CapabilitiesMaturityApproachDiagnosticAI & Tech ReviewContact
Weekly Review

AI & Tech Review

Emerging developments across AI and technology.

Week of April 25 to May 1, 2026
This is an archived edition. View the latest review →
Lead Story
Canada & Enterprise
CanadaEnterpriseAI

Federal Spring Economic Update References Six Pillars of National AI Strategy and the Existing Sovereign AI Compute Strategy

The Spring Economic Update tabled by Finance Minister François Philippe Champagne on April 28 referenced the six pillars of Canada's forthcoming national AI strategy, previously unveiled in a speech by Minister of Artificial Intelligence and Digital Innovation Evan Solomon. The update also referenced the existing Canadian Sovereign AI Compute Strategy, which is built on three complementary pillars: mobilizing private sector investment, building public supercomputing infrastructure, and establishing the AI Compute Access Fund. The call for applications under the AI Sovereign Compute Infrastructure Program launched in mid April. The package was tabled alongside a small and medium business procurement program intended to widen federal contract access.

  • Spring Economic Update tabled by Finance Minister Champagne on April 28 referenced the six AI strategy pillars previously unveiled by AI Minister Solomon
  • Existing Canadian Sovereign AI Compute Strategy spans private investment, public supercomputing infrastructure, and an AI Compute Access Fund; applications under the AI Sovereign Compute Infrastructure Program opened in mid April
  • Small and medium business procurement program tabled alongside, intended to widen federal contract access

Enterprise Impact: The Spring Economic Update is the clearest federal signal yet on the direction of national AI policy. Enterprises planning AI investment in Canada should sequence against BDC LIFT, the eventual full strategy text, and the existing sovereign compute access path. Vendors selling AI platforms, infrastructure, or services into Canada should expect "safe and sovereign" framing to surface in procurement conversations and supplier qualification. Boards should treat the volume and pace of federal AI activity as a directional commitment that will harden through 2026 and 2027.

Source: Government of Canada
AI Models & Platforms
EnterpriseAIInfrastructure

Big Tech Q1 2026 Earnings: Alphabet and Amazon Pull Ahead on AI, Combined Capex on Track for $700 Billion

Amazon, Meta, Alphabet, and Microsoft reported Q1 2026 earnings after the close on April 29, with Apple following on April 30. Alphabet’s Google Cloud revenue grew 63% year over year, well ahead of the 47% consensus, with the cloud backlog nearly doubling to $460 billion. Amazon Web Services grew 28% year over year to over $37 billion in revenue, posting its largest sequential quarterly increase ever. Combined 2026 capital expenditure plans across Alphabet, Amazon, Microsoft, and Meta now exceed $700 billion, nearly double 2025 spend and $100 billion above prior guidance.

  • Google Cloud Q1 revenue up 63% year over year, backlog nearly doubled to $460B
  • AWS Q1 revenue up 28% year over year to over $37B, largest sequential gain on record
  • Combined 2026 capex across Alphabet, Amazon, Microsoft, Meta tracking above $700B

Enterprise Impact: The quarter cements that hyperscaler capacity and pricing pressure will continue through 2026 and into 2027. Multi year cloud commitments under negotiation should include capacity reservations, price escalation caps, and named SKU continuity terms for AI workloads. Architecture teams should assume that frontier model capability is consolidating to providers with the deepest AI infrastructure spend, and plan accordingly for both vendor selection and contingency. Canadian buyers should pair platform decisions with explicit residency, lawful access, and data egress language.

Source: Bloomberg
EnterpriseAI

Microsoft Copilot Studio Multi Agent Orchestration Reaches General Availability Alongside April Microsoft 365 Copilot Release

Microsoft Copilot Studio multi agent orchestration reached general availability for all eligible customers in the April release wave, alongside connected experiences and faster prompt iteration improvements. The parallel Microsoft 365 Copilot April release added a Claude model option in Word, image model selection in PowerPoint across GPT Image, Flux, and an Auto setting, Copilot Chat in Teams chats, channels, calling, and meetings, and Copilot Notebooks gains across SharePoint and OneNote referencing, Word and PowerPoint generation from notebook content, and Microsoft 365 Group sharing in the Frontier preview.

  • Multi agent orchestration in Copilot Studio reaches general availability for all eligible customers
  • Microsoft 365 Copilot April release adds Claude model option in Word and image model selection in PowerPoint
  • Copilot Chat extends to Teams chats, channels, calling, and meetings with mobile support coming
  • Copilot Notebooks gains SharePoint and OneNote referencing plus Word and PowerPoint output

Enterprise Impact: Enterprise IT and AI governance teams should re scope Copilot rollouts to include the new agent orchestration capability and surface area. ISO/IEC 42001 aligned governance should now formally cover multi agent designs, including agent identity, intent traceability, and decision logging. DLP and information protection policies should be reviewed for the new image generation surface and for cross Teams Copilot Chat data flows. Procurement should validate license entitlements against the new feature set before renewal cycles open.

Source: Microsoft Copilot Studio Blog
EnterpriseAIInfrastructure

Google Commits up to $40 Billion to Anthropic, Locks in 5 Gigawatts of TPU Capacity Coming Online From 2027

Google announced on April 24 a commitment of up to $40 billion to Anthropic. The structure is $10 billion in cash now at Anthropic's existing $350 billion valuation set in February, with up to $30 billion more contingent on performance milestones. Google Cloud is locking in 5 gigawatts of TPU capacity for Anthropic, with capacity coming online from 2027. The deal lands days after Amazon expanded its own commitment to as much as $25 billion. Anthropic's annualized revenue has scaled from approximately $9 billion at the end of 2025 to $30 billion as of April 2026.

  • $10B cash now at Anthropic's existing $350B valuation, with up to $30B more tied to performance milestones
  • 5 gigawatts of TPU capacity from Google Cloud, coming online from 2027
  • Anthropic annualized revenue scaled from approximately $9B at end of 2025 to $30B by April 2026
  • Lands alongside Amazon's expanded commitment of up to $25B disclosed days earlier

Enterprise Impact: Frontier AI is now being financed like utility infrastructure rather than software, with hyperscalers locking in compute for equity arrangements with the labs they also distribute. For enterprise buyers, model availability, pricing, and continuity will increasingly be shaped by which hyperscaler stack they sit on. Architecture teams should reassess multi cloud and multi model strategies against the new reality and request explicit model continuity terms, capacity reservations, and price protection clauses in any AI contract under negotiation.

Source: Bloomberg
EnterpriseAI

Anthropic Ships Nine Claude Creative Connectors, Embedding Frontier AI Directly Into Adobe, Autodesk, Blender, Ableton, SketchUp, Resolume

Anthropic released nine Claude Creative Connectors on April 28, embedding the model directly inside widely deployed creative software through the Model Context Protocol. The full list: Adobe Creative Cloud, Autodesk Fusion, Blender, Ableton Live and Push, Splice, Affinity by Canva, SketchUp, Resolume Arena, and Resolume Wire. All nine connectors are available immediately across all Claude plans. Rather than copy data between apps, users invoke Claude directly inside the host tool to generate, modify, batch process, or query documentation.

  • Nine Claude Creative Connectors shipped April 28 across Adobe, Autodesk, Blender, Ableton, Splice, Affinity, SketchUp, Resolume Arena, Resolume Wire
  • Built on Model Context Protocol; available immediately across all Claude plans
  • Pattern signals frontier AI moving from chat surface into embedded production tooling

Enterprise Impact: Information protection, DLP, and licence management policies should be reviewed for embedded AI patterns now, not after creative or design teams have adopted at scale. Architecture teams should treat embedded model integrations as a distinct category in AI system inventories, separate from standalone chat or API deployments. Procurement should request explicit data residency, training data, and incident notification commitments where Claude plans are bought through indirect channels. Boards should request evidence that AI governance covers integrations into host applications, not only standalone deployments.

Source: Anthropic
Regulation & Infrastructure
EnterpriseAIRegulation

EU AI Act Digital Omnibus Trilogue Ends Without Agreement, August 2026 Deadline Remains Operative

The second political trilogue on the Digital Omnibus between the European Parliament, Council of the EU, and European Commission ended on April 28 without agreement. A further trilogue is scheduled for May 13. The Commission proposal published in November 2025 seeks to defer the high risk AI Act compliance deadline from August 2 2026 to December 2 2027. Until the agreement is published in the Official Journal, the August 2 2026 deadline remains the only operative legal deadline. Member States must still establish at least one national AI regulatory sandbox by August 2 2026.

  • Second trilogue April 28 ended without agreement, third scheduled for May 13
  • August 2 2026 high risk compliance deadline remains operative until agreement is published
  • National AI regulatory sandboxes still required in each Member State by August 2 2026

Enterprise Impact: Enterprises with EU exposure should plan against the August 2 2026 deadline and treat any deferral as upside, not baseline. Anchor AI risk management, post market monitoring, and event logging architectures to ISO/IEC 42001 so that EU specific obligations become a configurable overlay. Customer due diligence on AI governance posture is intensifying regardless of trilogue outcome. Boards should request a single status view of AI regime exposure across EU, Canadian, and US obligations rather than tracking each separately.

Source: European Commission
EnterpriseCybersecurityInfrastructure

Itron 8-K Discloses Intrusion in Internal Corporate IT, Customer Hosted Systems Reported Unaffected at Initial Disclosure

Itron, a major supplier of smart meters, grid edge devices, and utility software to electricity, gas, and water utilities, filed an SEC Form 8-K on April 24 disclosing that intruders had been present in its internal corporate IT systems before being expelled. The company stated at initial disclosure that it had not identified unauthorized activity in the customer hosted portion of its systems, and that operations continued in all material respects. News outlets picked up the disclosure on April 27. The intrusion sits inside a Q1 2026 regional threat picture in which Cyble's Americas Threat Landscape Report recorded 1,138 publicly claimed ransomware attacks across the Americas.

  • Itron 8-K filed April 24 disclosed intrusion in internal corporate IT systems; no unauthorized activity identified in customer hosted portion at initial disclosure
  • News outlet coverage followed on April 27; OT environments were not the reported scope of the intrusion
  • Cyble Americas Threat Landscape Report recorded 1,138 publicly claimed ransomware attacks in Q1 2026 across the Americas

Enterprise Impact: Utilities and any organization with Itron components in the stack should request a structured incident briefing under existing contractual rights and update third party risk records accordingly. The reported scope is corporate IT, not Itron's grid edge or meter operating environments, but vendor incidents at this layer can evolve as investigation progresses, and contractual visibility into subsequent disclosures is rarely automatic. ISO/IEC 27001 Annex A 5.21, 5.22, and 5.23 supply chain and cloud security controls apply. Boards should request evidence that vendor compromise scenarios are exercised, including supply chain variants in tabletop runs.

Source: TechCrunch
View all insights
Previous Reviews

Get the Weekly Briefing

AI & technology analysis delivered to your inbox every Monday. No noise, no hype — just what matters.