland (Ivanhoe Mines), Jakob Stausholm (Rio Tinto), and Mark Bristow (Barrick Gold) are identified as key executives driving the intersection of artificial intelligence infrastructure and energy supply. The article highlights the surge in demand for copper and uranium to support data centers and nuclear power, noting specific strategies employed by these leaders to address supply deficits and meet the electrification needs of the tech sector.
Leonardo acquires Iveco Defence Vehicles and ASTRA brands
Leonardo has acquired Iveco Defence Vehicles (IDV) and the ASTRA brands from the Iveco Group. This transaction strengthens Leonardo's leadership in the European land defence sector by combining its electronic systems expertise with IDV's vehicle design and production capabilities. The acquisition enables Leonardo to offer integrated, turnkey military platforms featuring advanced digital architectures. This move consolidates control over the industrial value chain and supports the company's broader strategy for global security and multi-domain integration.
Threat actors abuse Hugging Face and ClawHub for malware distribution
Acronis reports that threat actors are exploiting Hugging Face and ClawHub to distribute malware via trojanized shared files. Attackers use social engineering and indirect prompt injection to trick users into downloading malicious code targeting Windows, macOS, Linux, and Android systems. On ClawHub, nearly 600 malicious skills were identified across 13 developer accounts, including payloads like AMOS. Similar campaigns on Hugging Face host malicious files designed to install infostealers and trojans, leveraging user trust in these AI distribution platforms.
SonicWall patches three SonicOS flaws in Gen 6, 7 and 8 firewalls
SonicWall released urgent firmware updates to address three vulnerabilities in SonicOS affecting Gen 6, 7, and 8 firewalls. The flaws, rated high and medium severity, could allow attackers to bypass security controls, access restricted services, or crash devices. One vulnerability is an improper access control issue, while others involve path traversal and buffer overflow. Affected appliances running specific firmware versions are urged to apply fixes immediately or restrict management access to SSH only. No evidence of exploitation in the wild exists at this time.
Anthropic faces conflicting government directives over Mythos cybersecurity model access
The US White House opposed expanding access to Anthropic's Mythos model to 120 organisations due to security and compute concerns, while simultaneously drafting an executive action to allow federal agencies to bypass Pentagon supply chain restrictions. The Pentagon previously designated Anthropic a national security risk after the company refused to permit autonomous weapons use. Meanwhile, China launched a campaign against AI misuse, and OpenAI released a broader cybersecurity model. Anthropic, valuing over $900 billion, must resolve these geopolitical tensions before its potential IPO.
Severe Linux kernel vulnerability allows attackers to gain root access
Security firm Theori, assisted by AI Xint Code, discovered a high-severity Linux kernel vulnerability named Copy Fail (CVE-2026-31431). The flaw, affecting distributions since 2017, enables local privilege escalation to root access. Researchers reported the issue on 23 March 2026, with patches committed to the mainline kernel by 1 April. The vulnerability impacts web servers, cloud platforms, and container clusters, particularly where untrusted code execution is possible.
Samsung revenue jumps 43% in Q1 2026 amid AI boom and memory shortage warning
Samsung Electronics reported a 43% revenue increase in Q1 2026, driven by strong demand for AI chips and record earnings. The company warned that the global memory chip shortage could persist until 2027 due to rising AI and data centre requirements. Samsung officially discontinued production of LPDDR4 memory to focus on advanced technologies like LPDDR5 and HBM. Growth was led by the semiconductor division, while the MX and Networks division showed stability through flagship smartphone sales and 5G infrastructure demand.
KnowBe4 research finds 86% of phishing attacks are AI driven
KnowBe4 released its Phishing Threat Trends Report Volume Seven, revealing that 86% of phishing attacks are now AI-driven. The report highlights a 49% increase in calendar invite phishing and a 139% surge in reverse proxy usage to steal Microsoft 365 credentials. Jack Chapman, SVP of Threat Intelligence at KnowBe4, noted a shift towards multi-channel orchestration and targeted social engineering, including internal team impersonation. The findings indicate a seismic shift in attack vectors beyond traditional email, impacting collaboration tools like Microsoft Teams.
Encryption breaking technology is now 20x cheaper and CEOs should be very worried
Recent research from Google and Caltech indicates that resources required to break traditional cryptography are materially lower than previously estimated, increasing the risk to internet security and cryptocurrencies. Experts warn that practical quantum-enabled attacks are no longer theoretical, prompting organizations like Google to accelerate quantum-secure transition timelines to 2029. The primary challenge is execution risk, as migrating to post-quantum algorithms across heterogeneous enterprise environments is complex and time-consuming. Leaders are urged to establish continuous cryptographic visibility and adopt crypto-agility to reduce exposure before regulatory deadlines.
Big Tech AI spending reaches $725 billion despite record profits
Big Tech companies Meta, Alphabet, Microsoft, and Amazon have collectively committed to $725 billion in AI infrastructure spending for 2026, representing a 77% increase from 2025. Although Meta reported record quarterly profits, its stock fell over 6% after the company raised full-year capital expenditure guidance to $125-145 billion. Markets remain concerned about capital misallocation and the risk that infrastructure buildout will outpace near-term monetisation, with returns on this massive investment not yet visible on income statements.
OpenAI faces security risks from human network rather than ChatGPT model
OpenAI's ChatGPT, accessed over 300 million times weekly since early 2026, faces significant security vulnerabilities not within the model itself but within the surrounding human network. The article argues that developers, partners, employees, and end-users create threats through misconduct, errors, and attacks. Consequently, OpenAI's automatic security measures require a complete system redesign to address these human-centric weaknesses.
UK NCSC warns organisations to prepare for AI-driven patch wave
On May 1, 2026, the UK National Cyber Security Centre (NCSC) warned organisations to prepare for a surge of software updates driven by artificial intelligence. CTO Ollie Whitehouse stated that AI will exploit technical debt at scale, forcing a correction of vulnerabilities across open source, commercial, and SaaS solutions. The NCSC recommended enabling automatic secure hot patching, activating automatic updates for embedded devices, and adopting an update-by-default policy. The agency also republished its Vulnerability Management guidance version 2.1, emphasising asset identification and risk ownership.
White House intensifies efforts to engage tech companies on AI cyber defense
The White House has intensified efforts to engage technology companies on defending against AI-driven cyberattacks as part of a broader strategic shift announced in March 2026. The administration aims to deploy comprehensive cyber operations and incentivize private sector involvement to address threats from adversaries using AI for sophisticated cyber operations. This initiative reflects a proactive stance on AI-enabled defense mechanisms within federal strategies.
Taiwan first quarter GDP surges 13.69% to 39-year high driven by AI exports
Taiwan's National Development Council announced on 30 April 2026 that first quarter GDP growth reached 13.69%, a 39-year high. This figure significantly outperformed China, Singapore, and South Korea. The growth was primarily driven by a 51.10% increase in exports, fueled by global demand for AI servers and chips, alongside increased capital investment and consumer spending.
Security risks challenge Model Context Protocol adoption in production environments
Security researchers warn that the Model Context Protocol (MCP), designed to simplify AI agent integrations, introduces significant risks including tool poisoning, prompt injection, and arbitrary file access. While MCP offers convenience for connecting AI assistants to external tools, these features expand the attack surface, making it unsuitable as a default for day-to-day coding tasks. Experts recommend safer native alternatives such as project instructions, skills, slash commands, and direct APIs to maintain control and reduce exposure in production software teams.
Australia data centre boom faces power test
Australia's data centre industry faces scrutiny over a widening gap between announced projects and likely operational capacity by 2030. While operational IT capacity is projected to reach 3.1GW to 3.2GW, many proposals remain early stage or dependent on power and water approvals. New South Wales and Victoria account for the bulk of the pipeline, with energy constraints and local impacts becoming central points of debate. Major commitments, such as Microsoft's A$25 billion investment, reinforce growth, but independent forecasting highlights risks of overbuilding infrastructure for uncompleted projects.
Taiwan's ASE expects strong demand to boost advanced chip packaging sales in 2026
Taiwan's ASE Technology Holding expects revenue from its leading-edge advanced packaging business to rise 10% to over $3.5 billion in 2026, driven by strong customer demand for AI chips. The company raised capital expenditures by $1.5 billion for buildings, infrastructure, and machinery to support services in 2026 and 2027. ASE reported first-quarter revenue of $5.50 billion, up 17.2%, with net income increasing 87.3%. Shares rose 95% this year. The company broke ground on a new chip testing campus in Kaohsiung with an investment exceeding $3.43 billion, with operations starting in 2027.
SkyfireAI raises $11M for autonomous drone coordination
SkyfireAI, a US-based startup focused on AI-powered drone autonomy, has secured $11 million in seed funding led by Mucker Capital. The company aims to develop a software platform enabling multiple drones to operate autonomously in real-world, mission-critical environments for first responders, law enforcement, and defense teams. Founded by veterans from the US Navy, intelligence community, and DARPA, the firm seeks to scale drone operations without increasing human pilot requirements. The funding will accelerate product development and team growth to address regulatory and staffing limitations in the sector.
Wall Street analyst sets $1,000 price target for Micron
DA Davidson analyst Gil Luria initiated coverage of Micron Technology with a buy rating and a $1,000 price target, nearly double the current trading price. Luria argues that the AI infrastructure build-out will extend the memory chip cycle, driven by sustained demand for high bandwidth memory (HBM) for graphics processing units. He projects Micron could generate $393 billion in revenue by fiscal 2030. Despite the stock falling on the day of the report, the analyst believes the company's strategic shift to five-year HBM deals and low valuation multiples present a significant opportunity.