Viska · strategy read · report date 2026-08-05

Six positions the research took today

Bullish on

$8.78M

68.0% of invested capital · gold & silver, AI power, AI semis, energy

Bearish on

$4.13M

32.0% of invested capital · China base metals, industrial services

BULLISHGold & silver — positioning squeeze

Every gold and silver document published today points one way, and the argument is that the buyers have not arrived yet.

NEM$732,975+0.5% B$745,095+3.8% SBSW$521,249−5.2% FNV$437,720−0.1% HL$384,750−2.2% WPM$346,230+3.7% AUGO$288,950−8.8%

$3,456,969 · 26.8% of invested capital · 8 documents, every one bullish

The Market Ear — gold has broken above the post-highs downtrend, and China central-bank buying keeps the structural bid intact.

The Market Ear — CTAs are still short gold and speculative positioning remains depressed, leaving room to rebuild; a held breakout is squeeze fuel.

The Market Ear — silver shows the same setup: light positioning, CTAs short, a vol reset and short-covering convexity above the breakout level.

So: the case for the fund's largest cluster is not a demand forecast, it is an empty-boat argument — the systematic and speculative money that would have to buy is currently short or absent. Five of these seven positions sit at or below cost. The fund owns the exposure the thesis says is about to be chased, and owns it flat.

BEARISHChina industrial demand — base metals

China's services economy printed its worst month in nearly two years, and no house published a single bullish word on the fund's copper and aluminium.

SCCO$780,640+15.8% CENX$753,920−1.4% FCX$673,000+7.0% LUN$576,810+13.5% AA$655,620−14.0% RIO$365,650+4.1%

$3,805,640 · 29.5% of invested capital · 0 bullish documents

Goldman Sachs — China Services PMI fell to 50.4 in July from 54.1, its lowest since September 2024, against Goldman's own 53.9 forecast and 53.7 consensus.

Goldman Sachs — July CPI forecast trimmed to 0.9% and PPI to 4.0%, a fourth consecutive year of factory-gate deflation.

Goldman Sachs — the one metals margin story of the day, Gerdau, runs on US tariffs and price hikes in North America, not on Chinese demand.

So: nearly a third of the fund's capital is priced off Chinese industrial activity, and the only Chinese datapoint published today missed by three full points. SCCO, FCX and LUN are carrying gains of 7% to 16% into that print — the gains are the exposure. This is the sleeve where the fund is long and the evidence is absent.

BULLISHAI data-centre power buildout

The buildout of power for AI is the single loudest theme in the day's research, and the fund owns the equipment that gets bought.

CCJ$930,900−0.3% POWL$634,140−3.8% NVT$486,720+17.6% VRT$485,874+4.1% GEV$305,559+18.4% ETN$275,757+32.0% AZZ$152,710+51.5%

$3,271,660 · 25.3% of invested capital · ~24 documents across four sub-themes

Goldman Sachs, the day's highest-conviction note — Siemens Energy beat on EBITA, orders and free cash flow, FY2026 margin guidance raised to the top of range; Buy, target €212, 40.5% upside, more than 60% earnings growth to FY2030 — the thesis resting explicitly on closing the "GE Vernova valuation gap."

JP Morgan, European Capital Goods — Q2 was a broad beat-and-raise across electricals, data-centre and grid, with the market punishing heavily-owned names on "good results, poor reaction" optics, Vertiv named among them.

Goldman Sachs — gas-turbine capacity expansion and grid-technologies backlog growth both scored strong-bull at long and future horizons, the only such reads in the corpus.

Goldman Sachs — Rockwell Automation beat on semiconductor, data-centre, e-commerce and warehouse automation demand; target raised to $510.

So: the fund holds the picks-and-shovels of the theme the sell-side is most committed to, and GEV is the benchmark multiple in the day's strongest argument rather than its subject. VRT's softness is named as crowding rather than earnings. CCJ, the fund's uranium position, is its most direct claim on AI baseload and sits flat on cost.

BULLISHAI semiconductors — into the drawdown

Two houses independently call the current semiconductor drawdown a buying window, and the fund's two positions are sitting flat on cost.

MU$446,335−0.3% AMD$311,148−1.4% GOOGL$377,650+7.1% GLW$319,780−0.6%

$1,454,913 · 11.3% of invested capital · 19 documents, AMD the only fund holding named by any document

Goldman Sachs on AMD — Buy reiterated, 12-month target $640, CY26-28 EPS estimates raised 13%, 2027 datacentre growth well over 100% on the MI-4XX ramp and server-CPU TAM compounding ~50% through 2030. The house expects the shares lower first, against very high expectations — that is the entry, not the thesis.

JP Morgan (Gokul Hariharan), house view — the third 20%+ Asian Tech and SOX drawdown of this AI upcycle is not the start of a cycle peak; scaling laws, token demand, capex and cluster-efficiency data all still point up. He turns buyer.

Citi — carries a Korea memory HBM shortage structural note, bearing directly on MU.

So: the fund is being told by the publishing house that its position will get cheaper and is worth more. AMD carries a $640 target against a position marked 1.4% below cost, with the bear case at $264 and the bull at $875. Estimates are going up while the price is going down, and that gap is the whole opportunity in this sleeve.

BULLISHMiddle East risk & refining margins

Middle East escalation is showing up in unrelated companies' guidance, and refining margins are being marked up across two houses.

EQT$448,545−6.5% RIG$144,855+0.7%

$593,400 · 4.6% of invested capital · 15 documents on refining and disruption

JP Morgan — Asian refining margins bullish across five reads; Goldman Sachs marks refining-margin strength bullish at every horizon it scored.

Goldman Sachs on BP — Buy, target raised to 660p, on a 2Q26 beat driven by trading and refining strength, accelerating deleveraging and a re-accelerating upstream pipeline.

Goldman Sachs — Philippine input and output prices are reaccelerating partly on renewed Middle East conflict escalation; Booking Holdings cut Q3 and FY26 estimates citing Middle-East conflict headwinds. Mizuho is sceptical the de-escalation holds.

Goldman Sachs on EOG — 2Q26 free-cash-flow, cash-flow and production beats with capex below plan and guidance maintained-to-raised: US upstream is converting at these prices.

So: conflict risk is no longer confined to energy notes — it is being written into the guidance of travel and consumer companies, which is what a durable premium looks like early. The fund's smallest sleeve is the one the tape is validating, and EQT sits 6.5% below cost while the upstream cash-flow read is positive.

BEARISHIndustrial services & contracting

Industrial margins are compressing in every document that touched them, and the fund's three worst positions are exactly there.

MYRG$255,248−19.0% AMRC$33,528−32.5% VMI$25,175+27.7% PRIM$9,994−33.7%

$323,945 · 2.5% of invested capital · no house wrote about any of these names

Goldman Sachs — the electrification bear leg is margin, not demand: acquisition integration costs create a 2H headwind implying roughly flat EBITDA margin year on year.

Goldman Sachs on UL Solutions — target cut to $86 from $100, Industrial organic growth decelerating from 8.2% to 7.2% on advisory weakness, margin contracting.

Goldman Sachs on BrightView — Sell; revenue, EBITDA and EPS all missed, and management cut full-year guidance across revenue, EBITDA and free cash flow. Services execution is where the misses are landing.

So: the fund's deepest drawdowns — PRIM at −33.7%, AMRC at −32.5%, MYRG at −19.0% — sit in the one part of the electrical complex where the research is negative, while the equipment makers beside them beat and raised. The distinction the market is drawing is manufacturers versus installers, and the fund is on both sides of it. These positions are small enough that the drawdown has cost little; the reason to act is that nothing published today argues for owning them.