February 2026: So this year is off to quite the start

I’m not sure if you’ve been reading the news, but it’s been a wild few weeks.

We, like the entire internet, fell down the Clawdbot rabbit hole before deciding that just maybe, the safeguards we would need to protect the value of our data (and privacy) might mean it isn’t worth it. The Winter Olympics are on (did you see that backflip on ice skates?), and the Beckham drama caused a lot of debate (#TeamVictoria). Plus, we’ve finally converted Geordie (our AI superstar) to Claude. Goodbye Manus.ai and ChatGPT. Proving that the best way to use AI is to pick a problem to solve, not a model. Because while models change, the problems don’t!

Oh, and we’ve opened an additional client spot for the next 12 months... So if you have some data problems you want solved or just a DPO that actually understands your business, now would be the time!

In other news, Emma has been in Singapore and Melbourne eating all the delicious food, drinking all the good coffee, and talking about all things data governance. She might not know the timezone but if you’d like to discuss data, she’s down (and will be in Sydney next week for our NSW friends).

Lauren’s been holding down the fort in the UK with Geordie who made the trip all the way from Tweed Heads to keep her entertained. Luckily, he has a new found love of British weather. Perhaps jet lag has contributed to this delusional thinking?

It’s a packed newsletter this month. From London’s startup paradox, cyber chaos and China’s genius factories. Read on!

No Ads, No Problems

What’s really been captivating team Friday for the past week is Anthropic’s (Claude) SuperBowl ad that’s thrown down the gauntlet to OpenAI (ChatGPT). If you haven’t seen the ads, here’s one for your viewing pleasure. Not only funny, the ads sent a clear message. Anthropic won’t put advertisements in its Claude chatbot, a declaration that came shortly after OpenAI said ads were coming to ChatGPT. OpenAI CEO Sam Altman, quickly hit back and accused Anthropic’s ads of being misleading and criticised their competitor’s business model. Unfortunately our key takeaway is that he has no sense of humour.

Despite this seeming so inconsequential and potentially even gossipy, it’s actually important. Anthropic is setting up a global debate over the future of AI and how it’s monetised. Considering OpenAI needs to raise about US $200b by 2030 to stay afloat, it couldn’t be more timely. For Anthropic, the ads are a home run shaping the wider corporate narrative around Claude as principled and enterprise-safe (it’s giving Apple v Google, and we love it). We could go on about this forever, so maybe just read this FT deep dive into Anthropic’s extraordinary rise, which charts how the five-year-old startup went from $1 billion in annualised revenue at the start of 2025 to over $9 billion by year-end (with guidance suggesting it’ll exceed $30 billion by end of 2026).

London Calling (for Founders, Politicians, and Anyone Who’ll Listen)

Talking about London, The Economist devoted significant attention to the city in January and the picture is a city that’s simultaneously thriving and stalling. The good news (for Friday...): London has produced more unicorns than Berlin, Paris and Tokyo combined and raised $17.7bn in startup funding in 2025. The magic mix is talent, culture (over half of Britain’s fastest-growing startups were founded by immigrants), and capital. The bad news: productivity is 2% below pre-2008 levels, the stock exchange fell out of the top 20 for IPOs (beaten by Mexico and Oman), housing eats 30% of disposable income, and 130,000 Londoners left in 2023.

We love London. Founders, regulators, clients and world class education institutions are all within a bike (or tube) ride. For anyone building a data business, that proximity to both the innovation side and the compliance side is hard to replicate. But you can’t run a world-class innovation hub on crumbling infrastructure forever. If politicians don’t fix the visas (Emma is taking the changes very personally), housing, ever increasing tax rates (Lauren is taking this surprisingly hard given her Marxist bent) and the creaking transport links, the next generation of founders may just take the Singapore, Dubai or San Francisco call instead.

The Genius Factory

The FT Magazine goes deep on China’s “genius classes”, elite talent streams at top high schools that identify and develop the country’s most promising young STEM minds. Roughly 100,000 Chinese teenagers are selected annually to enter these programmes, and the graduates are now powering the nation’s AI challenge. China graduates approximately five million STEM majors annually compared to about 500,000 in the US. This is the human capital story that doesn’t get enough airtime. While the West debates chip export controls, China’s been building a talent pipeline at a staggering scale. The strategic implication for anyone in data or AI; your competitive advantage isn’t just your technology or your data, it’s your people.

Au Revoir, Zoom

The FT reports that France is pushing millions of state workers off Zoom and Microsoft Teams and onto Visio, a domestically developed videoconferencing alternative. It’s part of a broader European anxiety about tech dependence on US platforms, heightened by Trump-era geopolitical threats and the realisation that America has weaponised tech leverage through export controls on Chinese AI.

Tech sovereignty sounds great in a speech. Making it work is a data problem. Migrating millions of users off embedded platforms means migrating their workflows, integrations, stored data, and habits. The French state’s track record on these projects is… mixed. But the underlying impulse, reducing dependency on a single foreign vendor for critical infrastructure, is sound risk management. Every organisation should be asking: if our key platform provider changed terms tomorrow, could we move?

Books, Bytes, and Breaches

The Times revisits the British Library cyberattack and the ongoing case for more funding. The Rhysida ransomware group hit the library in October 2023, demanding 20 bitcoin. When it wasn’t paid, 600GB of stolen data was dumped publicly. Recovery has eaten 40–50% of the library’s reserves (£6–7 million), and many services were offline for months. But the British Library is more than an arts and humanities institution, it’s part of the national data infrastructure.

That framing is the interesting bit. We tend to think of “data infrastructure” as cloud platforms and fibre optic cables, but the British Library holds 170 million items, runs one of the world’s largest digital archives, and underpins research across every sector from pharma to AI training data. When it went down, researchers, startups, and businesses lost access to datasets they depended on. The lesson isn’t just “fund cyber”, it’s that we need to rethink what counts as critical data infrastructure in the first place. If it holds data that powers research and economic output, it should be governed and protected like the national asset it is.

Hacked, Then Sued

The Times reports that M&S and the Co-op are now being targeted by claims firms following their 2025 cyberattacks. Both retailers were hit by the Scattered Spider group. M&S confirmed customer data was stolen and was reportedly losing £43 million per week in lost sales. Co-op saw 20 million members’ personal data compromised.

Here’s the part that should keep every board up at night: the hack is bad, but the claims firms circling afterwards turn a crisis into even more of a financial avalanche. The real cost of a data breach isn’t just the ransom, the recovery, or the lost sales, it’s the long legal tail. If your governance can’t prove exactly what was compromised and what consent you held, you’re fighting blind.

The Boring Problems Win

The Economist’s briefing identifies three categories of AI adoption pitfalls: behavioural (employees resisting AI for fear of losing credit or jobs), technical (agentic AI reading all website text, hidden code tricking agents into fraud, cloned websites extracting credentials), and organisational (struggling to find specialists with both AI and industry knowledge, lacking the right data). Success stories include Garfield, a British legal tech startup using AI for small-claims disputes, and Johnson & Johnson, which shifted from a thousand-flowers approach to focused AI projects with data councils and rigorous evaluation. Pizza Hut’s AI-enabled ordering and kitchen optimisation is described as “incremental, not transformative”.

This tracks exactly with what we see in client work. The companies that succeed with AI aren’t the ones with the fanciest model, they’re the ones that have solved the boring problems first: data access, evaluation standards, change management, and clear use-case definitions. Start narrow, measure properly, and scale what works.

Big Brother Gets an Upgrade

The NYT reports on ICE’s rapidly expanding AI surveillance machine. The agency has grown its AI portfolio by 37%, deploying Palantir’s ELITE targeting system with facial recognition accessing 200 million photos. Palantir secured a $30 million contract to build “ImmigrationOS,” combining passport data, Social Security records, IRS information, and licence plate reader data. ICE field agents now have access to facial recognition covering roughly 60% of the US population’s photos, with 24 new AI applications added including social media monitoring tools.

Whatever your politics on immigration, the data governance story here is enormous. When a single government agency can combine passport data, tax records, social media, and 200 million facial images into one targeting platform, you’re looking at a data integration project that would make most enterprise CDOs weep with envy (or terror). The question isn’t whether this technology works, it’s whether the oversight, audit trails, and accountability frameworks can keep pace. Spoiler, they aren’t.

Game, Set, Innovation

Fast Company writes that Tennis Australia has quietly built one of the more impressive innovation machines in sport. The Australian Open isn’t just a Grand Slam anymore, it’s running an in-house R&D lab, a startup accelerator that’s piloted 40 companies, and a $40 million VC fund.

We love Tennis, but we also love the data play underneath. When you’re piloting 40 startups, running an accelerator, and managing a VC fund alongside a massive live event, you’re generating rich, multi-layered data across fan behaviour, commercial performance, and technology adoption. Everything from personalising the experience for 1.2 million attendees to managing the opt-ins and data-sharing agreements that underpin a $565.8 million economic impact. So Tennis Australia, if you need a hand, give us a shout!

Cheap Seats, Smart Data

The Economist profiles Ryanair’s extraordinary success: Europe’s largest airline, with a fleet exceeding 640 planes, net margins around 15% (double competitor Wizz Air and nearly four times the global average of 4%), and expectations to carry 208 million passengers this financial year. O’Leary’s secret is ruthless cost discipline: a single aircraft type (Boeing 737) to minimise maintenance and training costs, maximum plane utilisation, and exploitation of cheaper regional airports.

But behind the cheeky PR and “self-loading cargo” rhetoric, Ryanair is a data operation. Dynamic pricing, route optimisation, load factor management, ancillary revenue targeting — it’s all driven by data. O’Leary may not dress it up in AI buzzwords, but the operational discipline that generates 15% margins in an industry averaging 4% is, at its core, a data and analytics advantage. There’s a lesson here for any business, sometimes the most powerful data strategy doesn’t look like a data strategy at all. It looks like operational excellence.

Watercooler Chat

A section of the things we like that keep us sane while running a small business…

• Kowtow — A Kiwi clothing brand that Emma’s only mildly fixated on. Australasian fashion really is excellent.

• Adrian Ritchie — Lauren’s new instagram obsession and new AI “IT” girl.

• Muster Dogs on the ABC — They’re very cute and we’d like one please.