CAIA

Global Competition & Governance Geopolitics and policy remained prominent, with China’s AI ecosystem surpassing 700 officially filed generative models, highlighting both the speed and scale of deployment in its domestic market. In parallel, US and UK policymakers continued to refine national AI strategies and regulatory blueprints, aiming to balance innovation, economic growth and safeguards on issues such as biometric surveillance and systemic model risk. The result is a more fragmented but assertive global governance environment in which companies must navigate differing rules while still racing to industrialise AI across sectors from healthcare to financial services. #AIPolicy #ChinaAI #USAI #UKAI #AIRegulation

Chips, Infrastructure & Industry This week also underscored that AI progress is now tightly coupled to hardware and infrastructure economics, from Furiosa’s AI accelerator entering mass production as a challenger to Nvidia, to ongoing “inference wars” among chipmakers and cloud providers. Strategic partnerships and large‑scale deployments, including new data‑centre build‑outs and robotaxi pilots using advanced in‑car assistants, signal that AI capability is increasingly defined by end‑to‑end systems rather than isolated models. For enterprises, the message is clear: competitive advantage in 2026 will hinge on access to performant, cost‑efficient AI infrastructure as much as on model choice alone. #AIInfra #GPUs #Cloud #Robotics #AIatScale

AI news, weekly roundup. A thread 🧵 #AI #News #Summary

Risks and constraints 2026 Even with a more positive trajectory, 2026 construction output faces material constraints. Skills shortages, high input costs, and tight margins continue to weigh on contractors’ ability to scale up delivery, while global supply chain volatility still poses a risk to programme certainty. Policy execution will also be critical: delays in planning reform, uncertainty over public capital spending, or weaker‑than‑expected economic growth could all drag on the sector, turning a forecasted gentle upswing in 2026 into a flatter, stop‑start year for UK construction. #ConstructionRisk #SkillsGap #PlanningReform #SupplyChain #MacroTrends #Construction2026

Key growth drivers 2026 In 2026, new work in infrastructure, energy transition and major regeneration schemes is expected to be the main engine of construction output growth, supported by long‑term capital commitments and a growing pipeline of transport, utilities, and digital infrastructure projects. Residential activity is also projected to improve as financing conditions stabilise and housing policy reforms begin to filter through, although delivery will still depend heavily on planning capacity, build‑cost inflation, and buyer confidence. #Infrastructure #Housing #EnergyTransition #Regeneration #ConstructionMarket #UKIndustry

Is the UK’s construction output going to increase in 2025? A thread 🧵 #UK #Construction #Output

Intel: The third major GPU/AI accelerator supplier Intel rounds out the top three, supplying AI accelerators via its Gaudi line and discrete data centre GPUs, positioning itself as a cost‑efficient alternative for large‑scale AI training and inference clusters. While its share is far smaller than NVIDIA’s and AMD’s today, Intel’s Gaudi3 is specifically optimised for large AI workloads and is being deployed in cloud environments that want open standards and tight integration with existing x86 infrastructure. Together, NVIDIA, AMD and Intel form the core trio of companies manufacturing the GPUs and accelerators that power most of today’s large‑scale AI systems, even as custom silicon from hyperscalers and specialist startups begins to nibble at the edges of the market. #AI #Intel #Gaudi #GPUs #AIAccelerators #DataCenter #Cloud #Chipmakers

AMD: The fast‑rising challenger AMD is the primary challenger to NVIDIA’s dominance, with its Instinct MI300 series (and successor MI400/MI450 roadmaps) targeting the same high‑end training and inference workloads in hyperscale data centres. The company is leaning on aggressive packaging and high‑bandwidth memory to offer competitive performance and, in many cases, better price‑per‑watt, making it attractive to cloud providers looking to diversify away from a single‑vendor stack. Analysts now view AMD as one of three core players in the AI accelerator arms race, with potential for tens of billions in annual AI revenue later this decade as more cloud and enterprise buyers adopt multi‑GPU vendor strategies. #AI #AMD #Instinct #GPUs #AIChips #Hyperscale #CloudComputing #Semiconductors

What 3 companies make the most GPU’s for powering AI? A thread 🧵 #AI #GPU #Nvidia #Intel #AMD

By 2026, AI’s macro impact will increasingly show up in productivity and GDP figures as much as in product launches. Major asset managers expect AI‑related capital spending to remain a key support for growth in 2026, with AI investment’s contribution to US growth estimated at roughly three times recent historical averages. At the same time, benchmark studies highlight an “execution gap”: while most firms have pilots, only a small minority (single‑digit percentages) have scaled advanced agentic or autonomous AI into production. The growth story for 2026, then, is less about new breakthroughs and more about who can standardise AI governance, close that execution gap, and convert infrastructure spending into durable productivity gains. #AIProductivity #MacroTrends #AgenticAI #Governance #AIEconomy #FutureOfWork

On the revenue side, forecasts suggest the global AI market could reach roughly the low‑$300B range by 2026, with multiple studies pointing to CAGRs north of 20–30% from mid‑2020s baselines. This growth is being pulled by enterprise adoption of generative AI in core workflows—marketing, software engineering, legal drafting and product design—as companies move from chatbots to embedded copilots and domain‑specific models. Survey data from large enterprises shows more than 70% already using gen AI, with over two‑thirds planning budget increases and the majority now formally measuring ROI and seeing positive returns. #GenAI #EnterpriseAI #AIMarket #Productivity #AIROI #Strategy2026

AI’s forecasted growth in 2026. A thread 🧵 #AI #Growth #Forecast

Regulation and governance caught up with AI’s systemic impact in 2025. – The EU AI Act entered into force with a risk-based regime, strict obligations for high‑risk systems, bans on unacceptable uses and upcoming rules for general‑purpose and systemic models from August 2025. – The US pivoted toward a more innovation‑focused federal stance, while the UK advanced a principles‑based, regulator‑led model, forcing global companies to navigate a complex, multi‑jurisdictional compliance landscape. #AIRegulation #EUAIAct #Governance #AIEthics #Policy

Frontier models in 2025 pushed reasoning, multimodality and science into new territory. – New generations of models improved long-context reasoning, code generation and multimodal understanding across text, image and video, enabling more capable agents and real-time workflows. – AI systems accelerated scientific discovery, from protein structure and genomics to oncology and theoretical computer science, cementing AI’s role as a “co-scientist” in R&D. #FrontierModels #MultimodalAI #AIResearch #LifeSciences #RAndD

AI in 2025, summarised. A thread 🧵 #AI #Summary #Headlines