Questions around AI continue to dominate headlines.
Last week saw Nvidia CEO Jensen Huang working hard in a CNBC interview to shore up market support for neoclouds.
But on the negative side, we saw i) Microsoft pull out of an anchor tenant lease at a planned AI factory in the UK on electricity delays, ii) neocloud Coreweave’s debt continue to trade down, and iii) AI-related capital market activity stalling and falling with public listings being delayed and cancelled. Elice, a South Korean AI cloud and data-centre provider, did proceed with its listing but fell over 30% on its first day of trading.
Meta’s quarterly results provided some very encouraging indicators on the benefits of and demand for AI, while SpaceX’s result beat expectations for the quarter and delivered guidance ahead of the market, driven by a surge in AI revenue.
However, at this point the market is focused on the funding side of the equation.
Elsewhere US quarterly earnings season remains broadly constructive, notwithstanding some pockets of cost inflation across energy, freight and other core inputs. Consumer and credit trends remain healthy and travel demand strength continues to surprise on the upside.
This week will be a big one, with earnings releases for heavyweights like Microsoft, Alphabet, Bank of America, Coca Cola, Visa and Amazon.
The S&P 500 shed 0.5%, but market movement was more a reflection of rotation rather than outright selling, as the equal weighted S&P 500 outperformed by 160 basis points (bps). The S&P/ASX 300 was down 0.2%.
US 10-year bond yields fell 8bps. Fixed income performing as a defensive asset during equity market weakness was a return to old school norms; let’s see for how long.
Brent crude oil rose 6.2% and is back well over US$110 despite flows returning close to prewar levels, driven primarily by the growing view that the conflict escalates post the US midterm election. Gold fell 4.0%.
Comments from the RBA on inflation shifted the domestic market, in contrast to US Federal Reserve officials who are in blackout ahead of the Federal Open Market Committee (FOMC) meeting this week.
It is a big week for central bank enthusiasts, with meetings at the Fed, European Central Bank, Bank of Japan, Bank of Canada and Reserve Bank of New Zealand, with the latter two priced for interest rate hikes.
Macro & policy Australia
RBA Deputy Governor Andrew Hauser was blunt and direct at last week’s Citi conference. His message: the inflation outlook is materially worse than the August projection.
The situation has deteriorated in each of the three key drivers of inflationary pressure since August:
- The Iran conflict has become more prolonged than expected and the oil price is 30% higher.
- AI demand has continued to grow, with potential upside to previous projections.
- There has been no relief on domestic capacity pressures.
He noted the threat of unanchored inflation expectations as “the central banker’s nightmare”.
In this regard they are “beginning to see some signs that some people in the community are not sure if we’re going to bring inflation back to target at all, or certainly not for an extended period”, and that “an increasing share of businesses” are perhaps seeing 3% to 3.5% inflation as the new normal.
He was adamant that the RBA can’t allow expectations to shift away from target, meaning hard choices and “laying shorter-term pain on top of pain” to prevent worse outcomes down the track if higher inflation expectations are locked in.
The implied chance of a rate hike at the meeting on 2nd/3rd November shifted from 27% to ~60% following Hauser’s comments.
It is interesting to note the recent impact of rates on the consumer. While foot traffic growth is strong over six and 12 months to the start of October in national chains including JB Hi-Fi, Flight Centre, Kmart, SuperCheap Auto, Rebel and Chemist Warehouse, there has been a material slowdown in the wake of the most recent rate hike by the RBA.
Macro & policy US
Housing
New housing starts dropped to 1.221 million in September which was down from 1.288 million in August and below consensus expectations of 1.305 million. New building permits at 1.400 million were a touch softer than the 1.403 million in August, but ahead of 1.385 million expected.
However, the key issue was pending home sales, which fell 7.3% in September, versus -0.5% expected.
The rapid increase in mortgage rates is flowing into transaction activity. That September figure for pending home sales implies existing home sales will fall to ~3.7 million in October – the lowest level since August 2010.
Notwithstanding that home prices have not really moved for three years, a new mortgage at current rates on the median house price would consume 40% of average after tax income, up from 35% in January.
Homebuilders are already sitting on considerable excess inventory. There is a stockpile equivalent to 8.7 months of sales of unsold new single-story family homes in the three months to August, compared with a long-run average of around six months.
Most homebuilders have ignored pressure from the Trump Administration to create more supply and have chosen to preserve margin instead.
It seems a revival in US housing activity still appears to be some way off.
Employment
The labour market remains in good shape. Initial jobless claims declined to 194,000 in the week ending 17th October, from 200,000 the previous week and below consensus expectations, also at 200,000.
Continuing claims were at 1.728 million in the week ending 10th October, from 1.703 million the prior week and above consensus at 1.710 million.
Despite higher energy and borrowing costs, firms continue to hold on to labour.
Forward indicators of job losses, such as the WARN measure, remain contained while hiring intentions remain soft, meaning the labour market remains stable.
It also appears that there has been a significant colling of concerns around near-term job-shedding as a result of AI.
AI-related layoff announcements have equated to less than 5,000 jobs in both August and September, after a spike of just under 40,000 in May.
There is some evidence that jobs growth in the fields most associated with AI have seen 10,000/month less jobs created than trend – but this has been offset by higher job creation in the data centre industry.
US earnings
Thus far we have seen 24% of S&P 500 companies report.
- The blended earnings growth rate for Q3 EPS is 29.5%, near the 29.6% expected at the end of the quarter.
- The blended revenue growth rate is 12.7%.
- 83% of companies have beaten consensus EPS expectations (versus 78% five-year average.
- 66% have beaten consensus sales expectations (versus five-year average of 70%).
- In aggregate, earnings are 2.1% above expectations, down from 13.8% one-year average and 7.5% five-year average.
- Aggregate sales are 1.2% ahead of consensus, versus 2.1% one-year average and 1.8% five-year average.
Meta
Meta reported and is something of a poster child for the benefits of AI adoption both internally and by its consumers. Enterprise use cases and the benefits of AI in Meta’s core businesses are unambiguous.
– It is taking huge share in advertising markets, growing at 26% – which is twice the industry growth rate – with global average price per ad up 14%.
– This is due to AI-driven content and can be seen in the use of its various platforms.
– Instagram has surpassed 3 billion monthly active users, with average time spent per user growing double digits and at all-time highs.
– Reels have also hit all-time highs, while Threads now have more monthly active users than X (previously Twitter).
– More than 10 million people in the US and Canada set up Muse personal agents in its first six weeks, with 2.5 million using it on any given day. Shopping is the leading use case, with 25% of weekly activity, with average purchases over US$200. This provides early evidence of commercial utility.
The handbrake on Muse is compute intensity, which is constraining global rollout schedules. Meta is targeting greater than 10x improvement in compute cost efficiency over the next year via more efficient models, inference optimisation and better utilisation of off-peak capacity,
Its goal is uniting the ~500 million small-to-medium businesses on the Meta platform with the user base to generate agent-to-agent commercial activity to rival the strength of Alphabet, which has been a player in this space for such a long period of time.
Meta’s comments on FY27 earnings were vague and left the market with the view that opex and capex estimates – currently US$140 billion for FY26 and ~US$210 billion for FY27 – need to be lifted.
The company also flagged a distribution of Class C (non-voting) shares as a precursor to tapping equity markets in CY27 for funding of the ever-growing business plan. Meta’s credit default swaps moved up, and the stock was off 8% the next day.
Space Exploration Technologies (SpaceX)
SpaceX also gave some line of sight on AI demand.
It grew Q326 revenue 161% year/year to US$14.6 billion while adjusted EBITDA was US$8.8 billion, up 306% year/year. This was 14% and 30% ahead of consensus, respectively.
It guided to 1Q27 revenue guidance of US$33-37 billion (versus consensus at US$23 billion) and expects to exit 2027 at an average recuring revenue (ARR) of US$500 billion-US$1 trillion per month – well ahead of consensus expectation.
SpaceX finished Q3 with a nameplate compute capacity of 1.7 gigawatts (GW), saying it has line of sight to 2.5-3.0GW by the end of CY2026 and is confident of adding 10GW in 2027.
This would leave it with ~13GW, versus consensus expectations at 6.1GW.
The company has flagged building 20GW worth of power and cooling capacity in 2027, to prevent this being a constraint.
Pricing for its compute is tracking slightly above the high end of its US$30-50/watt range – with recent prices coming in higher, not lower.
SpaceX is the largest provider of incremental compute and at this stage appears to have significant pricing power. That said, it is pursuing a very short-term rental model in preference to long-term contracts, so would be exposed to a softening demand environment.
The company’s rough goal is a half-and-half split between internal use and third-party rentals of its compute capacity.
The company is guiding to 4Q26 capex of US$35-40 billion, versus consensus at US$21 billion. On its current run rate of spending US$36 million per megawatt (MW), its plan to add 10GW in 2027 implies ~US$360 billion of capex, versus consensus at US$140 billion.
Management pointed to a sub-one-year payback on compute spend, operating cashflow that increasingly funds the build form 1H27, and multiple pools of debt capital while protecting its investment grade credit rating.
At current pricing the whole thing stacks up. However free cash flow gets materially more negative before it improves and, when asked, the company has flagged a mix of debt, vendor financing and customer prepayments to bridge the funding gap – and also flagged the importance of its investment grade rating.
Markets
Takeover activity saw Cleanaway (CWY, +21.1%) and Challenger (CGF, +21.0%) among the best performers in the S&P/ASX 100 last week. NextDC (NXT, -38.9%) was the weakest on an earnings downgrade.
The latter saw Technology (-4.7%) along the weaker sector, though Resources (-7.2%) fared worse as iron ore, copper and lithium prices remained soft. Communications Services (+4.0%) and Financials (+4.0%) did best.
Amcor (AMC, +6.3%), Transurban (TCL, +3.8%), APA Group (AP, +2.6%) and Lottery Corporation (TLC, +3.1%) benefitted from rotation to defensive Industrials in the US market as AI-related stocks sold off. AMC was also helped by stronger than expected volume growth at some of its customers including Proctor & Gamble and Nestle.
IMPORTANT INFORMATION: This document has been prepared by Enhance Financial Partners, ABN 45 146 707 173 AFSL 515518, based on our understanding of the relevant legislation at the time of writing. While every care has been taken, Enhance Financial Partners makes no representations as to the accuracy or completeness of the contents. The information is of a general nature only and has been prepared without consideration of your individual objectives, financial situation or needs. Before making any decisions, you should consider the appropriateness for your personal investment objectives, financial situation or individual needs. We recommend you see a financial adviser, registered tax agent or legal adviser before making any decisions based on this information.