Every year starts pretty much the same way. First, CES lights up Las Vegas with hardware, demos, announcements, and the reassuring (but slightly panicking) feeling that the future is already shipping. A week later, Davos lights up Switzerland with CEOs, ministers, economists, movie stars, philosophers and an equally reassuring feeling that someone, somewhere, has a plan.
If you are a CIO, January is a strange month. You get hit by an avalanche of plausible narratives before you even finish closing last year’s portfolio and setting this year’s priorities. You get “physical AI” and robots on one side. You get “geopolitics, productivity, trust, resilience” on the other. You get panels, headlines, bold claims, and enough shiny certainty to drown a risk register. So, the question I ask myself every January is simple (the answers a bit more complicated): what is signal, what is noise, and what do I do with it on Monday morning when my teams are dealing with real constraints, real dependencies, and very real expectations?
I have learned the hard way to treat CES and Davos as two different instruments. CES shows what becomes technically possible. Davos shows what becomes politically and economically inevitable. Neither tells you what will work in your organisation. That part is still on you.
Why these events matter, even when you (often) roll your eyes
Let’s be honest: both events are easy to caricature. CES can feel like a showroom for prototypes that may never scale. Davos can feel like a high-altitude star-heavy conversation disconnected from the messy world of execution. And yet, both events matter, because they shape the framing that lands on board agendas for the next 12 to 18 months. They influence what your CEO and board/Comex thinks is urgent, what your CFO thinks is investable, what your CHRO thinks is coming for skills, and what your CISO thinks will break next.
Davos 2026 ran under the theme “A Spirit of Dialogue.” CES 2026 leaned heavily into robotics and what the industry itself called “physical AI.”
Those labels are clearly marketing, but the underlying direction is very real: AI is moving out of chat windows and into workflows, devices, operations, and decision loops (and into risk management). That shift changes the CIO agenda, because it changes where value is created, where risk sits, and where governance is unavoidable.
Now, what did I take from this year’s January cycle, both good and bad (and a plethora of things in between) ?
Lesson 1: The AI reality check is now mainstream, and that is a good thing
If you only remember one data point from Davos week, remember this one: PwC’s 29th CEO Survey reports that 56% of companies say they have seen no significant financial benefit from AI so far. That number is very uncomfortable, but healthy. It marks the end of polite self deception. It forces better questions. Where exactly is value supposed to show up? Cost, revenue, risk reduction, cycle time, quality, customer effort? Who owns that outcome? What changed in the operating model so that AI is not just another tool layered on top of broken processes?
If 2023 and 2024 were the years of pilots, 2025 was the year of scaling attempts, then 2026 is shaping up as the year of accountability. Not a grand “AI strategy” slide. Accountability in the portfolio, in the metrics, and in the day-to-day decisions on what gets industrialised and what gets stopped.
The good news is that the conversation is maturing. The bad news is that maturity requires saying no more often, and most organisations (and boards) are not trained for that.
Lesson 2: CES is no longer about gadgets, it is about constraints becoming visible
CES headlines can look like toys, until you translate them into constraints. This year, CES pushed robotics and edge AI hard, including the “physical AI” narrative. You also saw a lot of attention on chips and on device side capability. For a CIO, the practical takeaway is never “we need robots.” The practical takeaway is that AI is increasingly constrained by compute, energy, data locality, and integration into real environments. The promise moves from “generate content” to “act in the world,” which immediately raises questions about safety, testing, liability, and governance.
The good part is that it makes AI less abstract. It forces you to think in terms of processes, sensors, quality gates, and operational control.
The bad part is that it exposes how unprepared many organisations are when software becomes more tightly coupled with physical reality. Your change management cycle time, your security posture, your incident response maturity, your data quality, and your architecture discipline suddenly matter more than your ability to procure licences.
Lesson 3: Workforce impact is no longer a side discussion
Davos this week carried a sharp warning from the IMF leadership on AI’s potential labour market disruption, especially for younger workers and entry level roles. Whether you agree with every number or not, the direction is obvious: AI reshapes work faster than job architectures are redesigned. That creates a gap. The gap is where frustration, inequality, and organisational drag happen.
As a CIO, this is not a PR topic. It is an operating model topic. The question becomes: which roles change, which tasks disappear, which skills become non negotiable, and what training is actually embedded in the workflow rather than delivered as a one-off webinar.
The good news is that companies are finally discussing this openly at the top.
The bad news is that most companies still treat reskilling as a communication plan, not as a redesigned workforce architecture with incentives, time allocation, and measurable capability progression. OKPI’s, anyone?
Lesson 4: Cyber is becoming an AI amplified problem, and an AI enabled response
WEF’s Global Cybersecurity Outlook 2026 highlights a risk landscape shaped by accelerating AI adoption and geopolitical fragmentation. This aligns with what most CIOs feel already: attacks are faster, more automated, and (even) less predictable. At the same time, defenders have an opportunity to become faster too, if they modernise detection, response, and identity foundations.
The good lesson is that resilience is now the language that boards understand. “Can we continue operating when something breaks” is a better executive conversation than “are we compliant.”
The bad lesson is that resilience still requires investment in unglamorous fundamentals. Identity, segmentation, backup integrity, recovery exercises, third party posture, and incident playbooks. These are rarely the projects people cheer for, until the day they need them.
Lesson 5: Davos is a geopolitical and supply chain reminder, not a tech conference
A lot of Davos 2026 coverage focused on political unpredictability, shifting alliances, and calls for de-risking and diversification. For a CIO, this matters in a very practical way. Your technology stack is a supply chain. Cloud concentration, chip dependencies, software vendors, outsourcing footprints, data sovereignty constraints, regulatory divergence. These are not abstract risks. They show up in contract terms, latency, cost spikes, compliance exposure, and delivery disruption.
The good part is that executives are finally seeing technology as a strategic dependency, not just an efficiency lever.
The bad part is that “technology sovereignty” discussions can quickly become hand waving unless you translate them into concrete actions: supplier risk mapping, exit plans that are realistic (if any), architectural modularity, and prioritised investments in what truly reduces lock in.
My January rule: one year, three bets, and relentless basics
CES and Davos are important in one specific way: they compress the narrative for the year. CES tells you what vendors want you to believe is ready. Davos tells you what leaders want you to believe is inevitable. Both create pressure. That pressure can be useful if you channel it into clarity, and destructive if you let it turn into a frenzy of initiatives.
What I try to do every January is to convert “announcements” into three internal questions.
First: which of these trends changes our business model or our operating model in the next 12 to 24 months, not in five years.
Second: what do we need to stop doing so we have capacity to do the few things that matter.
Third: what foundation work is missing, the boring work, the work that makes everything else possible at scale.
That last question is where most of the value sits.
If I take a step back, what I learned again this year is that execution wins, not narrative. The organisations that will get real benefit from AI are not the ones with the loudest announcements. They are the ones that pick a small number of measurable outcomes, redesign the process around them, fix the data contracts, put governance in place, and industrialise responsibly.
They will also be the ones that treat cyber resilience and third party risk as part of business continuity, not as a technical afterthought. CES and Davos are useful because they remind you where the world is moving. They are dangerous when they trick you into believing movement equals progress. The CIO job, especially in a global organisation, is to translate motion into traction.
January is loud. The work starts in February.
