Amazon is the only one of the big seven that’s been disappointing, unfortunately. I do own it in my portfolio, and it’s been my favorite, but it hasn’t played out too well yet. Every stock has outperformed Amazon, but it’s getting too cheap. It’s trading at 14× price to operating cash flow, which for a mega cap like Amazon is pretty crazy. The company’s involved in everything—healthcare, retail, you name it—and its main drag right now is Amazon Web Services. While AWS is growing around 22% year-over-year, it’s lagging behind Azure and Google Cloud. That could change soon thanks to Amazon’s 30% stake in Anthropic, now worth about $30 billion, which heavily uses AWS. Analysts expect Anthropic to add 100 to 400 basis points of growth per quarter to AWS once Claude 5 ramps up.

Operating cash flow growth has slowed from 82% to 12%, likely due to inventory and tariff effects, but projections show it rebounding to ~20% annually for the next three years. With a valuation under 1× its growth rate, Amazon looks too cheap to ignore, trading at a discount while the rest of the big tech names rally. It may lag for now, but could catch up by 2026.

SoFi is down 7% over the last 5 days, but it’s had an amazing year — up 82% since I started talking about it around $13–14. I bought more at $8 when it dipped, and again at $21 in August, and it’s now sitting at $25. I may even average up again because I believe the stock is still extremely undervalued relative to its potential. The fundamentals have improved massively — earnings per share are expected to rise 150% by 2026.

SoFi now has 11.7 million members, only a fraction of its total addressable market, and revenue growth has accelerated 44% year-over-year with margins improving — adjusted EBITDA margin at 29% and net income margin at 11%. The company’s shift toward fee-based revenue (44% now, expected to reach 80–90%) will reduce risk and increase profitability long term.

Additional tailwinds include the return to crypto, a potential housing rebound as interest rates fall, and strong execution from management. The CEO expects to exceed 25% compounded annual revenue growth from 2023–2026, with EPS between $0.55 and $0.80 in 2026 — likely toward the higher end given their track record of beating guidance. Based on these assumptions, a 40× P/E implies roughly 168% upside over 5 years, or about 21.8% annualized returns. In my view, $25 still isn’t expensive considering SoFi’s growth, margin expansion, and long-term potential.

Revenue dropped 18% year-over-year, margins shrank, and losses deepened, yet BigBear.ai ended the quarter with $391 million in cash, its highest balance ever — only because it raised over $300 million in new equity, heavily diluting shareholders. In other words, they bought time, not traction. Revenue is declining, guidance was pulled, and one of their biggest Army contracts was disrupted. The $380 million backlog looks impressive but is highly concentrated in government contracts, making the business vulnerable to delays or freezes.

The company also wrote off $70.6 million in goodwill and took a $135.8 million hit on derivative liabilities, effectively detonating its balance sheet. Margins are contracting (gross margin down to 25%), losses are expanding, and there’s no evidence of a turnaround.

The bull case: BigBear.ai is embedded in national security, with growing AI defense demand and enough cash to survive for years. The bear case: no moat, no proprietary platform, shrinking revenue, and severe dilution (share count jumped from 251M to 360M). The bottom line — this is a high-risk speculation play, not a long-term compounder. The company has cash and time, but until execution improves and growth returns, it remains a reset story, not a recovery story.