News Corp is no longer a newspaper company. It is becoming an AI one

News Corporation reported its fourth quarter and full-year fiscal 2026 results on August 5th. Full-year revenues rose 7% to $9.03 billion. Net income from continuing operations increased 15% to $743 million. Total segment EBITDA - earnings before interest, taxes, depreciation and amortisation, a rough measure of underlying cash profit - grew across all three reportable divisions. These are solid but unremarkable numbers for a company that has now posted 12 consecutive quarters of profitability growth. The headline from a quarterly earnings report is never the point. The point is what the numbers reveal about the shape of the business.

The shape is changing. News Corp's three core segments - Dow Jones, Digital Real Estate Services, and Book Publishing - are being recombined by a force that Rupert Murdoch's successors identified earlier than most of their rivals: the demand for verified content in an artificial-intelligence economy.

Dow Jones is the crown jewel. Its revenues grew 8% each in the second and third quarters of fiscal 2026, to $648 million and $619 million respectively. But the more interesting growth is internal. Risk and Compliance, the business that sells regulatory data to financial institutions, grew 20% in the second quarter and 19% in the third. Digital advertising revenues at Dow Jones set records. Segment EBITDA margins reached nearly 30%, a figure that would not look out of place at a specialist B2B data company, not a publisher of the Wall Street Journal. Management has drawn a line to $1 billion in annual Dow Jones segment EBITDA within five years, up from $588 million in fiscal 2025. If achieved, that trajectory would more than double the segment's earnings power.

The Wall Street Journal's subscription model provides the underlying moat. Digital-only subscriptions reached 4.29 million in the second quarter and grew 11% year on year by the third, now accounting for 92% of all Journal subscriptions. Total subscriptions, including print, averaged 4.68 million. That is a recurring-revenue base built over decades, not a growth hack.

The real estate segment tells a different, cyclical story. Move, operator of Realtor.com, grew revenues 10% in both the second and third quarters, to $143 million and $148 million. But it was REA Group, the Australian operation, that drove the headline: revenue growth of 20% in the third quarter to $325 million, helped by a positive foreign-exchange swing. Combined Digital Real Estate Services revenue grew 17% in the quarter. The Australian property market's recovery is real, and the benefit for a firm that controls the listing platform is direct. But housing is cyclical, and the U.S. housing recovery at Move remains mired by elevated mortgage rates. The segment's strength should not be mistaken for structural permanence.

Book Publishing, anchored by HarperCollins, grew 6% to $633 million in the second quarter and 8% to $555 million in the third. A quarterly record, but also a reminder that publishing is, at its core, a function of having the right title at the right time. It is a cash-generating business, but not the strategic centre of gravity.

The centre of gravity is shifting toward something the market has yet to fully digest: AI content licensing. News Corp is no longer merely selling subscriptions and advertising. It is selling provenance.

The chief executive, Robert Thomson, has adopted a neat framing. "We're essentially an input company," he said at a Morgan Stanley technology conference in March. Semiconductors are an input. Data centres are an input. Energy is an input. So, increasingly, is verified content. The argument is that as artificial-intelligence models scale, the scarcity is not computation - it is trustworthy training data. News Corp's unique mix of real-time financial data, breaking news, and specialised property listings is hard to replicate with synthetic material.

The deals are starting to flow. A multiyear licensing agreement with Meta, announced in March, is worth up to $50 million a year for at least three years. News Corp already has a partnership with OpenAI, signed in 2024. The company expanded its existing partnership with Bloomberg to include AI rights for Dow Jones content. Mr Thomson said the firm is in advanced negotiations with several additional AI platforms and expects to receive its share of a $1.5 billion settlement with Anthropic later this calendar year.

These revenue streams are not yet material to reported earnings. But they represent high-margin incremental income on content that has already been produced. The marginal cost of licensing a digital archive to a technology company is close to zero. If the pipeline continues, what begins as a footnote becomes a structural shift in the earnings profile.

It is tempting to dismiss the AI-licensing angle as management enthusiasm. After all, the media industry has cycled through similar hopes before - from the early internet to social media to native advertising. Each time, publishers overestimated their leverage and underestimated how quickly platforms would capture the value.

Yet the dynamics are genuinely different this time. The previous cycles were about distribution: newspapers needed platforms to reach readers. The current cycle is about inputs: AI companies need publishers' content to train their models. The leverage has flipped, at least temporarily. Whether it holds depends on whether the content monopoly is real. Dow Jones' financial data and verified news are genuinely scarce. The rest of the stable - tabloids, regional newspapers - is less so. The value accrues to the parts of the business that own unique information, not to the parts that merely report on it.

To be sure, there is a risk that AI companies develop alternative data sources or that regulatory intervention forces licensing on less favourable terms. Mr Thomson's "woo or sue" strategy - embrace deals where possible, pursue litigation where content is taken without permission - is aggressive and not without risk. It works so long as the supply of premium content is genuinely constrained. If the content moat proves shallower than claimed, the licensing revenue will plateau rather than compound.

There is also the broader valuation question. As of late June, News Corp's shares traded around $25, giving the company a market capitalisation of roughly $14 billion. The stock trades at approximately nine times next-twelve-month EV/EBITDA and around 21 times forward earnings. Those multiples are low for a business compounding segment EBITDA at double-digit rates and growing its highest-quality division at 8% a year. Thomson Reuters, whose professional information business is the closest analogue to what Dow Jones is becoming, commands a premium valuation as a B2B data compounder. News Corp is still priced like a newspaper.

The buyback programme suggests management sees the gap. The board authorised a new $1 billion repurchase in July 2025, on top of $303 million remaining from an earlier programme. $459 million was repurchased in the nine months to March 2026. Mr Thomson described the pace as running at over four times the prior rate. Share buybacks are not inherently value-creating - they are value-creating only when shares are genuinely undervalued. Mr Thomson's conviction that the stock trades at a significant discount to intrinsic value is either right, in which case buybacks are rational capital allocation, or wrong, in which case the company is burning cash on an expensive delusion. The earnings trajectory so far lends weight to the former interpretation.

AInvest's aggregate signal labels the stock a Buy, with a fundamental rating of 9.31 out of 10. That score is opaque - the underlying methodology, contributing analysts and rating distribution are not disclosed - but it aligns with the broader analyst picture: eight analysts cover the stock, with five Buy ratings, two Outperform ratings and one Hold, and zero Sell ratings. Consensus has a mean price target of roughly $35, implying about 40% upside from June levels.

The danger is not that anything is broken at News Corp. The danger is that the reclassification is too slow. If the market continues to price Dow Jones as a declining newspaper rather than an ascending data business, and if it assigns zero value to the AI-licensing pipeline, the stock will languish despite improving fundamentals. The catalyst could be a sustained move toward Thomson-Reuters-like multiples for the Dow Jones segment, or a surprise in the scale of AI licensing revenue that forces analysts to rebuild their models.

For investors, the case for News Corp rests on three pillars: a Dow Jones unit that is quietly becoming a B2B data compounder, a real estate platform that benefits from cyclical recoveries, and an AI-licensing pipeline that is genuine but not yet priced. The break condition is clear. If housing turns sour, Dow Jones growth stalls, and AI licensing fails to scale beyond the initial deals, the stock's low multiple may prove fully justified. If any one of those pillars holds, the market's current classification is too narrow.

Content is king only when someone else owns the castle. News Corp is building its own.