A Tech Trust's NAV Doubled in a Year — Yet Its Shares Still Trade at a ~10% Discount

Every trading day, a London-listed company called Polar Capital Technology Trust files a release with possibly the dullest headline in finance: "Net Asset Value(s)." Retail investors scroll past it for a reason. But in a closed-end fund — an investment vehicle that locks a fixed number of shares around a portfolio of stocks — that daily figure is the whole game. It is the one number that separates what the trust actually owns from what you have to pay to own a slice of it, and right now it is telling a specific story about who is capturing the AI infrastructure cycle.

The reason to care this month: in the year to 30 April 2026, the trust's net asset value rose 102.2%, against a 55.0% gain for its benchmark, the Dow Jones Global Technology Index. Total net assets roughly doubled, from £3.8bn to £7.3bn. That is not luck and it is not a price target — it is the delivered, audited result of a very specific capital-allocation judgment, and it is worth unpacking because it tells you what the manager believes about where the AI cycle's value lives.

A bet, measured in net asset value

Ben Rogoff, who has run the trust for twenty years, runs an "AI maximalist" strategy. The interesting part is not the enthusiasm; it is where the money was put and, just as important, where it was withheld.

Over the year, the trust trimmed the "Magnificent Seven" to roughly 29% of the portfolio — versus more than 50% in the benchmark — and carried its largest US underweight in years. The money went instead into the physical buildout of artificial intelligence: memory makers like SK Hynix and SanDisk, networking and photonics names like Lumentum and Celestica, and the power-and-cooling layer feeding data centers. In other words, Rogoff bet that the early, measurable value of the AI cycle accrues to the hardware and infrastructure that gets built to train and run models — not to the software incumbents whose economics AI undercuts.

That last call is the contrarian core. The trust deliberately underweighted Microsoft, Apple, and the classic software franchises, on the argument that AI collapses the cost of producing code and erodes the seat-per-seat software model. Whether or not that thesis holds over years, it is already visible in the NAV: weight the portfolio toward memory and power four quarters after hyperscaler capital-expenditure forecasts were climbing from roughly $314bn at the start of 2025 to about $751bn by April 2026, and the accounting follows.

NAV is not your price

Here is where the second layer of the decision lives, and it is the part most people miss. The NAV tells you what the portfolio is worth; it does not tell you what you pay.

A closed-end fund's shares trade on the market, so their price can sit above or below the value of the underlying holdings. That gap — the premium or discount to NAV — is its own source of return or risk, independent of whether the stocks inside go up. Right now PCT trades at close to a 10% discount: as of 8 September 2026 the shares changed hands around 645p with the market quoting a discount of –9.77% to net asset value.

That discount is a double-edged number. On the way in, it means you are effectively buying a portfolio of AI-infrastructure stocks for about 90p on the pound. Closed-end investors have made real money simply by buying a discount and watching it narrow — PCT bought back 55.8 million of its own shares, about 4.8% of shares outstanding, at an average discount of 10.1% during the year, which mechanically supports the price. The discount did narrow from 11.3% at the start of the fiscal year to 8.3% at the April year-end, before drifting back out toward 10%.

But the reverse is equally true, and it is the risk a beginner has to see. A discount can persist for years or blow out if the manager's thesis wobbles. The trust's own disclosures flag the levers that could widen it: disappointing returns on AI investment, slower model progress, energy and semiconductor supply constraints, or a weaker dollar against a dollar-denominated portfolio. The discount is not free money; it is the market pricing the possibility that the bet misfires, and it only pays off if the NAV holds up while the discount closes.

What to carry out of it

For anyone trying to figure out whether a vehicle like this belongs on a watchlist, the clean separation is the useful part. Buying an index fund, you make one decision: which stocks. Buying a closed-end trust, you make two: whether you trust the manager's allocation, and whether the premium or discount is a fair price for that judgment.

The NAV doubling makes the first question easy right now — the allocation to memory, networking, and power was validated by delivered results, not by consensus. But an intact thesis does not by itself justify the trade at any price. The second question is the live one: at close to a 10% discount you are being paid to wait for the thesis to keep working, and the same discount that looks like an opportunity is the market's own hedged view that the AI trade has runs left but also room to disappoint. Own the portfolio at ninety pence if you believe the cycle, but never confuse what the trust owns with the price to own it.