Cloud Enable examines the options available to IT departments considering adopting or managing cloud-based architectures in their business environment.
• Tencent continues to advance its Hy AI model, focusing on cost-efficiency, performance, and specialization around productivity tasks.
• With Tencent WorkBuddy, Tencent is helping enterprises leverage AI to raise workforce productivity to benefit from the AI technology.
Tencent Cloud has been advancing its AI solutions despite not grabbing headlines internationally. The company is preparing the release of Tencent Hy4, the fourth generation of the Tencent Hy (Hunyuan) AI model. Tencent Hy4 is a 770 billion parameter, open-source, mixture-of-expert (MoE) model with 49 billion active parameters per token (i.e., optimizing for speed and compute resources by activating only a subset of the full model parameters). It is designed to improve reasoning efficiency and to support productivity tasks and visual understanding. The company has also designed Tencent Hy4 to support its AI products and applications by bringing expertise across Tencent including software engineering, gaming, finance, and security. As cost is one of the major concerns for enterprises, Tencent Cloud will also be adding Tencent Hy4 Flash, a more cost-effective model that carries Tencent Hy4’s visual understanding and agentic capabilities while reducing computational overhead. This is also responding to the broader industry trend toward a two-tier strategy, with a deep reasoning and a highly optimized, low-latency version (e.g., Gemini 3.8 Flash, GPT-4o mini, Claude 3.5 Haiku, DeepSeek-V4.1-Flash, and Qwen3.8-Flash).
• US investment in fiber is accelerating at an all-time high as AT&T, Zayo, Verizon, and Lumen expand their networks to support AI and data-center growth.
• Building more fiber does not automatically translate into greater profits; customer demand, network utilization and investment discipline will determine the winners.
The US fiber market is in the middle of an AI gold rush, with telecom providers appearing determined that the machines never have to wait for bandwidth. AT&T’s recent partnership with Global Infrastructure Partners (GIP) and CPP Investments adds momentum to an investment race that already includes Zayo, Lumen, and Verizon. Each company is reshaping their networks to capture future demand from AI workloads and data center expansion. The enthusiasm is understandable, but with so many carriers reaching for their shovels, a more fundamental question remains: will AI demand deliver the returns the industry needs or will some investors be left digging for profit?
AT&T’s deal points to a shift in how communication service providers (CSPs) fund the next phase of their network expansion. By partnering with investment and private equity firms, AT&T can combine external funds with its existing fiber assets, accelerating deployment while sharing the financial risk. The venture’s wholesale, open-access model also widens the potential customer base to include operators seeking to support enterprise connectivity, data center interconnections, and AI workloads. The arrangement carries a trade-off, however. External funding eases pressure on AT&T’s balance sheet and speeds up the build, but AT&T must share future revenue and profit with its investment partners.
AI has now become the industry’s preferred justification for fiber spending, and every major carrier is investing in fiber networks to support the next phase of digital infrastructure growth. The expansion of hyperscale data centers and AI computing facilities is increasing the demand for those high-bandwidth connections between data centers, cloud platforms, and enterprise locations, which makes low latency and network resilience more important. Capacity alone will not guarantee success. Carriers must identify the routes and markets with the greatest potential, secure customer commitments, and maximize utilization to generate sustainable returns. AI is accelerating the investment cycle, but commercial discipline will determine which providers turn infrastructure spending into profitable growth.
As investment in US fiber networks continues to gain pace, providers must ensure that spending is backed by real customer demand rather than expectations of future growth. With AT&T, Verizon, Zayo, and Lumen all expanding their networks, competition will also grow across key markets, putting pressure on revenues and profitability. Partnerships like these can help spread investment costs and attract additional customers, but success will come down to how effectively CSPs use their networks. Routes that connect to established data center hubs may offer more immediate opportunities, while investments in emerging markets may bring greater uncertainty. Carriers must balance opportunity to capture growth against the risk of building capacity which may take years to generate adequate revenue.
AI is creating a significant opportunity for the US fiber market, but scale and timing of the future demand is less certain. As these four expand their networks, CSPs must look beyond growth and focus on profitable returns. AT&T’s partnership highlights one way to fund network expansion, while sharing the risk, however all carriers must have the customer demand to support long-term revenue growth. With competition increasing, more fiber will not automatically translate into greater profits. That leaves the question of whether success will come to those with the largest networks, or those making the smartest investment decisions.
• Technology vendors are investing in forward deployed engineers to adapt AI solutions to meet customer needs.
• Systems integrators are working alongside their technology partners to help enterprises accelerate their AI adoption and ensure business outcomes are met.
It is widely reported that only a small percentage of firms that have conducted AI pilots are moving them into production. Even fewer enterprises have been able to scale their AI initiatives across multiple business functions. Based on GlobalData’s observations and discussions with enterprises, there has been greater success in AI implementations among large enterprises that have engaged a systems integrator or consulting firm to work alongside their technology partners, including hyperscale cloud providers and AI model companies (e.g., Anthropic and OpenAI).
• Agentic orchestration has been dominating vendor narratives as growing complexity in the CX space is driving the need for a sophisticated coordination system.
• CX agentic orchestration remains fragmented as there are still infrastructural gaps, and trust remains a major issue for customers.
Agentic orchestration has been increasingly occupying CX vendor narratives in recent years, so it is time to take a deep dive into what orchestration is and why it is becoming relevant. Simply put, orchestration is bringing people, workflows, and technology together to achieve an outcome. Depending on how far we stretch the concept, orchestration is not new, and even a deterministic call routing can be described as orchestration. The question is, what is new today and why is orchestration increasingly dominating vendor narratives?
• New fears of AI breaches and vulnerabilities significantly threaten AI adoption
• Salesforce and rivals launch aggressive control plane portfolios as part of the new layer of the AI stack
On the heels of broad industry doomsday reports of AIs going rogue, Salesforce launched its mega Dreamforce conference in San Francisco, California (US) last week. CEO Marc Benioff used the stage to reposition the company as the industry platform to provide the oversight and governance of both Salesforce and third-party agents under a single view.
• New features from Zoom reinforce its recently reconstituted platform vision, extend its ongoing transformation, and support an important, contemporary trend in the team collaboration space.
• This latest batch of improvements underscores Zoom’s place among the top team collaboration and contact center vendors such as Microsoft, Cisco, and Google.
Earlier this month, Zoom held the latest installment of an ongoing series of customer webinars highlighting new platform features. In keeping with tradition at Zoom, the enhancements were wide-ranging, covering various tracts of its platform.
• Slackforce Surfaces are front doors into data and context that enable users to grasp a deeper understanding of their business and make more informed decisions.
• Slackforce Surfaces represent the latest chapter in the renaissance of Slackbot, advancing Slack as a key part of the Salesforce organization.
Salesforce has deepened its investment in Slackbot, the AI-driven personal work agent built into Slack, with the debut of Slackforce Surfaces. The introduction will raise the utility of the Slack platform and further cement the starring role Slack has come to play at Salesforce.
• Zayo will build 8,000 miles of new long-haul fiber across key AI corridors, with Nvidia becoming its anchor customer.
• Nvidia’s extends beyond GPUs and compute, with partnerships spanning the optical and networking ecosystem underpinning AI infrastructure.
Until now, the AI Infrastructure race has predominately been focused on GPUs, data centers and access to reliable power, yet beneath all three sits a less visible, but increasingly critical, layer – connectivity. As AI workloads become larger, more distributed and dependent on moving large volumes of data between locations, fiber is emerging as a fundamental component of the AI Stack. Zayo’s recent announcement to build 8,000 miles of new long-haul fiber across key AI corridors, backed by Nvidia as its anchor customer, could mark the new beginning of a new investment cycle for terrestrial networks. The bigger question is whether AI-related companies could become the anchor customer needed for the next generation of fiber investment?
• Zoom CX is a credible option for enterprises looking to transform their customer engagement with omni-channel and AI capabilities.
• Working with a broader partner ecosystem is pivotal for Zoom to win in CX space since this involves workflows and different business applications.
Zoom is well-known for its conferencing solution, which is used extensively in modern workplace, but it has gone well-beyond conferencing in recent years. The core business of Zoom has been the enabler of conversations within the workplace. To go beyond communications, the company sees new opportunities by expanding its role to help enterprises automate workflows during and after conversations (i.e., meetings and phone conversations). This vastly enlarges the value Zoom can deliver to enterprise customers, especially with the application of AI. For example, AI can eliminate many manual tasks such as generating documents from meeting with summaries and next steps, or updating CRM records after a discussion within the sales team.
• AI is drastically changing the fabric of the traditional data center, prompting fundamental changes in design and architecture.
• The biggest challenge is that AI infrastructure requires simultaneous scaling across multiple constrained layers: electricity, cooling, networking, chips, facilities, capital, and operations.
The rise of AI workloads is pushing data centers through a major architectural shift: from relatively general-purpose, virtualized compute environments toward high-density, network-intensive AI infrastructure. For example, rack density is rising sharply, because traditional data centers were not designed for the power and thermal profiles of dense AI server clusters. This means power distribution, floor loading, cable management, and thermal design are becoming central architectural considerations. Power availability has now become a core design constraint. Energy availability is starting to influence where data centers are built, with land and power constraints pushing some infrastructure development into new or remote regions.
You must be logged in to post a comment.