The most interesting takeaway from this earnings season is that AI demand isn’t slowing—the market is simply becoming more selective. CRWV and NBIS were rewarded because their numbers show real demand: accelerating revenue, massive backlogs, and improving profitability. SMCI also benefited because AI demand is translating directly into stronger revenue and margins. Meanwhile, COHR and CBRS tell the other side of the story. COHR delivered a strong quarter, but after a huge run-up, a simple beat was no longer enough. CBRS had impressive future commitments, yet investors focused on weak hardware revenue. That tells us where the market is heading: AI stories are cheap. AI earnings are valuable. Going forward, I’d focus less on who has the most exciting AI narrative and more on who can convert

SNDK can reach $2,200, but I wouldn’t chase it blindly. What makes this rally different is that the story is shifting from simply “NAND prices are going up” to better earnings visibility, supply discipline, long-term contracts and AI inference potential. If management can deliver the targeted margins and FCF while HBF becomes a real product by 2027, the market could start valuing SNDK less like a traditional cyclical memory stock. But after a 467% YTD rally, expectations are already sky-high. At this level, the risk isn’t that SNDK has a bad business—it’s that the business performs well while investors expect perfection. So I’m closer to B: bullish, but waiting for a pullback. For me, $2,200 is achievable, but the next 30–40% won’t come from hype. It has to come from real earnings growth,

The STI rally looks impressive, but I think the easy money has already been made. A 24% YTD gain is difficult to ignore, especially when banks have been doing much of the heavy lifting. Yangzijiang’s record earnings show that this isn’t purely a liquidity-driven rally — real earnings are supporting parts of the market. But that also raises the bar. After such a strong run, valuation expansion alone probably won’t be enough. The next leg needs stronger profits, dividends and guidance. I’d be more selective here: banks for cash flow, shipbuilders like Yangzijiang for earnings momentum, and REITs if the rate environment becomes friendlier. My biggest takeaway: the STI may still have upside, but 2026 is shifting from an index-buying market to a stock-picking market.

Buffett’s “Slow Philosophy” Still Works The market is chasing AI, semiconductors and quantum computing, but Buffett’s portfolio offers a different lesson: durability beats hype. Coca-Cola and Apple may look boring compared with high-flying tech stocks, yet their strong brands, pricing power and recurring cash flow give them powerful moats. Berkshire’s move into Alphabet also shows that Buffett’s philosophy isn’t about avoiding technology—it’s about finding technology with a durable moat and strong cash generation. For ordinary investors, I’d keep it simple: 1️⃣ Moat — Can competitors easily replace the business? 2️⃣ Cash Flow — Does it consistently generate real money? 3️⃣ Long-Term Value — Would you still own it if the market closed for five years? The biggest lesson? Don’t chase whateve

A. SNDK — The story is bigger than NAND. I’d still pick SNDK. What caught my attention isn’t just the 13.7% rally, but the potential change in its business model. Long-term agreements could make earnings and cash flow much more predictable, while HBF gives SNDK another angle on the growing AI inference market. MU has broader exposure across HBM, DRAM and NAND, and WDC offers an interesting data-center HDD story. But SNDK currently has the most interesting combination of AI storage demand, long-term contracts, high-margin targets and shareholder returns. The only thing I wouldn’t do is chase the stock after a huge one-day move. At this valuation, expectations are already high. For me, the real test is whether SNDK can hold the gains and prove that those ambitious 2030 targets are actually

I lean toward B. To me, this looks more like a healthy reset in expectations than the beginning of a new downcycle. After such a strong rally, memory stocks were priced for near-perfect execution, so even solid earnings and guidance weren't enough to satisfy investors. The bigger question isn't whether NAND is slowing—it's whether that weakness spreads to DRAM and HBM. So far, AI demand hasn't changed. Hyperscalers are still investing aggressively, HBM supply remains tight, and AI servers continue to require more high-performance memory. That's why I think Micron is in a different position from pure NAND players. Its AI growth is increasingly driven by DRAM and HBM rather than NAND alone. Unless we start seeing analysts cut DRAM/HBM forecasts or AI capex slows meaningfully, I'd view this

My favorite stock from this week’s list is $RKLB.  I think Rocket Lab has an interesting long-term story, not only because of its launch business, but also its Space Systems segment and the potential of Neutron. The valuation is not cheap and execution risk is still high, but if the company continues to deliver, I believe the upside could be substantial. I also like $CSCO as a more established choice. AI data centers are creating strong demand for networking infrastructure, so I’ll be watching its EPS, revenue growth and management guidance closely. For dividends, $IBM stands out to me because it offers a combination of income and exposure to AI/software growth. Overall, I don’t think investors should focus only on whether EPS beats estimates. Guidance, margins, cash flow and future

8,000 Is Possible, But Earnings Must Deliver I’m bullish on the S&P 500 reaching 8,000, but the reason isn’t valuation—it’s earnings. Goldman’s thesis is straightforward: EPS growth, especially from AI infrastructure, can push the index higher without requiring a much higher P/E. Cooling inflation also gives the Fed room to stay patient. But there’s one big catch: expectations are already very high. If AI spending keeps translating into real profits, 8,000 looks achievable. If oil spikes, the Fed turns hawkish, or AI earnings disappoint, the rally could lose momentum quickly. So I wouldn’t chase the headline target. Watch earnings, inflation and Fed policy. If those three stay friendly, the bull market still has fuel. @Capital_Insigh

I’d pick B) Cybersecurity. AI is creating a huge productivity wave, but it’s also expanding the attack surface across cloud, identity, data and AI agents. As enterprises deploy more AI, cybersecurity becomes less of a discretionary expense and more of a necessity. Among the names, PANW would be my top pick. Its platform-consolidation strategy, strong ARR growth and exposure to multiple areas of enterprise security give it a compelling long-term setup. CRWD is also attractive, especially with its strong platform ecosystem and recurring revenue model. That said, both stocks have already rerated significantly, so I wouldn’t blindly chase new highs. I think cybersecurity still has legs, but the next phase will need to be supported by earnings, cash flow and sustainable growth—not just the AI

I think Nvidia’s $500B financing push is more genius than gamble — at least for now. Jensen Huang is effectively bringing Wall Street capital into the AI infrastructure boom without putting the entire burden on Nvidia’s own balance sheet. More financing means customers can build more data centers, buy more GPUs, and accelerate AI deployment. That creates a powerful cycle: capital → infrastructure → Nvidia chips → AI revenue. But the risk is obvious. If AI data centers struggle to generate enough returns, leverage could work in reverse, putting pressure on lenders, infrastructure valuations and eventually Nvidia’s growth expectations. So I wouldn’t call this a circular bubble yet. I’d call it a massive bet on AI economics. My view: Jensen may have found a brilliant way to scale AI demand —