Pento Versus Roberts, Garrib Tested
2026-10-05 · Nordic week ahead · Garrib on capex and European deindustrialisation, Pento versus Roberts on 2027
How it is made. Aavistus Daily is written and voiced by our own systems, every morning, from data we gather ourselves. The voices are synthetic. Every fact is traced to a source on this page, and every dated call is settled in public on the calibration page. The presenters, Nora and Callum, are named characters, not people.
Aavistus is a new framework, still under development. We will make mistakes, and every dated call is settled here in public so you can see them; the aim is to keep calibrating until the record beats the average forecaster, not to sound right.
Two bearish podcast calls — Pento saying 2027 is a disaster, Garrib on hyperscaler capex pulling the US long end higher — tested against Lance Roberts arguing the opposite.
Garrib on capex, Europe tested
Aidan Garrib at PGM Global on Forward Guidance argues the US long end is being pulled higher not by inflation expectations — anchored at 2.4% on 10-year breakevens — but by expectations of a higher terminal policy rate driven in turn by hyperscaler AI capex. The chain is a volume shock to real investment that lifts nominal growth without producing a visible inflation panic. His scale point is specific: high-bandwidth memory is projected to consume roughly 30% of hyperscaler data-centre budgets in 2026, up from 8% in 2023–24, with US nominal GDP printing 6.6% against domestic credit growth of 3%.
The direction is defensible: the mechanism is consistent with our standing view that the US 10-year holds 5% to 5.5% through year-end, now around 5.25% and up roughly 50 basis points on the month. On the ECB side, euro area September HICP at 3.8% (up from 3.2%, with France leading the acceleration from 2.6% to 3.4%) reinforces our view that the deposit rate reaches at least 3.25% by mid-2027. The euro trades at a 17-month low this morning and our read that EUR/USD stays below 1.15 on a six-month horizon is on side.
Where Garrib runs ahead of the data is on the structural European call. French HICP reaccelerating supports him on inflation. But French unemployment just moved from 8.3% to 8.2%, and German unemployment is flat at 4.0%. That is cyclical weakness with reacceleration risk, not structural fracture — deindustrialisation at the speed he describes shows up as a layoffs wave through the industrial supply chain, not as prints that drift sideways. The direction is right; the pace is too fast for what the labour market is showing today. The read that would lend weight to his framing is a cluster of Nordic or German industrial profit warnings of comparable size inside the next five weeks.
A quiet Nordic calendar
The policy calendar does not move a Nordic position this week: the ECB meets 29 October, Norges Bank 5 November, the Riksbank 12 November. The one running position the record is tracking is a Nord Pool call made 2 October: Finland's next weekly average stays below €100 per MWh. The week ending Sunday printed in the high forties — wind back on the eastern side of the Scandinavian ridge, against a stubbornly expensive western side with the Danish zones well above €170 and southern Norway and southern Sweden around €140. Delta Air Lines reports this week and is the Nordic-relevant print, because it anchors transport comparables for Maersk ahead of its own release. The case for chasing Maersk at a fresh 52-week high is not there; the read changes only if the third-quarter release shows contract rates re-priced higher into 2027 despite the Suez return.
Pento's disaster, Roberts's buy, our read
Michael Pento on Thoughtful Money with Adam Taggart said 2027 is a disaster for both stocks and bonds across all durations and moved his portfolio to 100% market-neutral. His three-legged mechanism: the Fed under a new chair tightening to prick an AI bubble, a fractured global bond market he frames via a US debt-to-revenue ratio of 720%, and a market-breadth picture in which much of the S&P 500 is already deeply off its highs. The next phase, he says, is deflationary contraction followed by severe stagflation.
Lance Roberts on the same show one day earlier argued the opposite: 5%+ yields are an exceptional entry point for bonds; the broader equity market is deeply oversold under the megacap headline; Q4 brings rotation into value, utilities and small caps. His composition of the US 10-year yield: 2.6% growth and inflation, 1.9% real yield, 1.0% term premium — and he notes the yield sits three standard deviations above its 50-month moving average, a reading reached only about five times since 1994 and preceding sharp drops in yield every time.
On direction, Pento and we are not far apart on equities — both say the broad US equity index settles lower from here. On the Fed cracking the AI bubble, we are closer to Roberts than to Pento: if the Bravos balance-sheet picture tested on Saturday is right, the balance sheet is quietly expanding while the policy rate is hiked, which partly offsets the policy restriction and makes Pento's crash case less supported on mechanism, not more.
On bonds, Pento and Roberts converge on the short end — the US 2-year near 4.8% is Pento's only haven and Roberts's near-sure thing. That is also the record's standing read: short-dated euro and US government bonds over US equity. On long duration, Roberts says buy now; the record says wait for the labour market signal — the Fed still has room to hike without breaking it. The three-standard-deviation flag is real, but in each prior case the yield drop followed labour-market fracture rather than preceding it. The high-yield option-adjusted spread rose to 3.24% this week from 2.80%, 44 basis points on the month — term premium rebuilding into supply, not a credit accident building. The read that would validate Pento is a convergence: the high-yield spread breaking 400 basis points, the US 10-year closing above 5.4% on five consecutive days, and the equal-weighted broad index underperforming the cap-weighted by more than 3% over the same window. One of the three is not enough; three together is.