This is a demonstration built on sample data. Nothing here is a real account, holding or balance.
The scouting desk
Picks
Every name on this screen arrives with its complete written case: the argument for it, the argument against it, and what must be true for the idea to work. When one earns your conviction, the watchlist starts its clock, and the written gates decide what happens after that.
iShares Global Tech Index: -0,3% over the same stretch
Sell
The company
Alphabet owns Google Search, YouTube, Android, the Google Cloud arm and the Gemini AI models. It sits across advertising, cloud and AI, and is the cheapest of the mega-caps on earnings. Liked for the moat and valuation; watched for an antitrust overhang and an earnings date this week that can move it sharply either way.
Where this sits in the cycle
An advertising-plus-cloud giant geared to ad budgets and cloud demand. It reports first of the mega-caps this week, so it sets the tone; the risk is a soft cloud or capex number that confirms the AI-spend worry now hitting the group.
Buy this for real: begin the walk-throughClose this idea
iShares Global Tech Index: -3,9% over the same stretch
Buy
The company
CrowdStrike sells cloud-based security that protects company laptops, servers and cloud accounts from attack, all run from one platform called Falcon. It sits on the software side of the remit, not the AI-hardware side: a subscription business with very sticky customers and strong cash generation. We like it because it is the clear leader in its field, it keeps selling more to the customers it already has, and security is one of the last budgets a company cuts.
Where this sits in the cycle
Security spending holds up better than most through a slowdown, which makes this steadier than a chip name. The catch is the price: fast-growing software is priced richly, so it falls hardest when interest rates rise. There is mild seasonality, with the strongest sales late in the year as companies spend what is left of their budgets.
iShares Global Tech Index: -3,9% over the same stretch
Buy
The company
Palo Alto Networks protects company networks, cloud systems and security teams, increasingly sold as one bundled platform rather than separate tools. Like CrowdStrike it is a software and subscription business, hard for rivals to break into and a strong generator of cash. We like it because customers are standardising on fewer, bigger security vendors, and Palo Alto is one of the winners of that shift.
Where this sits in the cycle
Security budgets are resilient through the cycle, so demand is steadier than for hardware. The risk is the valuation: it is priced for growth, so a jump in interest rates or any slip in growth hits it hard. It is not very seasonal beyond the usual year-end budget flush.
iShares Global Tech Index: -6,3% over the same stretch
Hold
The company
Nvidia makes the chips that train and run almost every serious AI model, plus the software that keeps developers tied to them. It is the engine room of AI compute infrastructure, as central as a single company gets. We like it for its near-monopoly on AI training, its enormous marginsThe share of every sale kept as profit after all costs. Margins of 46% means 46 øre of each krone sold stays in the company., and an ecosystem rivals struggle to copy.
Where this sits in the cycle
It sits at the very centre of the AI spending boom, which is also its risk: it is the most volatile way to own the theme, so it runs hardest and falls hardest. If the cloud giants ever pause their buildout, this feels it first. It is driven by product cycles rather than the calendar, so there is little seasonality.
iShares Global Tech Index: -6,3% over the same stretch
Hold
The company
Broadcom makes the networking chips and the custom AI chips that the big cloud companies design with it, and it also owns a large software business in VMware. It is picks and shovelsSelling tools to the gold rush instead of digging for gold: suppliers who get paid whichever miner wins. for AI compute: it gets paid as the data centres get built, whoever wins. We like it for its huge cash generation, its grip on custom AI silicon, and a growing dividend.
Where this sits in the cycle
Half the story rides the data-centre spending cycle, the other half is steady software revenue, which softens the swings a chip name usually has. Chips are cyclical, but AI demand has turned that into a longer, steadier run for now. The risk is a pause in cloud spending, which would hit the chip side first.
iShares Global Tech Index: -6,3% over the same stretch
Buy
The company
Arista builds the high-speed switches that move data between the chips inside an AI data centre, the plumbing that lets thousands of accelerators work as one. This is core AI compute infrastructure on the networking side. We like it because it leads in the fast switching that AI clusters need, it sells to the largest cloud builders, and it earns fat marginsThe share of every sale kept as profit after all costs. Margins of 46% means 46 øre of each krone sold stays in the company. doing it.
Where this sits in the cycle
It rides the cloud giants' building spree, so it does well while that spending grows and stalls when it pauses, which makes it more cyclical than a software name. A large share of its sales comes from a handful of huge customers. We are still early to middle in the AI buildout, the part of the cycle that suits it.
iShares Global Tech Index: -0,4% over the same stretch
Hold
The company
Vertiv makes the power and cooling kit that keeps data centres running, including the liquid cooling that dense AI racks now need. It is the unglamorous but essential part of AI compute infrastructure: once a data centre is built, someone has to power it and cool it. We like it because heat and power have become real bottlenecks for AI, and Vertiv sits right in that gap.
Where this sits in the cycle
Its order book sits directly downstream of the cloud giants' spending, so it is more industrial and more cyclical than the chip names. It benefits from two long trends at once, the AI buildout and broader electrification. The clear risk is a cut to data-centre budgets, which would show up in orders quickly.
iShares Global Tech Index: -0,4% over the same stretch
Buy
The company
ASML makes the machines that print the circuits onto chips, and it is the only company in the world that makes the most advanced kind, called EUV lithographyThe machines that print microscopic circuits onto chips. ASML is the only company in the world that makes the most advanced kind.. It sits upstream of the entire AI buildout: every leading-edge factory has to queue for its machines. We like it as the monopoly nobody can route around, able to set its own prices, with superb returns.
Where this sits in the cycle
Its sales follow the equipment cycle, the years when chipmakers add capacity, so it is a longer and lumpier cycle than the chip names below it. The main risk is China export rules removing a big slice of demand. It is driven by its order book rather than any season.
iShares Global Tech Index: -0,4% over the same stretch
Hold
The company
TSMC is the factory that actually makes the most advanced chips in the world, Nvidia's and Broadcom's among them. It is the manufacturing heart of AI compute infrastructure, with a near-monopoly on the leading edge. We like it because almost the entire AI chip supply chain has to come through its doors, whoever designs the chip.
Where this sits in the cycle
Chip-making is cyclical, but AI and heavy computing have stretched that into a longer run. The overriding risk is geopolitical: it is in Taiwan, so any serious tension over the strait is the thing that breaks the idea. It is capital-heavy and tied to the whole industry's cycle rather than the season.
Buy this for real: begin the walk-through
The coloured tag is the live read, moving with the price. Before you own a name it follows the market: Buy in a clear uptrend, Sell in a clear downtrend, Hold in between. Once you own it, your own take-profit and kill take over, and it turns to Sell when either is reached. It is a prompt, never an order.
The record
Closed ideas, scored against iShares Global Tech Index over the same stretch and kept on the books.
SMCI2026-05-12 → 2026-06-18
-6,5% · fund +12,9%
Kill criterion fired: the accounting questions did not clear, so the idea was closed and scored rather than waited on.
How this works
Where picks come from. Once a month, or whenever you ask for one, a scout run sweeps the hunting ground: AI infrastructure and compute, software and SaaS, and commodities as diversifiers, always in companies worth more than two billion dollars. It proposes at most three names, and each one arrives written up in full, with its tax picture attached. One exclusion is absolute: no growth built on circular AI financing, where suppliers fund their own customers.
The two tests. Every name is judged through two lenses. The Buffett lens asks whether this is a good business at a defensible price, weighing growth, profitability, the balance sheet and valuation. The trend lens asks whether the market currently agrees, measured plainly by the price against its own recent history. A wonderful business in a broken trend is something to watch rather than buy, because the trade goes with the market, never against it.
What happens to a card. The newest scout run sits on top in full. Names from earlier runs that you have not yet acted on stay below in the compact rows, kept current with today's price beside the price at suggestion, and sink as newer arrivals land above them. Move a name to the watchlist and its price is frozen from that day; from then on it is tracked exactly as if you had bought it, and at six and twelve months it is scored against the same money put into iShares Global Tech Index, the fund the long-term money buys back into and the bet each pick has to beat.
When you would sell. Every pick carries kill criteria, written at suggestion time while heads are calm. They are read weekly: the Monday letter carries each open idea's current price beside the words that would end it, so a broken thesis waits days, not a month. When a criterion fires the idea is closed and scored rather than quietly forgotten; the fuller scoring against the index stays with the monthly review.
The rule above all. A pick is information, never an instruction. Money moves only through the written gates: the thesis written down, the cooldown observed, the size checked and the wrapper chosen.
Nothing can be confirmed before the clock runs out. That is the point of it.
Close this idea
Nothing can be confirmed before the clock runs out. That is the point of it.
Close this idea
Nothing can be confirmed before the clock runs out. That is the point of it.
Close this idea
Nothing can be confirmed before the clock runs out. That is the point of it.
Close this idea
Nothing can be confirmed before the clock runs out. That is the point of it.
Close this idea
Nothing can be confirmed before the clock runs out. That is the point of it.
Close this idea
Nothing can be confirmed before the clock runs out. That is the point of it.
Close this idea
Nothing can be confirmed before the clock runs out. That is the point of it.