Microsoft beats clean, Alphabet stumbles as ETF investors buy the AI-dip

Marcus Weyerer, Director of ETF Investment Strategy, Franklin Templeton, said: “Microsoft and Alphabet’s latest earnings reinforce our view that the AI investment cycle remains broadly intact, but progress from here is likely going to be less uniform than last year. Microsoft reported Azure growth of 48% y-o-y (vs. 45% guided), which is a key indicator of AI monetisation and that capex can earn a return. It also beat on the top and bottom lines. Alphabet’s Google Cloud expanded 103% y-o-y (vs. 82% in Q2), equally demonstrating substantial commercial demand for AI.

“The investment required to sustain that growth remains extraordinary. Alphabet’s capex is soaring from projected ~US$200bn this year to ~US$300bn in 2027, meaning FCF will likely remain negative for the foreseeable future, testing investors’ patience. The firm also reported lower-than-expected EPS as its investment portfolio value took a hit, sending the stock lower double digits after-market. Microsoft now expects to spend US$170bn in 2026 – approx. US$5bn less than previously guided but still a substantial amount and expected to increase further, although management also expects FCF to remain positive. This should alleviate some concerns in the market about the sustainability of capex. The question has long moved from whether AI can attract investment to which companies can turn that investment into profits. Based on today’s reports, Microsoft and Alphabet are demonstrating growing evidence of monetisation, but in our view, the jury will still be out for several quarters to come.

“Meanwhile, our proprietary ETF flow analysis suggests that investors continued to buy into recent market weakness rather than abandoning the AI trade. Developed-market technology ETF flows recovered from 2.4 standard deviations (σ) below their historical trend in August ($3.4 billion in outflows) to marginally below trend in September ($5.3 billion in inflows, -0.4σ) and 1.3σ above trend in October month-to-date ($15.8 billion in inflows). If sustained through month-end, October’s reading (+1.3σ) would surpass April as the strongest reading this year. The semiconductor selloff earlier in the month may have shaken share prices, but it has done little to undermine ETF investors’ appetite for AI exposure.

“We observe a similar pattern in technology-heavy Asian markets, whose semiconductor industries are closely linked to hyperscaler spending. Taiwan-focused ETF flows recovered from -1.2σ in August ($1.7 billion in outflows) to -0.4σ in September and +1.0σ in October month-to-date ($10.9 billion in inflows). South Korea also improved markedly, from -1.7σ in September ($2.6 billion in outflows) to essentially at-trend in October ($5.5 billion in inflows).

“Bigger picture, we have now seen three of the four hyperscalers report convincing AI business growth; upstream, SK Hynix just offered a record print, and Anthropic’ s S-1 filing contains clues as to rapidly expanding revenue opportunities further downstream. Looking towards year-end, we remain constructive but see reasons for greater portfolio diversification. Fed policy, US election uncertainty and potential geopolitical or energy-market disruptions could all contribute to volatility. US mega-caps remain indispensable sources of structural growth. But the tide may not lift all boats anymore. Equity income and actively managed strategies can complement higher-beta exposures and help manage concentration risks while staying fully invested. The AI trade is far from over, but its next chapter may reward selectivity as much as conviction.”