My pick: C, between -5% and +5%. I am bullish on the NAND cycle, but less bullish on the immediate stock reaction. SK hynix confirms that AI infrastructure demand and memory pricing are exceptionally strong, which should be favourable for SanDisk’s enterprise SSD business and margins. The problem is expectations. After the huge run in memory stocks, a normal beat may no longer be enough. I think investors will care more about forward NAND pricing, enterprise SSD growth and margin guidance than headline EPS. For another major leg higher, SanDisk needs to show that AI storage demand is broadening while supply remains disciplined. If management signals that pricing and margins can keep improving into 2027, I would become more bullish. Great earnings are expected now. The real surprise would
My pick: IBM. Among these names, I prefer a company where the upside does not depend mainly on falling rates or a turnaround going perfectly. IBM already has a relatively clear path: grow higher-margin software, expand Red Hat and convert enterprise AI adoption into recurring revenue and free cash flow. MRK also has strong fundamentals, but the longer-term Keytruda patent cliff makes pipeline execution especially important. MMM’s margin recovery is encouraging, although sustaining growth after restructuring will be the next test. I am more cautious on the turnaround group. NKE, BA and HD could have greater upside if conditions improve, but much more has to go right. For 2027, I would rather own earnings improvement that is already visible than pay today for a recovery that still needs to
A. AMZN, but I would not rush to catch the first dip. AWS growing around 39% suggests the underlying AI/cloud demand story remains strong. My concern is the other side of the equation: enormous infrastructure spending and deteriorating free cash flow. I would want evidence that AWS growth can eventually outrun the capex required to support it. I am less tempted by INTC because its weak forward guidance points to an operating problem rather than simply market disappointment. SNDK has excellent AI-storage momentum, but after such a huge run this year, strong results may already be heavily priced in. I do not think the AI rally is necessarily running out of steam. The market is becoming more selective. Earlier, AI revenue growth itself was enough. Now investors increasingly want AI growth +
Free cash flow matters most to me this earnings season. A revenue or EPS beat tells us what happened this quarter, while guidance tells us what management expects. But with AI capex reaching extraordinary levels, FCF helps answer the harder question: is that growth actually generating enough cash to fund itself? Microsoft makes me more confident that AI investment can generate attractive returns, but Alphabet makes me more selective about assuming every AI capex dollar will do so. Strong demand alone is no longer enough. Amazon is the next report I would watch most closely. AWS sits at the centre of the same debate. If AWS growth accelerates while Amazon maintains healthy cash generation despite rising infrastructure spending, it would strengthen the argument that Microsoft’s result is pa
C. Prefer Microsoft for now. Google Cloud doubling is impressive, so I would not interpret Alphabet’s fall as evidence that its AI strategy has failed. The concern is the gap between AI growth and cash returns. With capex surging and quarterly FCF turning negative, Alphabet still needs to demonstrate that this investment produces sufficient incremental cash flow. Microsoft currently gives a clearer picture: strong Azure growth while still generating substantial FCF despite heavy capex. That makes the AI investment cycle easier to justify. I would keep GOOGL on the watchlist rather than automatically buy a 10% dip. If Cloud maintains strong growth while capex growth moderates and FCF recovers, the risk-reward becomes much more attractive. Great AI demand is not enough. Eventually, AI has t
For me, confirmation would require more than another earnings beat. I would want to see HBM demand remain supply-constrained, NAND and DRAM pricing continue improving, and data-centre demand broaden beyond a handful of hyperscalers. SanDisk will be particularly useful for testing whether strength is spreading from AI memory into enterprise storage. Rising enterprise SSD volumes and prices, improving margins and strong free cash flow would strengthen the case. The biggest signal would be capacity growth remaining disciplined despite high prices. Memory cycles often turn when strong pricing encourages excessive supply. If SK Hynix, Samsung and Micron can expand HBM while avoiding an uncontrolled increase in conventional memory capacity, the upcycle could have considerably further to run.
For me, the key is cash conversion, not just AI-driven revenue growth. Amazon needs AWS growth to accelerate without capex consuming the resulting operating cash flow. Intel needs AI/server demand and foundry investment to translate into improving margins and operating cash flow, rather than simply higher utilisation. SanDisk needs data-centre and enterprise SSD demand to support pricing, margins and sustainable FCF. The common test is simple: Is incremental AI revenue growing faster than the cash cost of supporting it? Strong revenue or order growth alone is insufficient. I would watch FCF after capex, operating margins and ROIC over several quarters. If those improve while AI investment remains high, that is much stronger evidence that the AI build-out is becoming a cash-generating cycle
Microsoft gets my vote. Azure is already translating AI demand into strong cloud growth while Microsoft continues generating substantial cash flow, despite heavy infrastructure spending. That makes its AI monetisation path look more mature than peers still seeing capex overwhelm free cash flow. Meta is a close second because AI is improving its core advertising engine, but its capex is also rising rapidly. Amazon’s 39% AWS growth is impressive, yet its TTM free cash flow has swung negative as AI infrastructure spending surged. Alphabet’s Cloud growth is exceptional, but negative quarterly free cash flow makes the near-term return question harder. My ranking: MSFT > META > AMZN > GOOGL for proving that AI capex can translate into sustainable cash returns first.
For the bond portion of my portfolio, I’d prioritise capital stability before chasing the highest yield. A 3.60% YTM with roughly 2.3 years of duration is interesting because it offers a higher prospective yield while limiting interest-rate sensitivity compared with long-duration bonds. But I wouldn’t treat the 3.60% as equivalent to a guaranteed 3.60% return. Credit spreads, defaults, portfolio changes, fees and interest-rate movements still matter. The comparison I’d make is against T-bills, SSBs, fixed deposits and money-market funds. If a short-duration bond ETF offers only a small yield premium, I may prefer the simpler and lower-risk alternatives. If the premium becomes meaningful, taking some duration and credit risk becomes more attractive. For me, bonds should primarily stabilise
B. Hold rates steady. A hold is already heavily priced in, so I think the bigger market mover will be the guidance for December and the reaction of Treasury yields. If the Fed signals another hike, I wouldn’t automatically buy the first tech dip. I’d watch the 10-year yield and earnings first. If yields rise but profitable tech companies with strong free cash flow continue delivering earnings growth, I’d view a valuation-driven sell-off as a potential buying opportunity. But if higher yields coincide with weakening earnings, shrinking margins or AI capex rising much faster than cash generation, I’d stay patient. For me, the opportunity is not simply “tech fell, buy”. It is whether fundamentals remain strong after the discount rate moves higher.