SOUNDINGS

A daily reading of the South African investment case

The read, 30 Aug 2026

Fixed Investment (GFCF) as % of GDP D2
13.20%
Bearish
SA–US 10-Year Yield Spread X1c
409bps
Bullish

State capacity D3

Can the grid and the freight system carry an investment boom? Direction matters more than level.

Horizon

Can the grid carry an investment boom? D3a

Energy availability factor% · weekly
60.0%70.0%80.0%70.3%Apr 25Jul 25Jan 26Apr 26Jul 26

The public fleet only. Eskom’s live CSV holds one financial year at a time and the older folders 404, so this history accrues a week at a time as we capture it — which is also why the composite below cannot yet be normalised against five years.

Table view — Energy availability factor
DateEAF
24 Aug 202670.25%
17 Aug 202670.74%
10 Aug 202672.42%
3 Aug 202672.09%
31 Jul 202675.50%
27 Jul 202677.47%
20 Jul 202679.03%
13 Jul 202675.72%
6 Jul 202673.85%
30 Jun 202666.80%
29 Jun 202670.17%
22 Jun 202667.07%
15 Jun 202668.61%
8 Jun 202664.61%
1 Jun 202665.84%
31 May 202665.50%
25 May 202666.69%
18 May 202668.10%
11 May 202666.93%
4 May 202661.65%
30 Apr 202659.90%
27 Apr 202661.48%
20 Apr 202656.49%
13 Apr 202658.15%
6 Apr 202662.26%
31 Mar 202666.60%
30 Mar 202663.61%
28 Feb 202668.30%
31 Jan 202671.0%
31 Dec 202568.90%
30 Nov 202568.20%
31 Oct 202565.70%
30 Sept 202570.20%
31 Aug 202566.20%
31 Jul 202563.80%
30 Jun 202560.20%
31 May 202559.0%
30 Apr 202556.10%
31 Mar 202556.90%

Bullish EAF 70.25% for the week to 2026-08-24, up from 63.8% a year ago. The week ending 2026-08-31 prints in the coming days; today's re-fetch found no restatements. The grid is the binding constraint the client weighted at half the composite, and it is the component improving.

Can the freight system carry it? D3b

Transnet rail volumesmt · half-year
767880828120232024202420252025
Container throughputTEU per month
300k350k400k450k500k411kOct 24Apr 25Oct 25Jan 26Jul 26

Rail is a step with dots and never a smooth line: it prints twice a year, and interpolating between those prints would draw six months of data we do not have. Turnaround times: he accepted the annual figure (2026-08-30) — the Ports Regulator publishes one number a year, months late, and that is the honest best. The series is sourced and the next cycle captures the published years; until then the board lists it as awaiting capture.

Table view — Transnet rail volumes
DateRail volumes
30 Sept 202581.4mt
31 Mar 202582.1mt
30 Sept 202478.0mt
31 Mar 202476.1mt
30 Sept 202375.6mt
Table view — Container throughput
DateContainer throughput
31 Jul 2026411k TEU
30 Jun 2026454k TEU
31 May 2026342k TEU
30 Apr 2026305k TEU
31 Mar 2026452k TEU
28 Feb 2026387k TEU
31 Jan 2026354k TEU
31 Dec 2025350k TEU
30 Nov 2025366k TEU
31 Oct 2025329k TEU
30 Sept 2025484k TEU
31 Aug 2025410k TEU
31 Jul 2025436k TEU
30 Jun 2025374k TEU
31 May 2025346k TEU
30 Apr 2025291k TEU
31 Mar 2025397k TEU
28 Feb 2025357k TEU
31 Jan 2025333k TEU
31 Dec 2024321k TEU
30 Nov 2024348k TEU
31 Oct 2024310k TEU
30 Sept 2024439k TEU
31 Aug 2024382k TEU
31 Jul 2024329k TEU

Bullish 81.4mt in the half-year to 2025-09-30, against 78.0mt in the comparable half a year earlier. Semi-annual; the next results window is around September. Direction positive at last print, but the print is 11 months old — read the date with the number.

The composite D3

73.7 provisional
0 its own range over the window, 9–74 100

Reads 30 Sept 2025 — the composite carries the staleness of its worst input, not its own cadence.

Electricity availability 2025-03 – 2026-08
Port throughput 2024-07 – 2026-07
Rail volumes 2019-09 – 2026-07

Each component is normalised to 0–100 against its own full available history, and that window is printed above it. The registry asks for five years; the histories are shorter than that and will be for years, so the index says provisional rather than implying a settled scale. Weights are his call, made 30 August 2026 — "electricity must dominate" — so electricity carries half the index, equal to ports and rail combined, which split the rest.

Table view — The composite
DateValue
31 Jul 202673.7
30 Jun 202656.7
31 May 202638.7
30 Apr 202618.8
31 Mar 202651.2
28 Feb 202650.7
31 Jan 202653.1
31 Dec 202547.4
30 Nov 202547.7
31 Oct 202536.8
30 Sept 202568
31 Aug 202549
31 Jul 202546.4
30 Jun 202529.4
31 May 202522.9
30 Apr 20258.5
31 Mar 202524.2

Neutral 73.7 of 100 at 2026-07-31 (provisional), under the client's new weights — electricity 0.5, ports 0.25, rail 0.25. Components: electricity 100, ports 62.2, rail 32.4. First reading under 'electricity must dominate' (equal thirds read 64.9). Electricity's 100 is min–max pinned at its own 17-month maximum — the short-window caveat DESIGN §5 warns about, so the level flatters. The mix is genuinely split: grid strong and improving, ports declining, rail held flat from its 2025-09-30 print, whose staleness the composite inherits.

Where these numbers come from

SeriesSourceCadenceLatest
Electricity Availability (public + private)Eskom Data Portal — Weekly generation capacity breakdown CSVweekly · quantitative2026-08-24 · current
Transnet Rail Freight VolumesTransnet annual & interim results PDFs — semi-annualsemiannual · quantitative2025-09-30 · lagging
Port VolumesTNPA Port Statistics (transnet.net)monthly · quantitative2026-07-31 · current
Port Turnaround TimesPorts Regulator of South Africa — annual Port Performance PDFs (research cycle)annual · quantitativeawaiting first capture
State Capacity IndexComputed at build time (src/utils/)monthly · derived2025-09-30 · lagging