IT Connection

AI Wars Intensify via Major LLM/Agentic Releases

June 29, 2026June 29, 2026 by Charlotte Dunlap, posted in Big Data

Summary Bullets:

• Cycles between advanced AI model rollouts are significantly shortened among leaders in this space

• Developers are gaining access to agentic-injected integrated development environments (IDEs); while knowledge workers gain access to agentic AI assistants.

The second quarter marks a momentous period in the industry’s ongoing AI efforts. Platform leaders shipped next-generation agentic runtimes including autonomous and other advanced capabilities, all while managing a more compressed cycle of new AI models, which are rolling out in a matter of weeks versus months.

A few notable announcements highlight this structural shift in how enterprise AI is built, deployed, and presented to enterprises.

Microsoft’s long-awaited private review of its first in-house reasoning model, MAI-Thinking-1, an enterprise-grade medium-weight model that promises to shake up the industry in a number of ways. Microsoft is going up against the industry’s strongest models based on the strength of its mathematical and scientific reasoning abilities, for improved training loops, citing numerous Microsoft-backed engineering benchmark tests. It is taking on Claude Sonnet 4.6 and Opus 4.6 by claiming lower token costs and smaller inference footprint. For the first time since the beginning of its relationship with OpenAI, Microsoft is able to break into the enterprise space with its own AI model, on par with leading rivals. Microsoft’s win will inevitably be at the expense of OpenAI.

Expanding its AI portfolio further was the June release of Microsoft Copilot Studio – Computer Use, revamping AI assistants to perform further up the agentic AI stack. The release supports the use of computer-use agents directly in Copilot Studio, helping bypass integrations with APIs in order to develop workflow automations.

To keep pace with top rivals Google and Anthropic, OpenAI announced its biggest model release yet, GPT-5.5, emphasizing its strengths in agentic coding, scientific research, and the ability to automate tasks associated with knowledge work. As the industry’s early GenAI leader, OpenAI has been challenged to maintain its innovative prowess.

OpenAI’s newest advancements are mere weeks following its last GPT release, demonstrating the staggering breakneck pace AI model providers are compelled to maintain to keep up in this highly competitive segment. OpenAI is hoping to win back the loyalty of professional coders who have moved to Anthropic Claude in droves for its accuracy in coding. OpenAI has been most popular among consumers, while competitors, including Google and Anthropic, have gained more traction in the enterprise space.

AWS’s latest AI announcements demonstrate a deliberate pivot towards agentic AI amidst an increasingly competitive landscape. Under mounting competitive pressure, Amazon is investing heavily in tools that span developer and non-developer audiences.

The newly announced Amazon Quick agentic AI assistant is a revamp of the GenAI assistant Q Business platform, providing knowledge-based workers with insights while also being able to act and automate repetitive workflows. Quick connects internal data across AWS services, third-party platforms, and on-premises systems. Other key announcements were Kiro agentic IDE built on Code OSS and powered by Claude models, via Amazon Bedrock; and Bedrock AgentCore, serverless runtime, and AgentCore Harness, which let developers build and run production-grade AI agents quickly without needing to code custom orchestration loops.

UK Plans Teen Social Media Ban, but the Action Raises Questions About Enforceability – and Privacy

June 26, 2026June 26, 2026 by Amy Larsen DeCarlo, posted in Communicate

Amy Larsen DeCarlo – Principal Analyst, Security and Data Center Services

Summary Bullets

• Following the lead of other nations including Australia, the UK is getting set to restrict access to social media sites for minors 16 years and younger, starting in 2027.

• Comparable rules in other countries have proven to be difficult to execute, with teenagers finding workarounds. However, UK government officials say their efforts will leverage highly effective technology to enforce the ban, including biometric facial age estimation.

When UK Prime Minister Keir Starmer announced plans to introduce legislation later this year that will bar teenagers and children from social media access, critics offered immediate comparisons to a similar action by Australia in 2025, which has largely been deemed a failure. Though initially Australia touted the fact that 4.7 million accounts held by children under 16 had lost access to social media platforms like TikTok and Snapchat, research shows that just months later, the ban had very little effect. A University of New Castle study of 408 12 to 17 year olds found that due to “limited implementation, incomplete compliance, and substantial circumvention of social media restrictions, the ban has been largely unsuccessful”

UK officials countered criticism with claims that the legislation will require strict age enforcement controls which the platform vendors themselves will be responsible to execute. These include sophisticated and potentially controversial age verification technology. Technology Secretary Liz Kendall directed Ofcom, the UK government-approved regulatory authority for the broadcasting, internet, telecommunications and postal entities, to conduct an evaluation by October of what accurate systems can be used to validate a user is over 16. Possible solutions include biometric face-age estimation or third-party identity checks.

Educators and researchers have long raised red flags about the potential harm social media poses to children and teenagers. Research from the US Surgeon General and the American Psychological Association reported that minors spending more than three hours a day on social media experience double the depression and anxiety of those who don’t. Other issues linked to social media use in children include poor body image, sleep deprivation, and even structural impact on brain development in areas like impulse control.

As alarming as the byproducts of social media can be on anyone, the almost draconian efforts to thwart access introduce some other serious concerns. Once biometrics are introduced en masse, what controls will be in place to ensure there is no manipulation or other misuse of the content? And at what point is the boundary set between what is and isn’t social media? With many news and other information sites incorporating forum content, will students be prevented from accessing this as well?

However, with as many as nine of ten UK parents approving of the ban, it is likely to be carried forward. Only time will tell how inventive UK children will be in circumnavigating around social media restrictions.

Google Cloud Highlights the ROI of Agentic Workloads During UK Summit

June 26, 2026June 26, 2026 by Beatriz Valle, posted in Communicate

Summary Bullets:

• The Google Cloud UK Summit was held in June 2026, at Tobacco Docks, in London (England). The company will release Gemini 3.5 Pro and Gemini Omni, a multi-modal system, in late-2026.

• Google announced the UK availability of Gemini 3.5 Flash, through the Gemini Enterprise Agent Platform and the Gemini Enterprise app.

Google Cloud held its UK Summit in London with the strategic intent of demonstrating the ROI of agentic AI. With more than 4,000 participants, it followed the thread commenced in 2025 (please see Google Cloud Focuses on Agentic AI During UK Summit, July 15, 2025), only this time the industry has moved decisively towards the implementation of the technology in practical cases. The company showcased its considerable presence in the UK, with a new data center that will be opening in Waltham Cross, Hertfordshire, later in 2026 as part of a GBP5-billion investment in AI programs in the country.

Continue reading →

HCLTech Hones Its Application Development Practice, Reflecting the Disruptive Impact of AI

June 12, 2026June 26, 2026 by Rena Bhattacharyya, posted in Cloud Enable

A woman with long, wavy dark hair and a warm smile, wearing a gray blazer and a silver necklace, posed against a neutral background. — R. Bhattacharyya

Summary Bullets:

• AI has disrupted traditional developer teams and tasks, and new processes and talent will be required to responsibly implement the intelligent automation and probabilistic nature of agentic systems.

• As enterprises drive towards a mature application landscape that is built using AI and for AI-infused applications, intelligent orchestration and integration are critical.

Although AI offers the promise of greater efficiency across a myriad of enterprise workstreams, one of the use cases with the greatest benefit is application modernization. GenAI’s effectiveness in writing and refactoring code has already been highly touted in mainstream media; less known is its use in other aspects of the software development lifecycle (SDLC). It can be used for discovery, documentation, quality assurance, autonomous testing, intelligent orchestration, and other tasks as well. Furthermore, AI is doing much more than accelerating application development; it is changing how software is engineered. Intelligence and analytics are no longer add-ons that are layered onto existing applications. Today’s applications have intelligence embedded into their workflows and decision logic, essentially creating modern apps that are designed to be AI-first.

Continue reading →

Why Quantum-Safe Connectivity Could Become the Next Differentiator for Wholesale Providers

June 9, 2026June 9, 2026 by Brendan Swan, posted in Cloud Enable, Secure

B. Swan

Summary Bullets:

Telecom Italia’s Sparkle has launched its quantum-safe interconnect (QSI) with Equinix across 20 International Business Exchange (IBX) positioned throughout Europe, the Americas, and Asia.
With the rise of quantum computing, it has highlighted the importance of cybersecurity, with growing concerns around future attacks where sensitive data could be compromised – store now, decrypt later.

The wholesale telecommunications segment has traditionally competed on scale, reach, latency and price. However, as enterprises accelerate the adoption of AI, hybrid cloud, and internationally distributed workloads, another factor is rapidly moving up the priority list: Security. With cyber threats becoming increasingly sophisticated and quantum computing edging closer to becoming able to break traditional encryption, quantum-safe connectivity is emerging as the next major differentiator for wholesale operators seeking to move beyond price-led competition. CSPs that can combine their global reach, low latency, and post-quantum security will be positioned to capture a greater share of the enterprise and hyperscaler demand. The broader question now emerging for the industry is whether security could soon become just as important as scale, reach, and price in wholesale telecommunications.

Continue reading →

Remote Desktops/VDI – A Persistent Bad Idea

May 29, 2026May 29, 2026 by Steven J. Schuchart Jr., posted in IT Leader

S. Schuchart

Summary Bullets:

• The nature of the problems changed from local to network, requiring expensive network support rather than using PC technicians.

• It’s important to remember that many concepts in modern enterprise IT come up again and again cyclically, only to be characterized as ‘paradigm-changing’ by enthusiastic marketing teams.

The idea of remote desktops or streaming an employee’s Microsoft Windows desktop to a more affordable device or even in a window on a remote employee’s self-owned hardware has been around for a long time. The reasons were simple – less spending on desktop computers, easier support both on-site and particularly for remote, better security, and remote desktops worked particularly well for temporary or project-based access for outsiders. Citrix and VMware both made hay with remote desktops, especially in the era where PC hardware and Microsoft Windows itself were considerably less reliable than today.

Continue reading →

FinOps Takes on the AI Explosion, Including Token Management

May 29, 2026May 29, 2026 by Charlotte Dunlap, posted in Cloud Enable

A smiling woman with shoulder-length blonde hair, wearing a white blouse, against a softly blurred background. — C. Dunlap
Research Director

Summary Bullets:

FinOps X conference takes place in San Diego, California (US) June 9-11, 2026.
Key themes will include how enterprises will operationalize AI-driven FinOps across platform engineering.

FinOps X conference in San Diego will take place in one week, and not surprisingly AI will dominate keynotes and discussions among FinOps practitioners. These experts will share insights into best practices for operationalizing AI-driven FinOps across platform engineering, including CICD, Kubernetes, and other cloud-native architectures.

Continue reading →

Verizon DBIR: Adversaries Weaponize AI in Stealth Attacks by Targeting Points of Exposure

May 22, 2026May 29, 2026 by Amy Larsen DeCarlo, posted in Secure

A close-up portrait of a woman with light brown hair, wearing a black blazer and a light-colored turtleneck sweater, smiling softly against a light blue background. — Amy Larsen DeCarlo – Principal Analyst, Security and Data Center Services

Summary Bullets:

Bad actors are raising their intelligence quotient with AI, tapping it to find vulnerabilities faster and to power mobile-centric phishing campaigns.
Supply chains are a weak link with partner network weaknesses linked to nearly half of all breaches.

An already volatile threat landscape is becoming even more dangerous as threat actors tap AI to accelerate and improve the success of their attacks on enterprises. Verizon’s 2026 Data Breach Investigations Report (DBIR) reveals how effective adversaries have become in using AI to capitalize on enterprise weaknesses. Exploiting software vulnerabilities was the initiating factor in 31% of all breaches, notable because this is the first time in almost 20 years that it has overtaken compromised credentials as the most frequent entry point for an attack.

Continue reading →

Zoom Enhancements Help Move Work Forward but Fall Short in Other Areas

May 20, 2026May 20, 2026 by Gregg Willsky, posted in Cloud Enable, Communicate

G. Willsky

Summary Bullets:

Zoom hopes that this latest round of capabilities will help drive greater platform adoption and distinguish itself from rivals.
With respect to achieving those goals, the capabilities deliver Zoom a mixed report card.

Zoom announced a mix of capabilities encompassing mobile access for Zoom My Notes, improved agentic search for Zoom AI Companion and upgrades to Zoom MCP Server. Through the new capabilities, Zoom intends to achieve two implicit goals: to drive platform adoption by adding appealing functionality that aligns with recent market trends and to distinguish itself in a competitively dense field.

Continue reading →