AI Reality Check — Issue 18

AI Reality Check - Issue 18

Friday 24 July 2026 | Kaye Nicholson | GrowthZone AI | growthzoneai.co.uk

Born analogue. Raised digital. 30 years of real business experience. Now explaining what AI actually means for work.

━━━━━━━━━━━━━━━━━━━━━━

1. Apple's Three-Front Week

What happened: Apple overtook Nvidia to become the world's most valuable company, with a market valuation of 4.88 trillion dollars. In the same period, Apple is in talks with the US Department of Justice to settle its 2024 antitrust lawsuit over an alleged smartphone monopoly, with multiple offers reportedly made this year. And Apple is simultaneously suing OpenAI, alleging that former Apple employees stole confidential hardware secrets to accelerate OpenAI's push into consumer devices.

Why it matters: three completely different kinds of pressure, all playing out for one company at the same time. Financial dominance at record levels. Government negotiation over antitrust concerns. Direct legal conflict with a fast-moving competitor. It is a useful snapshot of how complex operating at the top of the technology industry has become.

Who it's for: anyone tracking the competitive dynamics of big tech. Businesses evaluating long-term bets on Apple's ecosystem or OpenAI's products. Anyone interested in how antitrust regulation is evolving for the biggest technology companies.

━���━━━━━━━━━━━━━━━━━━━━

2. Grok 4.5 and GPT-5.6 Launched a Day Apart

What happened: xAI launched Grok 4.5, an Opus-class coding model priced at 2 dollars per million input tokens. Less than 24 hours later, OpenAI's GPT-5.6 had its full public launch, following the preview covered in earlier issues.

Why it matters: two of the biggest competitors in AI shipped significant products within a day of each other, with no apparent coordination. This is what genuine competitive pressure in the AI industry looks like right now, and it tends to benefit users through faster improvement and more competitive pricing.

Who it's for: developers and businesses using AI coding tools. Anyone managing AI spend across multiple platforms and comparing pricing and capability.

━━━━━━━━━━━━━━━━━━━━━━

3. Anthropic Is Eyeing a $10 Billion Compute Deal With Meta

What happened: according to reporting, Meta is in talks to lease data centre compute to Anthropic in a deal potentially worth up to 10 billion dollars over multiple years. Anthropic pitched the arrangement in June, and Meta is reviewing it as a way to monetise excess compute capacity. Either side could still walk away. Anthropic has also signed a separate 45 billion dollar compute deal and is hiring former Google executives to build its own data centres.

Why it matters: the compute infrastructure underlying AI is now being bought and sold between the largest technology companies at a scale that rivals entire national infrastructure budgets. Meta wants to recoup its heavy AI infrastructure spending. Anthropic wants guaranteed compute capacity to keep up with demand. This deal, if it happens, is a significant marker of how intertwined the major AI players' infrastructure has become.

Who it's for: anyone tracking the financial and infrastructure side of the AI industry. Investors and business leaders thinking about long-term platform stability. Anyone curious about where the enormous costs of running AI models actually go.

━━━━━━━━━━━━━━━━━━━━━━

4. AI Flagged 250,000 Suspicious Cancer Research Papers

What happened: an AI tool flagged more than 250,000 cancer research papers as showing writing patterns linked to paper mills, organisations that mass-produce fake or low-quality research for a fee. Suspicious studies rose from around 1% of published cancer research in the early 2000s to more than 16% by 2022. Three academic journals are now testing the tool as a pre-peer-review filter.

Why it matters: this is one of the clearest examples yet of AI doing something human review alone could not do at scale. It also reveals a substantial and growing integrity problem in published cancer research that most people assumed peer review had already caught.

Who it's for: anyone who relies on published research to inform business, health, or product decisions. Healthcare professionals and researchers. Anyone interested in how AI is being used for institutional oversight and quality control.

━━━━━━━━━━━━━━━━━━━━━━

5. Nadella's Warning: Proprietary AI Means Paying Twice

What happened: Microsoft CEO Satya Nadella warned that businesses using proprietary AI models are effectively paying twice: once with money, in subscription or usage fees, and once with their internal knowledge, processes, and workflows, all of which get fed into the AI system through everyday use. The AI provider, in Nadella's framing, keeps both.

Why it matters: this is a rare moment of a major tech CEO publicly naming a tension that most AI vendors would rather not discuss. It's a useful frame for any business thinking seriously about which AI tools to build critical workflows around, and what they may be giving away in the process.

Who it's for: business owners and decision-makers evaluating AI platform commitments. Anyone building significant workflows around a single proprietary AI tool. IT and data governance leads thinking about AI vendor relationships.

━━━━━━━━━━━━━━━━━━━━━━

Also this week: an AWS billing glitch briefly and incorrectly charged some cloud customers estimates in the billions of dollars for services they never used. SpaceX is reportedly in talks to sell the Pentagon AI computing capacity from its underused Memphis data centre. The US Department of Justice cleared federal employees to install TikTok on government devices following a new ownership structure. A survey found 86% of clinicians now use AI daily, with 83% doing so without employer guidance.

━━━━━━━━━━━━━━━━━━━━━━

This week in one sentence: Apple fought on three fronts at once, Grok 4.5 and GPT-5.6 launched a day apart, Anthropic eyed a $10B Meta compute deal, AI caught 250,000 suspicious cancer papers, and Nadella warned proprietary AI comes at a hidden cost.

Born analogue. Raised digital. 30 years of real business experience explaining what AI actually means for work.

— Kaye Nicholson | GrowthZone AI | growthzoneai.co.uk

Subscribe at growthzoneai.co.uk | Follow @GrowthZoneAI

#AIRealityCheck #GrowthZoneAI #UKBusiness #NorthEast