Published on LinkedIn and amitabhapte.com on8^th Mar 2026

Three signals this week. Barcelona’s biggest mobile show. Apple’s biggest product week in years. And the most honest labour economics report the AI industry has produced. Different stages, same underlying story: intelligence is arriving everywhere at once, and the gap between capability and consequence is widening.

MWC Barcelona 2026: AI Moves into the Pipe

The GSMA’s theme this year was “The IQ Era.” For once, the branding matched the floor. MWC 2026 wasn’t about device launches. It was about AI embedding into network infrastructure itself.

The most consequential announcements came from operators, not handset makers. The GSMA launched Open Telco AI, a collective industry effort to weave AI into carrier operations. Qualcomm’s new X105 modem embeds an AI processor directly in the chip, a 6G stepping stone that will shape OEM roadmaps for 2027 devices. Deutsche Telekom debuted an AI call assistant that lives in the network, not in an app. And AWS committed €33 billion to Spain, explicitly framing the country as its European AI epicentre.

My PoV: Telecom providers are quietly becoming AI infrastructure providers. When intelligence is embedded at the carrier layer, every device on that network gains capability it didn’t ship with. Your connectivity strategy and your AI strategy are now the same strategy. Most enterprise roadmaps haven’t caught up to that yet.

Apple’s Big Week: Intelligence at $599

Apple launched seven products in three days. Two stand out.

The iPhone 17e brings the A19 chip, 256GB base storage, and full Apple Intelligence to the $599 price point. The story isn’t the device, it’s the distribution. Apple’s AI stack just reached a much larger addressable base.

The MacBook Neo is more significant. A $599 laptop running an A18 Pro chip, the same silicon as the iPhone 16 Pro, with Apple claiming 3x faster on-device AI performance than comparable Intel machines. It is the first Mac powered by an iPhone chip. The architectural wall between Apple’s phone and laptop lines has come down.

My PoV: This week wasn’t about hardware. It was about what happens when AI-capable silicon reaches commodity pricing across every form factor. Combined with agentic coding tools like Claude Code, the barrier to building functional software has effectively hit zero. The question for technology leaders is no longer which devices to provision, it’s how to govern what a workforce of accidental developers builds with them.

The Anthropic Labour Report: The Gap Between Fear and Fact

Anthropic published “Labour Market Impacts of AI: A New Measure and Early Evidence” this week. It is worth reading carefully.

The paper introduces “observed exposure”, measuring what AI is actually being used for at work, not what it theoretically could do. The gap is stark: Computer and Math roles have 94% theoretical AI exposure but only 33% actual usage coverage today. Legal sits at 80% theoretical, 15% actual. The wave is real. The timeline is slower than the headlines suggest.

Computer programmers top the “actually happening now” list at 75% task coverage, followed by customer service at 70% and data entry at 67%. Yet unemployment in exposed occupations has not meaningfully risen since ChatGPT’s 2022 launch. The one signal worth watching: hiring of workers aged 22–25 into exposed roles has quietly slowed.

My PoV: That entry-level hiring signal matters more than aggregate unemployment data. The mid-level talent of 2028 is being shaped right now. Workers who are not hired into exposed roles today don’t disappear, they redirect. But the pipeline compresses. For enterprise leaders, the implication is concrete: talent acquisition strategies in software development, customer operations, and financial analysis need to account for a structurally thinner entry cohort arriving in the next two to three years.

My Takeaway This Weekend

MWC confirmed AI is now infrastructure, in the network, not on top of it. Apple confirmed AI silicon is now a commodity, at $599 in both your pocket and on your desk. Anthropic confirmed the labour disruption is real but the clock is slower than feared, for now.

The word “yet” is doing heavy lifting across all three stories. The period between “not yet” and “already happened” is consistently shorter than organisations plan for. The question is not whether these shifts are coming. It is whether your architecture, your talent pipeline, and your operating model are being built for the right horizon.

Published on LinkedIn and amitabhapte.com on 9^th Nov, 2025

---

This week in AI – When Ambition meets Arithmetic

The AI boom is now running on infrastructure, not imagination.

OpenAI’s CFO Sarah Friar spent much of the week clarifying that the company isn’t asking for a government backstop on its $1.4 trillion infrastructure plan, a figure so vast it rivals national energy budgets. Her comments drew a sharp response from Trump’s new AI czar, David Sacks, who declared there would be “no federal bailout for AI.” The exchange revealed the growing tension between private ambition and public patience. The industrialisation of intelligence is proving as capital-intensive as any past revolution, and just as politically fraught.

Yet the capital keeps flowing. OpenAI signed a $38 billion cloud-computing deal with Amazon, making AWS its primary engine for model training and deployment. For Amazon, it’s a strategic coup; for OpenAI, a hedge against the global shortage of compute and chips. The partnership underscores how the AI stack is consolidating, fewer players, bigger bets, tighter dependencies.

Apple talking with Google to power a new Siri using Gemini AI marks a pragmatic turn for the company once obsessed with control. Even Apple is realising that no single firm can build the full AI stack alone. The new race isn’t to own the model; it’s to own the infrastructure, the energy, and the ecosystem.

Tesla shareholders approved a record-breaking $1 trillion pay package, a figure that defies logic until you see the scale of his ambition: turning Tesla from an automaker into an AI-and-robotics platform that spans cars, humanoids, and autonomous fleets. His reward is tied not to quarterly profits, but to an $8.5 trillion valuation. The math may be extraordinary, but it captures the mood of the moment, an economy running on belief as much as balance sheets.

My takeaway this weekend

The age of AI infrastructure is here and it’s expensive.

Across these stories runs the same current: the strain is showing. Building intelligence at planetary scale demands not just algorithms and GPUs, but grids, land, and trillions in capital. The numbers are breath-taking, but so are the risks. The question is no longer can we build it, but how much can the world afford to spend chasing it. Musk’s trillion-dollar vision shows what’s possible when belief meets capital. Friar’s clarification reminds us what happens when optimism meets arithmetic. The future of AI won’t be limited by imagination. It will be priced by reality.

---

Beyond AI: my mindshare – Human Code Behind Technology Buying

Earlier this week, I joined fellow technology leaders Tom Clark (Everywhen), Rebecca Reynolds Jones (Institute of Directors), and Charlotte Walters (HSBC) on a panel at Computing’s Future of B2B Tech Marketing event, moderated by Computing editor Tom Allen.

Our discussion explored how technology buyers and marketers are navigating an era of automation, AI, and hyper-personalisation and what trust really looks like when so much outreach is now machine-generated.

My core message was simple: credibility begins with relevance.

The best outreach doesn’t just know your name; it understands your business, your pressures, and your purpose. The quickest way to lose a CIO’s trust? Treat technology like a product rather than a partnership.

As AI personalisation scales, human authenticity becomes the real differentiator.
Integrity, empathy, and business fluency still open more doors than algorithms ever will.

“The future of technology marketing won’t belong to those who automate the fastest, but to those who empathise the deepest.”