Tech News

Stay updated with the latest tech news that shapes our world. Discover trends, innovations, and insights in the tech industry.

A new, sophisticated phishing campaign is exploiting Microsoft’s own infrastructure—Active Directory Federation Services (ADFS) and trusted office.com redirects—to harvest Microsoft 365 credentials. Malicious actors deploy malvertising, leading users from legitimate ads (e.g. for “Office 365”) through trusted Microsoft domains like outlook.office.com, before redirecting them to attacker-controlled phishing sites. By leveraging ADFS configurations within a valid Microsoft tenant, these redirects appear authentic to both users and security filters, enabling credential theft and even bypassing MFA protections. Security researchers urge organizations to closely monitor ADFS redirect chains, analyze Google ad parameters directing to office.com, and deploy enterprise-wide ad blockers as part of effective risk mitigation.
Sources:
Bleeping Computer
,
Computing.co.uk
,
Cyber Security News
Key Takeaways
– The attack—dubbed “ADFSjacking”—abuses Microsoft’s ADFS infrastructure and valid redirects to masquerade phishing pages as authentic, thereby evading standard security tools.
– Conditional loading and intermediary domains (like faux travel blogs) further mask the redirection chain, returning non-target users to legitimate sites.
– Mitigation measures include monitoring for unusual ADFS redirects, inspecting Google ad traffic parameters for office.com, and deploying ad blockers and behavioral detection systems.
In-Depth
In a troubling escalation of cyber threats, attackers are now weaponizing Microsoft’s own identity infrastructure to probe and steal user credentials without drawing suspicion. Known as “ADFSjacking,” this strategy harnesses active directory federation services (ADFS) and trusted office.com redirects to construct a near-perfect phishing environment.
Users clicking on seemingly legitimate ads for Office 365 are funneled through outlook.office.com—giving the impression that everything is safe—before being funneled to cleverly masked credential-stealing pages. These redirects are made possible through attacker-controlled Microsoft tenants that orchestrate the redirection flow, manipulating ADFS to appear benign to security systems and users alike.
The sophistication is further evidenced by conditional loading: only the intended victim sees the phishing page, while others are harmlessly bounced back to the real Microsoft site. This stealth approach also taps into reverse-proxy techniques and token theft to sidestep multi-factor authentication.
Organizations must recalibrate their defenses accordingly: look for ADFS redirect chains targeting unknown domains, inspect Google ad-to-office.com traffic for suspicious parameters, enforce enterprise ad-blocking measures, and shift toward behavior-based detection tools that identify anomalies rather than relying solely on URL filtering. In the game of phishing, leveraging the trust of infrastructure yields powerful rewards for attackers—but informed vigilance can still level the playing field.

Microsoft has entered into a landmark agreement with Nebius Group N.V., an Amsterdam-based AI infrastructure provider, to lease GPU infrastructure capacity from Nebius’s forthcoming Vineland, New Jersey data center. The contract is valued at $17.4 billion over five years, with an option to expand it up to $19.4 billion depending on Microsoft’s future compute needs. Nebius plans to fund the build-out of the data center and GPU procurement through cash flow from the deal and debt secured against the contract. The announcement triggered a sharp rally in Nebius’s share price, underscoring investor confidence in the expanding AI infrastructure sector.
Sources:
Reuters
,
Financial Times
,
Investors.com
Key Takeaways
– Massive Scale and Strategic Growth: The deal underscores the intensifying race for AI compute capacity, spotlighting Nebius as a rising “neocloud” provider challenging traditional hyperscalers.
– Financial Confidence: Nebius will finance its expansion through long-term contract cash flow and debt backed by the agreement, signaling strong financial backing and planning.
– Market Validation: The stock market’s enthusiastic response—decimal-point jumps in Nebius’s share price—reflects investor belief in both the company’s potential and broader demand in AI infrastructure.
In-Depth
Microsoft’s latest multiyear agreement with Nebius marks a significant milestone in the escalating AI infrastructure arms race. At its core, the deal locks in $17.4 billion worth of GPU leasing over five years, with the potential to reach nearly $19.4 billion if demand continues to surge. The GPUs will come from Nebius’s upcoming data center in Vineland, New Jersey—an operation that Nebius plans to bankroll through the deal’s cash flow and debt backed against it. This careful financial structuring shows a pragmatic approach to growth, balancing opportunity with cautious leverage.
Nebius, initially spun off from Yandex’s international operations, has rapidly evolved into a full-stack AI infrastructure provider with assets and capabilities sprouting across Europe and the U.S. This deal not only provides Microsoft vital compute access—especially as GPU demand outpaces supply—but also positions Nebius as a credible alternative to large cloud incumbents. Investor reaction was immediate and overwhelming, with Nebius shares jumping sharply by up to 68% in after-hours trading, depending on the outlet cited. This rally reflects both confidence in Nebius’s execution and faith in broader AI infrastructure demand.
From a layman’s standpoint, Microsoft’s expansion of vendor relationships beyond traditional hyperscalers like Azure or AWS introduces healthy competition and resilience in its AI ecosystem. For Nebius, the partnership delivers both cash flow and credibility, validating its strategic direction. All told, the agreement exemplifies how enterprise-scale players are doubling down on purpose-built AI infrastructure—baking future demand into today’s build-out strategy.

Microsoft has unveiled a sweeping $30 billion investment strategy in the UK over the next four years, dividing the funds equally between capital expenditures (notably infrastructure and supercomputing) and expanding its UK operations. Half of the sum—about $15 billion—is earmarked for building up AI infrastructure, including the UK’s largest supercomputer, which will be built in partnership with British cloud company Nscale and powered by over 23,000 advanced GPUs. The other $15 billion will be used to grow Microsoft’s operational footprint in the UK—hiring, local development, and supporting developer ecosystems among others. The announcement comes amid improved regulatory tides in the UK and coincides with U.S.–UK diplomatic engagements, signaling a new chapter in technological cooperation.
Sources:
Reuters
,
MLQ
,
Ground News
Key Takeaways
– Microsoft is making one of its largest single-country AI investments ever: $30 billion in the UK through 2028, split evenly between infrastructure/supercomputing and operational expansion.
– A headline project is the development of the UK’s largest supercomputer, in collaboration with Nscale, using over 23,000 GPUs, aiming to significantly boost AI compute capacity in Britain.
– The investment reflects growing confidence in the UK’s regulatory environment, shifting business climate, and the importance of U.S.–UK cooperation in high technology and AI during a period of mounting global competition.
In-Depth
Microsoft’s newly announced $30 billion UK investment is a signal that the company is doubling down on its commitment to making the UK a central node in global artificial intelligence infrastructure. Over the next four years, Microsoft plans to pour half of that sum into developing physical compute capabilities—supercomputers, data centres, advanced hardware like GPUs—and the rest into growing its operational base in the UK, which includes hiring, partnerships, local innovation ecosystems, and supporting developers. The core infrastructural piece is the planned supercomputer in collaboration with Nscale, which with over 23,000 advanced GPUs should mark a major step forward in the compute capacity available for AI research, enterprise adoption, academic uses, and more.
On the regulatory and political front, this investment does not come in isolation. It follows the UK’s regulatory approval of Microsoft’s acquisition of Activision Blizzard in 2023, which had previously been held up by competition concerns. The improved business climate appears to have encouraged Microsoft’s leadership to believe that large-scale AI hardware and infrastructure investment is now more feasible. The timing—coinciding with state-level diplomatic engagements between the U.S. and UK—underscores broader strategic ambitions: both in economic growth and in technological sovereignty and leadership.
However, this kind of scale brings challenges. Building and powering a supercomputer with 23,000 plus GPUs is no small feat: energy supply and cost, cooling, site selection, and regulatory permissions (especially around data, environmental, planning laws) are likely to be substantial hurdles. Microsoft will have to ensure return on investment by making sure the investment translates into usable compute capacity, serves universities, startups, and industry, not just internal consumption. Also, with global tech competition heating up, this move by Microsoft may provoke or precipitate similar announcements or counter-measures from competitors, both in the UK and abroad.
If everything goes according to plan, though, the UK could emerge as a more formidable player in global AI, with stronger infrastructure, a deeper talent base, and more rapid AI adoption across sectors. The question going forward will be whether the promised infrastructure is delivered on time, runs efficiently, and is accessible in ways that benefit more than just large corporations.