OpenAI ended its exclusive cloud arrangement with Microsoft on April 27. The same week, Google, Microsoft, Meta, and Amazon committed a combined $725 billion in capital expenditure for 2026 — up 77% from last year. The Pentagon cleared six AI companies to operate inside its most classified networks. And the EU quietly delayed its AI Act enforcement timeline by 18 months. A lot changed in the space of seven days.
Google cloud revenue hit $20 billion in Q1, up 63% year over year. Microsoft posted $83 billion in quarterly revenue at 18% growth but free cash flow cratered 22% under the weight of infrastructure spending. The AI economy is producing growth. It is also burning through capital at a pace that has no historical precedent in the technology industry.
1. OpenAI and Microsoft End Exclusivity. The AI Cloud Wars Just Opened Up.
After four years of exclusive partnership, OpenAI and Microsoft restructured their deal on April 27. OpenAI can now offer its models on any cloud platform — AWS, Google Cloud, or anyone else. Microsoft retains a non-exclusive license to OpenAI’s intellectual property through 2032 and stays the default Azure-first partner. Revenue share payments continue through 2030 but are now subject to a cap. The AGI determination clause — which would have required Microsoft to decide how to respond if OpenAI achieved artificial general intelligence — was removed.
For Microsoft, the deal lands alongside eye-watering numbers. Calendar-year 2026 capex is set at $190 billion, well above the $152 billion analyst consensus. CFO Amy Hood attributed $25 billion of that figure directly to rising memory chip costs. Free cash flow fell 22% in Q3. The same week, Microsoft launched its first-ever voluntary severance program, offering buyouts to 8,750 US employees.
The strategic read is straightforward. OpenAI committed to spend $250 billion on Azure services when the original recapitalization closed in October. That anchor spend is still in place. But OpenAI can now go anywhere for distribution. Amazon and Google will both be competing for the same workloads Microsoft previously had locked up. The exclusivity wall is gone.
2. $725 Billion. Big Tech’s AI Capex Is Now Its Own Economic Event.
First-quarter 2026 earnings confirmed the number. Google, Microsoft, Meta, and Amazon are collectively spending $725 billion on capital expenditure this year — up 77% from the $410 billion record set in 2025. Google’s cloud revenue jumped 63% year over year to $20 billion. Brent Thill at Jefferies said the revenue growth justifies the spend. “The bear thesis is garbage.”
The chip constraints behind the number are significant. Nvidia’s Vera Rubin NVL72 racks are priced at up to $8.8 million apiece. HBM memory shortages are expected to persist until 2027. Samsung and SK hynix both warned that AI-driven memory demand is outpacing their production capacity. Half of planned US data center builds have been delayed or cancelled due to power infrastructure shortages and parts constraints tied to Chinese supply chains.
The pressure is landing on chip makers outside the US too. Chinese GPU maker Cambricon posted Q1 revenue of $423 million as domestic demand accelerated behind Nvidia export restrictions. Huawei is positioned to take the lead in China’s AI chip market in 2026 as Nvidia’s H200 shipments stall in regulatory limbo. The $725 billion is not just a spending number. It is a supply chain stress test.
3. The Pentagon Clears Six AI Companies for Its Most Classified Networks.
The US Department of Defense this week cleared Microsoft, Amazon, Google, Nvidia, SpaceX, and OpenAI to deploy AI systems inside its Impact Level 6 and Impact Level 7 environments — the networks that handle classified and top-secret data. LLMs are now authorized for “lawful operational use” inside the military’s most sensitive infrastructure. Palantir’s Maven AI system was separately formalized as a core military platform with multi-year funding.
Google’s move came with a carve-out. The company signed the classified Pentagon AI deal but exited a separate $100 million drone swarm program, drawing a line between classified computing infrastructure and autonomous weapons systems. That distinction matters as generative AI moves deeper into military decision-making.
For the AI companies, the Pentagon clearances represent a new revenue category. Government AI contracts at the classified level carry margins that commercial cloud work cannot match. The infrastructure investment required is enormous. So is the lock-in. Once a military network is running on your stack, switching costs are effectively infinite.
4. AI Layoffs, AI Officers, and the Corporate Reckoning That’s Just Getting Started.
Coinbase announced plans to cut roughly 700 employees while restructuring around an “AI-native” operating model. Sam Altman publicly warned against companies “AI-washing” layoffs — using automation narratives to cover broader cost-cutting. Snap cut 16% of its workforce in the same week and quietly ended its $400 million partnership with Perplexity before a broad rollout ever happened. Meta has disclosed plans to cut 8,000 jobs to fund AI infrastructure.
The organizational response is taking shape in parallel. IBM reported that most large companies are now creating Chief AI Officer roles. McKinsey called AI “the largest organizational shift since the industrial and digital revolutions.” Lloyds Banking Group became the first FTSE-listed blue-chip to deploy an AI tool in its boardroom. Bain estimated SaaS firms could realize nearly $100 billion in margin gains by converting labor costs into software spending through AI automation.
The gap between the companies moving fast and those still planning is widening. AI spending is compressing competitive timelines across every sector. A company that deploys an AI-native workflow this year does not give that advantage back. Coinbase framed its cuts as a move toward efficiency. Whether it is that or a cost play dressed up as strategy will show up in the next two quarters of product output.
5. The EU Blinks on the AI Act. High-Risk Rules Push to 2027.
EU lawmakers and member states reached a provisional agreement this week to dilute and delay several provisions of the EU AI Act. Rules governing high-risk AI systems will now apply from late 2027, an 18-month delay from the August 2026 deadline that had been in force. Certain industrial machinery applications are excluded from the act entirely. Mandatory watermarking for AI-generated content and bans on unauthorized sexually explicit deepfakes are retained.
The US is moving in the opposite direction on structure. The Trump administration published a National Policy Framework for AI pushing for a unified federal approach. In the UK, Technology Secretary Liz Kendall announced plans for a domestic AI hardware strategy covering chips and semiconductors. The House of Commons launched an inquiry into neuromorphic photonics as a potential path to reducing the energy intensity of AI workloads. Written evidence closes May 14.
The regulatory picture is fragmenting. Brussels delayed its hardest requirements. Washington went for federal coordination over state-by-state patchwork. London moved toward domestic hardware sovereignty. For multinational tech companies, the compliance map looks different in every major market. That divergence is not an accident. It is a competitive policy.
What to Watch: Week of May 12-18, 2026
UK neuromorphic photonics inquiry: written evidence deadline is May 14. The inquiry covers low-energy computing as a response to AI’s rising power demands. Any government signal on chip R&D funding commitments will be the first concrete output from Kendall’s hardware strategy speech.
Microsoft voluntary severance results: the buyout offer to 8,750 US employees closes this week. Uptake numbers will signal how much of the workforce restructuring Microsoft plans to follow with involuntary cuts. Watch for any announcement on headcount reductions in Azure infrastructure roles specifically.
Nvidia Vera Rubin NVL72 shipment schedule: at $8.8 million per rack, hyperscalers are locked into multi-billion delivery pipelines. Any delay announcement from Nvidia or commentary from AMD on its MI400 competitive position moves the semiconductor supply chain picture significantly.
EU AI Act implementation guidance: with the high-risk deadline now pushed to late 2027, the European AI Office is expected to publish transitional guidance for companies that had been actively preparing for August 2026. That guidance will determine whether the delay creates a genuine compliance window or just shifts the same preparation burden forward.
Samsung union situation: more than 30,000 Samsung workers are demanding bonuses averaging $400,000 per employee tied to AI profits. A prolonged action would directly affect HBM memory output at a time when every hyperscaler is already supply-constrained on advanced memory.
Sources
- CNBC: OpenAI shakes up partnership with Microsoft, caps revenue share, April 27, 2026
- Ad-Hoc News: Microsoft Pentagon breakthrough and OpenAI reset, the $190 billion cost, May 2026
- Tom’s Hardware: Microsoft and OpenAI end exclusivity agreement, May 2026
- Tom’s Hardware: Big Tech capex to hit $725 billion in 2026, up 77%, May 2026
- Tom’s Hardware: Microsoft attributes $25 billion of AI budget to memory chip costs, May 2026
- Tom’s Hardware: Pentagon announces AI deals with OpenAI, Google, Microsoft, Amazon, Nvidia, May 2026
- MarketingProfs: AI Update May 8, 2026: AI news and views from the past week
- CNBC: Do you need a chief AI officer? How AI is changing boardrooms, May 11, 2026
- TLT LLP: AI Brief May 2026, UK policy and regulatory landscape
- MIT Technology Review: 10 Things That Matter in AI Right Now, 2026
Editorial Disclosure
This roundup is based on a combination of earnings reports, press releases, regulatory announcements, and independent market research sourced from publicly available information. It covers developments during the week of April 28 to May 12, 2026, in the artificial intelligence, technology, and semiconductor sectors. Securities referenced incidentally in this article include Microsoft Corporation (Nasdaq: MSFT), Alphabet Inc. (Nasdaq: GOOGL), Meta Platforms Inc. (Nasdaq: META), Amazon.com Inc. (Nasdaq: AMZN), Nvidia Corporation (Nasdaq: NVDA), Coinbase Global Inc. (Nasdaq: COIN), and Palantir Technologies Inc. (Nasdaq: PLTR). None of these companies compensated aktiego.com for coverage. OpenAI is a privately held company referenced on the basis of publicly announced partnership agreements and regulatory filings. Financial performance data is sourced to Q1 2026 earnings releases and is current as of the date of publication. Forward-looking statements attributed to named analysts, executives, or institutions represent opinions at the time of publication and are not guarantees of future performance or regulatory outcomes. EU AI Act regulatory timelines are current as of May 12, 2026, and are subject to formal adoption procedures. Editorial forward-look commentary reflects the author’s own assessment. The information provided on this website is for informational and educational purposes only. Our content is derived strictly from verified online sources to ensure accuracy and objectivity. This analysis does not constitute financial, investment, or professional advice. Readers are encouraged to consult with qualified professionals before making decisions based on this information. For more information, please see our full DISCLAIMER.
