The Calver Chronicle

Published 05:00 BST · 99 stories

Markets

Market Daily Brief: Oil at $117, and AI comes for the consultants

How we score a company →

The weekend FT was full of big numbers: a $117 oil price, a $3.5tn merger idea and $10tn of AI spending. Big numbers make good headlines. They rarely make good investments on their own.

Oil and the AI bill

Brent touched $117.45 and settled at $116.72, up more than 10% this month, after Iranian strikes on tankers in the Strait of Hormuz. Some analysts talk of $150. Futures for April 2027 are still above $100. Separately, US executives quit the $25bn Alaska LNG venture over claims of hidden control by Russia's Novatek, which is sanctioned and not investable.

Robert Armstrong asked who pays for the AI build-out. Brookings puts the cost at $10.3tn through 2032, $466bn of AI debt has been issued this year, and the big cloud companies' free cash flow is expected to turn negative. When the spenders need to borrow, someone else is taking the risk.

The consultants' turn

Anthropic and Hg plan to launch a European AI consultancy in London and Berlin by the end of the year. That is a warning for Accenture and the Big Four, whose business is selling people's time.

Accenture: 13/20. It earns about 27% on equity, turns more than all of its profit into cash, and returned $11.5bn to shareholders last year. At $249 it trades at about 18 times adjusted earnings, with a cash yield of about 6% after stock pay. The question is whether AI shrinks the hours clients pay for. The shares have nearly doubled since July, so much of the comfort we would have had is gone.

HgCapital Trust: 11/20. This London-listed trust invests in Hg's funds: about 60 private software companies, which grew profits by 19% in the past year. The shares, at 398p, sit about 24% below the last reported asset value of 524.8p, and Hg itself is buying more. But the companies are valued at about 23 times cash operating profit (EBITDA), they carry debt of about seven times that profit, and fees are about 1.5% a year before the manager's share of gains. The discount is real. So is the leverage underneath it.

Johnson & Johnson and the spin-off

Merger talks between J&J's orthopaedics arm and Smith & Nephew collapsed over who would run the combined business. The FT says J&J is now leaning towards a spin-off of DePuy Synthes. J&J had already announced a separation, due by mid-2027.

Johnson & Johnson: 12/20. A great, diversified health business with strong finances, net debt of only about $27bn and a dividend raised year after year. But at $264 it trades at about 24 times this year's guided adjusted earnings, with a cash yield near 3%, and the talc litigation reserve stands at about $9bn. We admire it. We would buy it lower.

Smith & Nephew: 11/20, unchanged. Our score never assumed a deal.

Mention only

- Tesla: 6/20, too hard, unchanged. Elon Musk is backing a possible Tesla-SpaceX merger valued at about $3.5tn; one critic calls about $2tn of that puffery.
- Vonovia: 9/20, too hard, unchanged. Berlin's plans could cost it about €8.4bn, about 60% of its market value.
- Netflix: 13/20, unchanged. About 1,000 job cuts and content chiefs out after a fall of about $300bn from its peak.
- Stuart Kirk argues the next big theme is wellness, a $7tn market, but it is mostly private firms.
- UK founders are selling rather than scaling at an eight-year high; the Budget is on 28 October.

Also on our desk (since Friday)

- American Express 15/20 (Friday X sweep): a wonderful franchise at about 17.6 times adjusted earnings. Around $265 it would score 16.
- GE Aerospace 15/20 (weekly sweep): a superb engine aftermarket, but about 35 times this year's earnings.
- Fevertree 12/20: net cash and a director buying, but about 38 to 44 times earnings.
- Wise 12, Comcast 11, PHX Energy 11, Dynacor 11, Samsung Electronics 10, Fiserv 10, Block 8. Affirm 6, too hard.

Nothing cleared 15 today. When oil jumps and AI headlines swing prices, the patient investor's job is the same: know what a business earns, and wait for a price that leaves room for error.

Research only, not investment advice.

More from the Weekend edition, Sat 24 Oct 2026 →