Show_down

AVGO Q3 guidance: revenue $29.4B (+84% YoY), AI chips $16B (+200%+ YoY). The underlying demand is real. The problem is the stock priced in $18-20B AI revenue and only got $16B. That 10-20% disappointment vs buy-side bar = 12.9% selloff. The business is fine. The valuation was stretched.

With the AI infrastructure trade setting records every day this week (MRVL +32%, HPE +20%, NVDA +6%), does Broadcom HAVE to absolutely smash expectations tonight to keep the momentum going?? Or is a solid inline beat enough?? 🙋 wondering how high the bar actually is

My colleague who switched to "diversified value investing" in January to avoid the AI hype is having a very quiet day. Nasdaq up to 27,000 first time ever, AI software +8.7%. Sometimes the hype is the right call lah 😏

When S&P hits ATH but 8/11 sectors fall, do you treat it as a healthy AI mega-cap rally or an early warning sign of concentration risk? I genuinely can't decide which camp I'm in right now 🤔 what's your read lah

TSMC raising 3nm prices 15% this H2 then another 10% in 2027, and Apple, NVIDIA, AMD are all just nodding and paying. Meanwhile I'm already calculating how much my next MacBook Pro is going to cost next year 😭

The narrative around Meta's layoffs is broadly positive: cut headcount, redirect capital to AI, unlock efficiency. The stock market has historically rewarded this playbook. But I think it's worth asking whether the evidence actually supports the bullish framing, or whether we're pattern-matching to 2022 without examining what's different this time.

The case for the bulls

The 2022 "Year of Efficiency" is the template. Meta cut roughly 25,000 roles over 2022-2023. The stock went from around USD 90 to over USD 500. Operating margins expanded dramatically. The market rewarded every cut.

The logic for today is similar: AI systems can handle tasks that previously required human headcount. A leaner team working alongside AI can be more productive. Reinvesting the savings into AI infrastructure accelerates the competitive position.

Meta's financials support the context. USD 56 billion in quarterly revenue and strong operating margins mean these cuts are coming from a position of strength, not desperation.

The case for scepticism

What's different now is the scale of the AI bet. Meta is committing USD 115-135 billion to AI infrastructure in 2026 alone. This is not reallocating savings from headcount cuts. This is a massive capital expenditure programme on top of the cuts. The two things are related but not equivalent.

AI productivity gains are also unevenly distributed across job functions. Engineering roles augmented by AI coding tools see measurable throughput improvements. Roles in legal, policy, communications, and market operations are harder to replace or augment at scale. How Meta's specific mix of cuts maps onto genuinely AI-augmentable tasks is not yet public information.

There is also an institutional knowledge cost that doesn't appear in short-term earnings. When you cut 10% across the organisation, you lose a disproportionate amount of context, relationship capital, and tacit knowledge that took years to accumulate. This compounds quietly over multiple years and is very difficult to quantify until it's already happened.

What I'd be watching

It's too early to know whether the productivity thesis plays out at the scale Zuckerberg is betting on. The 2022 cuts were cost discipline. The 2026 cuts are a structural transformation thesis. Those are different claims requiring different evidence.

Revenue per employee over the next four quarters is the metric worth tracking. If it rises significantly faster than peers who haven't done equivalent cuts, the AI productivity thesis is gaining real evidence. If it doesn't, the narrative deserves more scrutiny than it's currently receiving.

The market may be right to be bullish. The history supports it. But "history supports it" and "the thesis is proven" are not the same thing.

Cerebras opened at $350 and closed Day 1 at $311.07 (+68.15%), valuing the company at approximately $95 billion. The $5.55B raise on 30M shares marks the largest US tech IPO since Uber's 2019 debut. T...

My take on Nvidia GTC

Anyone expecting a stock pop just from Jensen's keynote? That's pure hopium.

Here's the real debate Wall Street's wrestling with: Why isn't the stock running despite insane earnings?

☞ Two big question marks:

How long can AI infra spending stay hot?

Can anyone crack Nvidia's monopoly?

☞ So this GTC, Jensen needs to address:

Scenario A → Crystal-clear earnings visibility for 2027+ (earnings drive price)

Scenario B → Unshakeable dominance narrative (multiple expansion drives price)

Either way, market needs one of these to latch onto.

My take on the Israel-Iran conflict: why it stays contained, and my $NVIDIA(NVDA.US) options strategy for the aftermath. NOT ADVICE!!!

$SPDR S&P 500.US just sitting at $681.75 after that run from $481 low last year . $NVIDIA.US 7.61%, $Apple.US 6.94%, $Microsoft.US 5.03% making up 20% of the entire fund . When Mag 7 sneeze, SPY catches cold. But 503 holdings total means diversification still works. Just DCA and ignore the noise.

Bloomberg says $Coinbase(COIN.US) revenue could 7x under GENIUS Act. Monness says short with $120 target. Armstrong says if rewards get banned, "we'd actually be more profitable". Stock at $160 after -6.5% day. Three views, one stock. I know which one isn't priced in. 🎲

$AMD(AMD.US) just hit 41.3% server revenue share — highest in company history. Desktop revenue share 42.6%, up 14.6 points YoY. Q4 rev $10.27B, EPS $1.53 beat. First Shanghai raised target to $300. Stock at $205.94. Market focused on GPU catch-up, I‘m focused on EPYC eating the world. 🐎

Marex just added 560% to their $BitMine Immersion Tech(BMNR.US) stake, now 10M+ shares at ~$198M value . BlackRock also up 165% to 9M shares . Two of the biggest institutional players quietly doubled down in Q4 while retail was panicking. Stock at $19.74. Smart money rarely this loud. 🐋

$Sandisk(SNDK.US) tonight it doesnt even bother to rebound...

The core logic behind $Taiwan Semiconductor(TSM.US) recent performance remains the continued surge in AI demand, transforming it from a traditional cyclical stock into a kind of "quasi-infrastructure."

Many still understand it through the lens of smartphone and PC cycles, but what truly impacts valuation now is advanced process capacity and advanced packaging. Issues like CoWoS capacity shortages are essentially bottlenecks in AI computing power supply, not typical semiconductor inventory cycles.

The problem is that the market has already partially priced in long-term AI growth; going forward, stock prices will depend more on execution, such as N2 progress, yield rates at the US plant, and capital expenditure control.

The key issue isn’t whether taxes go up overnight, but how services gradually move from “growth” to “mature” tax brackets.

The key issue isn’t whether taxes go up overnight, but how services gradually move from “growth” to “mature” tax brackets.

Does Coinbase’s fintech narrative outweigh short-term crypto price risk?

Bitcoin sliding under $80,000 sucks for Coinbase’s trading fees, but the firm’s subscription and custody service diversification could soften the blow if traders stay.

Compared with flashy AI or meme trades, $Circle(CRCL.US) feels like a slow, infrastructure-style bet that only pays off if crypto adoption keeps maturing.

CRWV's latest model got decent reviews. But competing with Chinese EVs is a nightmare

Do you think AMD can meaningfully gain AI market share, or will NVIDIA’s ecosystem keep the gap wide?

$Coreweave(CRWV.US) Watching for a bounce, but wouldn't dare catch this falling knife

$Coreweave(CRWV.US) If you believe the AI compute shortage lasts for years, this is a direct bet. High risk, high reward.

$Coinbase(COIN.US) Earnings season coming, will be a check on retail engagement