Shyon

I’m cautiously bullish, but I think the market may be getting ahead of itself on the Fed. Waller’s dovish tone is definitely supportive for growth and AI stocks, but tonight’s jobs report could quickly change the narrative. With consensus around just 56K jobs, a weak number could reinforce September easing expectations and push yields lower, while a stronger-than-expected print could bring some of that optimism back down to earth. I’m watching yields and the labour data more closely than the index gains.

For individual names, $Snowflake(SNOW.US) Snowflake’s 16% jump shows that strong AI-driven growth is still being rewarded, while $Broadcom(AVGO.US) Broadcom’s sell-off despite a huge AI revenue target reminds me how high expectations have become. I remain constructive on Broadcom and Tesla over the longer term, but $Tesla(TSLA.US)Tesla’s Cybercab launch still needs to prove it can scale beyond an invite-only rollout before the market fully prices in the robotaxi opportunity. Meanwhile, Singapore banks and the STI $FTSE Straits Times Index(STI.SG) hitting fresh highs are encouraging, and I’d rather continue holding quality positions than chase prices after a strong run.

🟩 BUY THE DIP — I see the Cybercab launch as a positive step for $Tesla(TSLA.US) Tesla’s long-term story. The biggest value isn’t the hardware itself, but the potential to build a scalable Robotaxi network with a purpose-built vehicle designed around autonomy. If Tesla can achieve its low-cost-per-ride target, the economics could be very compelling.

I’m not too concerned about the lack of livestream or the initial Austin-only rollout. For a technology this ambitious, I’d rather see Tesla focus on real-world testing, safety and regulatory approval before aggressively scaling. The fleet-sales model is also interesting because it could allow businesses to participate in the Robotaxi ecosystem and potentially accelerate deployment.

For me, this is still an early-stage story, so I’m not expecting immediate revenue transformation. But I believe the launch moves Tesla one step closer to turning FSD + Robotaxi + Cybercab into a meaningful new business. The overnight dip looks more like a “sell the news” reaction to me than a reason to abandon the long-term thesis. 🚗📈

I think $Broadcom(AVGO.US)Broadcom’s reaction shows just how high the expectations around AI have become. An 86% revenue increase and AI chip sales above $16B are phenomenal, but when guidance comes in slightly below consensus, the market immediately punishes it. At this valuation, “good” is no longer good enough — investors are demanding constant upside surprises.

I’m also becoming more selective on AI infrastructure spending. Circular financing doesn’t automatically mean we’re in an AI bubble, but I want to see massive compute investments translate into sustainable revenue and cash flow. That’s why I’m still bullish on semiconductors over the long run, but I’m comfortable using pullbacks to build positions rather than chasing strength.

For now, I’m watching Friday’s non-farm payrolls closely because the jobs data could reshape expectations for the Fed and growth stocks. At the same time, $NVIDIA(NVDA.US)Nvidia’s strength and $Dell Tech(DELL.US)Dell’s AI-server momentum suggest underlying AI demand remains very strong. For me, the key question isn’t whether AI demand is real — it’s whether earnings can continue growing fast enough to justify the expectations already priced in.

Broadcom’s $Broadcom(AVGO.US) latest quarterly AI semiconductor revenue reached $16.7 billion, up 221% YoY and 54% QoQ. Next-quarter AI semiconductor revenue is expected to rise to $21.7 billion, up 236% YoY.

Broadcom expects AI semiconductor revenue to reach around $58 billion in 2026, $115 billion in 2027, and $230 billion in 2028—and management says actual demand is already higher than these estimates.

Growth is being driven by six frontier AI customers deploying customized XPUs. Google is scaling Ironwood TPU v7 and TPU v8i, with annual TPU supply potentially reaching tens of billions of dollars. Anthropic plans around 1GW of Ironwood this year, 5GW of TPU v8i in 2027 and another 10GW in 2028. OpenAI has started shipping its first customized XPU, Jalapeño, with deployments expected to exceed 5GW by 2028. Meta is also ramping MTIA, with around 3GW of cumulative deployment visibility through 2028.

Across these six customers, potential demand is approaching 30GW, giving Broadcom confidence in roughly $350 billion of AI semiconductor shipments over the next two years. However, the bottleneck is increasingly supply and data-center buildout, including advanced manufacturing, HBM, substrates, power and facilities.

Broadcom estimates that every 1GW of AI compute deployed represents roughly $20–30 billion of semiconductor content. In other words, future revenue growth will increasingly depend not just on AI demand, but on how quickly the industry can bring additional gigawatts of AI capacity online.

My call: −3% to +3%.

I’m bullish on Broadcom’s long-term AI opportunity, especially custom ASICs, networking and its growing exposure to hyperscalers. However, I think a lot of the good news is already priced in after Nvidia’s strong earnings. With AVGO still well below its June high, the market will likely focus heavily on forward guidance rather than just the headline numbers.

Even if Broadcom beats expectations and raises its outlook, I’m not expecting another Nvidia-style breakout immediately. The ~8% implied move shows elevated expectations, while investors may also use a strong report to take some profits. For Thursday, I’m therefore expecting a relatively volatile session but ultimately a −3% to +3% close as the market digests the results and guidance.

The oil $United States Oil Fund LP(USO.US) move is definitely something I’m watching closely, but I’m not rushing to rotate heavily into energy just because of the headlines. A prolonged US-Iran conflict could keep oil elevated and add another layer of inflation pressure, which would make the Fed’s September decision even more complicated. For me, the bigger risk is that higher oil prices push Treasury yields higher and pressure high-valuation growth stocks.

At the same time, Dell’s $Dell Tech(DELL.US) results are a strong reminder that the AI infrastructure story remains very healthy underneath the macro noise. The huge increase in its revenue guidance and record AI-server backlog suggest hyperscaler and enterprise AI spending is still accelerating. I’m particularly watching Broadcom tonight because its results and guidance could give us another important read on custom AI chips and networking demand, especially after the recent volatility across semiconductors.

Overall, I’m staying cautious rather than making a major portfolio rotation based on one geopolitical event. I expect volatility to remain elevated, especially if oil stays above $90 and the 10-year yield continues pushing higher. I’d rather use any meaningful pullbacks in fundamentally strong AI and semiconductor names to DCA gradually, while keeping some exposure to defensive areas in case the Iran situation escalates further.

I’m increasingly cautious about the AI infrastructure spending loop. Anthropic’s $35B Lambda deal is another strong signal that compute demand remains enormous, but with $NVIDIA(NVDA.US)Nvidia-linked capital and infrastructure involved, I think investors should look beyond headline contract values and ask where the ultimate economic risk sits. Circular financing doesn’t automatically mean an AI bubble, but it does make me more selective—I’d rather see AI companies convert massive compute spending into sustainable revenue and cash flow before assuming these valuations are fully justified.

The bigger near-term concern for me is the combination of rising rate-hike odds and elevated Treasury yields. A 64–65% September hike probability, coming so quickly after Warsh’s hawkish tone, could keep pressure on high-valuation growth stocks even if AI fundamentals remain strong. I’m not bearish on AI, but I would avoid chasing strength and prefer using volatility to accumulate quality names gradually. With $Broadcom(AVGO.US) earnings and Friday’s jobs report ahead, I expect the market to remain very sensitive to both rates and AI spending signals.

For me, $NVIDIA(NVDA.US)$442B one-day gain is impressive, but I wouldn’t chase the move blindly. The $108B Q3 guidance and 117% data-centre growth clearly reinforce that AI demand is still very strong, while Singapore’s 6.8% manufacturing growth and 11.2% electronics growth provide an interesting real-economy confirmation. I’d rather add selectively on pullbacks than buy after such a powerful re-rating.

Tonight’s Warsh speech is the bigger near-term risk for me. With PCE still elevated and the labour market relatively firm, I want to hear whether the new Fed chair is leaning toward easier policy or keeping inflation as the priority. I’m staying constructive on AI long term, but I’d keep some cash ready and avoid overexposure until we see how markets digest Warsh’s message.

I’m voting for Up more than 7%. $NVIDIA(NVDA.US) has already gone through seven straight losing sessions, so a lot of caution and negative expectations are arguably priced in ahead of the print. With consensus still looking for around 97% YoY revenue growth and another strong quarter, I think a solid beat could trigger a powerful relief rally.

What matters most to me isn’t just whether Nvidia beats the headline numbers, but whether its forward guidance confirms that AI infrastructure spending is still accelerating. If data-center growth, Blackwell demand and margins come in stronger than expected, the market could quickly shift from “expectations are too high” back to “AI earnings are still delivering.”

I know a 7%+ move in a $5T company sounds aggressive, but Nvidia has repeatedly shown that its earnings can move the entire semiconductor and AI complex. If the numbers and guidance both surprise to the upside, I think the combination of short-covering, sidelined cash and renewed confidence in the AI trade could make a double-digit upside move very possible. My vote: Up more than 7%. 🚀

I’m leaning toward staying invested in AI rather than chasing the inflation trade. Nvidia’s $108B guide and the strong reaction in Salesforce, Okta and CrowdStrike tell me the AI story is broadening beyond chips into software and recurring revenue.

That said, I’m not ignoring inflation. With PCE still elevated and Warsh speaking Friday, I expect volatility to remain high, so I’d rather add selectively on pullbacks than chase the post-earnings rally.

For Singapore, the 2% core inflation reading is definitely worth watching, but strong exports and surging electronics demand give me some confidence. For now, I’m staying focused on AI growth while keeping some dry powder for the next Fed-driven pullback.