LearnBits

Big Tech’s relentless spending on artificial intelligence infrastructure is starting to reshape the profit picture in ways that worry even the most bullish analysts at Goldman Sachs. Hyperscalers such as Microsoft, Amazon, Alphabet and Meta are pouring hundreds of billions into data centers, chips and related gear, with projections nearing $770 billion in capital expenditures across the sector for 2026 alone, roughly matching their entire operating cash flow.
$AMZN
This capital intensity marks a sharp departure from the asset-light model that powered these companies for years. Higher depreciation charges and declining sales-to-asset turnover are already pressuring return on equity at names like Microsoft and Amazon. Goldman points out that while current earnings look strong, the bill for all this buildout could weigh on future margins and shareholder returns if monetization lags behind the spending.
$MSFT
At the same time, investor expectations have raced ahead. AI-exposed stocks such as Nvidia, Broadcom and the major cloud providers now trade at premium valuations, and the S&P 500’s performance has become dangerously concentrated in this small group. Goldman sees signs of market exuberance but stops short of calling it a full bubble. Still, any slowdown in AI investment or disappointing revenue growth from these massive outlays could trigger a painful repricing.
$NVDA
The caution comes as the AI trade shifts from hype to heavy lifting. For companies like these hyperscalers and their suppliers, the real test will be whether the enormous upfront costs translate into sustainable profits before the momentum fades.

Investors pulled more than a billion dollars from Bitcoin exchange-traded funds in a single week recently, marking the largest weekly exit of 2026. At the same moment, capital flowed steadily into XRP products and surged into brand-new funds tracking Hyperliquid’s $HYPE token.
This isn’t a broad retreat from crypto. It’s a rotation. After years of Bitcoin dominating institutional portfolios, allocators appear hungry for assets that offer real activity and revenue. Hyperliquid, a decentralized perpetuals exchange, has captured attention with its high trading volumes and token buyback mechanics that funnel fees back to holders. Its ETFs have already drawn over $150 million in net inflows shortly after launch, outpacing early Bitcoin products on a size-adjusted basis.
$XRP funds, meanwhile, continue their steady climb. They added nearly $23 million in one recent week, extending a streak of positive flows even as the broader market wavered. Cumulative inflows for XRP ETFs now exceed $1.4 billion since their late 2025 debut, reflecting persistent belief in Ripple’s utility for cross-border payments.
The shift highlights a maturing market. Big players who once piled into Bitcoin as the safe entry point are now testing higher-upside names tied to actual usage. Whether this rotation sparks broader altcoin strength or remains a niche move will shape the next leg of crypto’s cycle. For now, the flows tell a clear story: diversification is back in fashion.

The Dow Jones US Dividend 100 Index is up about 10% in 2026, quietly beating the major indexes while everyone else stayed glued to AI headlines. That gap between attention and performance is exactly the setup worth understanding before chasing dividend yields blindly.
The instinct to just buy the highest yield available is the first mistake. Sunoco (SUN) currently pays a 6.88% yield, above its sector average, (Changelly) but a number that high usually means the market is pricing in real risk, not free money. Sustainable income comes from companies with the cash flow to keep paying and raising that dividend for decades, not the ones simply offering the biggest number today.
That's where Dividend Kings earn their reputation. Procter & Gamble (PG) has raised its dividend annually for 70 years, (CoinGecko) a streak matched by only a handful of companies anywhere, and it still plans to return $15 billion to shareholders through dividends and buybacks this year alone. PepsiCo (PEP) has extended its own streak to 54 consecutive years, (WorldCoinIndex) even while trading at a discount tied to unrelated fears about weight-loss drug demand denting snack sales, a disconnect that's arguably created an entry point rather than a warning sign.
Real estate and infrastructure plays round out the picture differently. Realty Income (O) owns more than 15,500 properties globally and pays monthly rather than quarterly, appealing to investors who want income that mirrors a paycheck rather than a quarterly lump sum. Brookfield Infrastructure (BIPC, BIP) has grown its payout at a 9% annual clip for 17 straight years, backed by contracts indexed to inflation rather than consumer sentiment, which matters in an environment where rate expectations keep shifting.
None of this is about picking the single best stock and walking away. It's about understanding why a company can keep paying, not just whether it currently does, because a dividend that gets cut erases years of compounding in a single announcement. The companies quietly compounding through five and seven decades of economic cycles tend to reward patience over excitement, and 2026 is shaping up as a reminder of exactly why that boring approach keeps working.
$MUB
$NVDAB
$SPCXB

#SamsungSKHynixSharesRiseYTD
For 26 unbroken years, Samsung Electronics held the top market cap spot on South Korea's KOSPI. On June 22, SK Hynix took it away, briefly. Shares jumped 5.6%, pushing the company's common stock valuation to 2,080 trillion won, about $1.35 trillion, edging past Samsung's 2,067 trillion won for the first time since November 2000.
It lasted roughly one trading day. On June 23, both stocks fell more than 12% in a KOSPI crash severe enough to trigger circuit breakers twice, and Samsung's lead was restored by the afternoon close. Samsung also disputes the original ranking on a technicality, arguing its market value should include preferred shares, a calculation that would have kept it on top the whole time.
The brief crossing matters less than what caused it. $SKHYNIX has become essentially a pure bet on AI memory, with virtually every dollar of revenue tied to chip sales and no other business line diluting the swings. That focus paid off in spectacular fashion this year. Shares are up somewhere between 250% and 300% year to date depending on the tracker, after first-quarter results that set records across nearly every metric: revenue of 52.6 trillion won, up 198% year over year, and an operating margin of 72%, reportedly higher than Nvidia's own margin over the same stretch. High-bandwidth memory now makes up more than 40% of the company's total revenue, and SK Hynix controls something like 57% to 62% of the global HBM market.
Samsung's rally looks different by comparison, climbing somewhere between 158% and 163% this year, still enormous but more muted next to SK Hynix's run. That gap traces back to business structure rather than any execution gap. Samsung spreads its revenue across foundry work, logic chips, displays, and consumer electronics alongside memory, which softens the cyclical swings but also dilutes the pure upside SK Hynix is capturing right now.
What both companies are actually riding is a DRAM shortage that's reshaping pricing across the entire industry. Contract prices for DDR5 memory surged 90% to 95% quarter over quarter in early 2026, according to TrendForce, and that shortage is expected to persist through at least mid-2027. SK Hynix has reportedly chosen to slow its shift toward next-generation HBM4 production specifically to keep harvesting those elevated DDR5 margins a while longer, a decision that keeps enterprise memory prices high for buyers who'd assumed costs would normalize by now.
None of this comes without risk. Sharp pullbacks like the one on June 23 are a reminder that valuations built on a few months of extraordinary pricing power can unwind just as fast as they built up. But for now, the memory shortage underwriting both stocks shows no sign of easing, and the real question for investors isn't really Samsung versus SK Hynix. It's whether the entire memory chip supercycle keeps running at this pace, or whether 2026's record margins turn out to be the peak rather than the new normal.