NeutralAMZNAmazonAmazonGOOGAlphabetAlphabetGOOGLAlphabetAlphabetMETAMSFTMicrosoftMicrosoftORCLOracleOracle

1w ago on

$AMZN $GOOG $ORCL $MSFT $META All going today Read more here: https://t.co/VkTTZJI4VK https://t.co/bQG4TYTR2Y

BullishAMZNAmazonAmazonGOOGAlphabetAlphabetGOOGLAlphabetAlphabetMETAMSFTMicrosoftMicrosoftORCLOracleOracle

1w ago on

$AMZN Always encouraging to see this price action after a rough ER reaction I continue to believe that $ORCL $GOOG $AMZN $META and $MSFT will present nice opportunities in the long term. My strategy is and has been to slowly DCA into each of them Latest add on $AMZN was at $225 on the ER dip

NeutralAMZNAmazonAmazonLITELumentum Holdings Inc.Lumentum Holdings Inc.MRVLMarvellMarvellMSFTMicrosoftMicrosoftNVDANVIDIANVIDIAORCLOracleOracleSPCXSpaceXSpaceX

1w ago on

Optical Earnings, AI Spending and the Midterms: What Could Move Stocks This Week

Last week, I wrote that the AI trade was starting to feel fragile. This week only strengthened that feeling. Companies are still spending enormous amounts on AI infrastructure, but investors are asking harder questions about when that spending will pay off and how long the growth will last. So my goal is to continue to find the best positioned AI companies at the best prices while also exploring and gaining exposure to industries outside of tech. This is what I want to cover today: A look back at what I covered the past 6 days OpenAI news Hyperscaler Capex projections + implications Midterms & their influence on the markets + a basket AI Medicine / Healthcare companies that interest me that have been performing well Big Week For Optics (Earnings + LITE/MRVL) A Discussion on My Portfolio + Plans Charts + Tickers & Levels of Interest This is going to be a very long and comprehensive article. Please take your time reading it and ask any questions in the subscribers chat! Let’s get started. What I Wrote About This Week The post volume was lower this week, but I packed much more into each article. I would love some feedback on what readers prefer. Higher volume/single topic or lower volume/multiple topics? Let me know in the comments or sub dm please! Antrhopic S-1 provided a nice boost to the market and seemed to calm some nerves. At least for the short term. Marvell/Lumentum Deal / Hyperscaler ER & Capex / SpaceX. We got a very unexpected potential deal, some positive & some negative results from hyperscalers, and some interesting read throughs for Capex. No one is being rewarded for any kind of ER in the tech world right now. But there were some really great results worth talking about. There is lots of panic about Google Search growth. I unpack the concerns, what mangement is doing about it, what else they could implement, and what happens if they don’t succeed. With that, let’s jump into the OpenAI developments that might impact the market. Financial Disclaimer This article is for informational and educational purposes only. It reflects my personal research and opinions as of the date of publication and should not be considered individualized investment, financial, tax, or legal advice. Nothing in this article is a recommendation to buy, sell, or hold any security. Investing involves risk, including the potential loss of some or all of your investment. Stocks can be volatile, and past performance does not guarantee future results. Estimates, forecasts, valuations, and forward-looking statements are based on information available at the time and may prove inaccurate. Actual results can differ materially from expectations. I may own securities discussed in this article and may buy, sell, or change my positions at any time, including without updating this article. My interests may differ from yours, and my decisions may not be suitable for your financial situation or investment objectives. I use sources I believe to be reliable, but I cannot guarantee that all information is accurate, complete, or current. Company statements, financial data, and third-party estimates may change, and I have not independently verified every claim. Please do your own research and consider speaking with a qualified financial professional before making investment decisions. This article is not an offer to sell or a solicitation to buy any security. OpenAI One question behind all the spending concerns is whether the customers can fund the infrastructure they have agreed to use. OpenAI is one of the biggest. Suppliers are committing money to capacity today, expecting OpenAI’s spending to support years of business. After the close on Fri, Reuters reported that OpenAI had reduced its latest fundraising target from $30 billion to about $20 billion. The valuation being discussed fell from $1.4 trillion to $1.25 trillion, with the WSJ reporting roughly $11 billion of commitments so far. This private round is still being arranged. March’s financing brought $122 billion of commitments at an $852 billion valuation, so the proposed valuation is still considerably higher. If OpenAI raises less than planned, the question is how much its own business can cover and whether that changes its spending schedule. Reuters reported that the company was telling investors growth accelerated after GPT-6 Astra. September revenue was approaching a $50 billion annualized pace, compared with OpenAI’s reported pace above $20 billion exiting 2025. It was targeting $70 billion by year-end. For the actual first two quarters, the WSJ reported revenue of $5.7 billion and $6.7 billion, alongside deeper Q2 operating losses. We still need newer cost figures to judge how much of its expansion that growth can fund. Look at what happened on OpenRouter after Astra arrived. OpenAI passed 50% of combined OpenAI and Anthropic spending in the week of September 7, its first time above that level since February 2024. Astra alone took roughly a third the following week. That is a substantial change from OpenAI’s share of around a quarter earlier in the year, among customers using those two labs through OpenRouter. Weel’s October 22 study gives us a look at actual switching among 4,000-plus Australian and New Zealand businesses. In September, roughly three moved from Anthropic to OpenAI for every two going the other way. Plenty were keeping both with 34.8% of businesses paying either lab paid both. The more recent numbers are stronger for Anthropic though. Village’s October 15 analysis put its share of the two labs’ OpenRouter spending at 65–71% since October 7, following Opus 5.5. Ramp’s separate Token Spend Management sample had Anthropic at 60.5% and OpenAI at 35.1% for the week ended October 23. A month earlier, they were at 51.3% and 44.2%. Total spending in that Ramp sample grew 8.9% over the same period. Ramp’s October 14 chart, covering prices through October 11, shows businesses paying less per million tokens on average. Model choice and discounts affect that average. Customers can use more AI while paying less for each unit, so the revenue going back to the labs can grow more slowly than usage. Amazon’s AWS expansion covers $100 billion over eight years, on top of its existing $38 billion agreement. Amazon also funded a separate $50 billion investment in OpenAI by September 30. It benefits from more AWS spending and has money invested in the customer doing that spending. Microsoft recorded $10 billion of commercial revenue from OpenAI in the September quarter, including revenue sharing. That works out to about 11% of Microsoft’s total revenue. OpenAI receivables rose to $7.2 billion from $6 billion at June-end. How that balance converts into cash will be worth following as the relationship grows. NVIDIA has agreed to up to $105 billion of conditional guarantees for limited portions of PORTS lease and power obligations if specified OpenAI defaults occur. Those obligations activate in phases as leases begin and service conditions are met, with first service expected in NVIDIA fiscal 2029. The arrangement helps secure future demand for NVIDIA infrastructure while creating potential costs if OpenAI cannot meet those obligations. If financing delays push deployments back, cloud revenue could arrive later and equipment suppliers could miss shipment forecasts while carrying the cost of capacity already built. We may lower earnings estimates or pay a smaller multiple for them before that happens. A completed funding round and capacity arriving on schedule could ease those concerns. When does hyperscaler spending start to slow? I want to continually monitor this because a slowdown in spending will change where I want to be in the AI trade. Google or Microsoft could start keeping more of the cash their businesses generate. Their equipment suppliers would have less growth coming from ever-larger budgets. The timing of that shift is worth understanding if we own both. The latest calls still point to another big year of investment in 2027. Google expects $195–205 billion of capex in 2026 and a significant increase next year. Amazon expects about $220 billion of cash capex this year and a meaningful increase in 2027. Microsoft expects around $170 billion for calendar 2026, including more than $50 billion in the December quarter. Spending should accelerate in January through June 2027. Meta expects $135–145 billion this year, including finance-lease principal payments, and continued aggressive investment next year. Its numerical 2027 budget is still to come. The accounting differs. Microsoft includes new finance leases in its capex number, while Amazon reports cash spending. Microsoft also lowered its calendar-year estimate because some third-party capacity is arriving later. That spending moved into 2027. Management still says demand exceeds supply. Further out, the forecasts start to flatten. Goldman Sachs’s September estimates were $800 billion of hyperscaler capex in 2026, $1.2 trillion in 2027 and $1.4 trillion in 2028. Growth drops from 50% to about 17%, even though spending rises another $200 billion. A supplier can have a huge market to sell into and still face much slower growth. The Vanguard chart below uses Bloomberg estimates from July 31, before these earnings calls. It shows $1.08 trillion in 2027, $1.19 trillion in 2028 and $1.15 trillion in 2029. That forecast has spending flattening after 2028. Meta gives us one reason to pay attention to 2028. It wants to maximize capacity through 2026–2027, then has more flexibility to adjust server purchases to how much people use its AI products. It could slow spending if adoption disappoints, or keep investing if demand justifies it. Management has left itself that choice. Oracle has put dates around its recovery. At its October 28 investor day, it said capex should peak in fiscal 2027–2028 and step down in fiscal 2029. It expects positive free cash flow during December 2027 through May 2028, the second half of fiscal 2028. It then targets at least $30 billion of free cash flow in fiscal 2029, which runs from June 2028 through May 2029. Oracle expects to fund its own growth from then on. That would reduce its need to keep raising capital, which is a useful change for shareholders. The plan also uses vendor financing, customers supplying hardware and customer prepayments to reduce what Oracle has to pay upfront. Equipment suppliers could still be selling into those projects even as Oracle’s own spending falls. The other companies are starting from different places. Microsoft already generated $18.6 billion of free cash flow in its September quarter and expects to stay positive for fiscal 2027, ending June 2027. Google had negative $9.1 billion in Q3 but was still positive $19.7 billion over the trailing year. A $5.2 billion fine payment and higher TPU inventory also weighed on that quarter. Google has not given a date for quarterly free cash flow to recover. Amazon had negative $42.4 billion of trailing-year free cash flow. Andy Jassy has said much of its 2026 AWS investment should start being monetized in 2027–2028. That is when more of the new capacity should bring in revenue, though he has not promised free-cash-flow breakeven in either year. Meta had negative $2.1 billion in Q3 and has not given a recovery date. For the hyperscaler stocks, the encouraging outcome would be revenue and cash generation catching up with the buildout. Capex can keep rising and free cash flow can still improve, provided operating cash flow increases by more dollars than capex does. The stocks could respond before cash flow fully recovers if each quarter makes that outcome look more likely. How much upside that creates depends on how much recovery investors already expect. If capex is cut because customers are buying less, the revenue outlook weakens too. Lower investment also helps cash flow sooner than earnings. The servers already bought keep being depreciated, and the data centers still cost money to run. We could see cash flow recover while those costs continue weighing on reported profit. For our AI infra naes, the risk is investors paying for growth that the next round of budgets cannot support. A supplier can beat last year’s revenue and still disappoint the market. Faster connections, more optics per system and winning customers can help individual businesses grow through a spending plateau. The stock reaction will depend on how those results compare with the growth the market already expects. What could the midterms bring? I want to think through the possible stock reactions before Tuesday. At the time of this writing Kalshi had Democrats at about 93% to control the House and 70% for the Senate. Read more