Apple prepares a new way to buy iPhones. Plus, making sense of Google's new AI models

Every weekday, the CNBC Investing Club with Jim Cramer releases the Homestretch — an actionable afternoon update, just in time for the last hour of trading on Wall Street. Stocks are rebounding on Tuesday, with the S & P 500 on pace to snap a three-session losing streak. The steady climb in oil prices and interest rates is not deterring the market rally, at least for today. The artificial intelligence buildout theme traded sharply higher Wednesday as the market prepared for upcoming earnings. We wrote last Monday that earnings season could be the catalyst that puts the AI trade back on track, and that thesis will face its first major test Wednesday evening when Alphabet reports results. As we noted then, if Alphabet, Amazon , Microsoft and Meta Platforms raise their capital expenditure outlooks for this year, while continuing to signal investment growth into 2027, it would reinforce investor confidence in the size and durability of the current AI investment cycle. Apple shares rose slightly after Bloomberg News reported the company is partnering with buy now, pay later firm Klarna to launch a new device leasing program, called Apple Upgrade. According to the report, the new service is expected to launch July 28 and support most iPhone, Mac, iPad, and Apple Watch models. Instead of buying a new device outright, the subscription program lets users upgrade early to a new model or keep the device at end of the leasing term. Apple Upgrade will initially be available in the U.S. and accessible both online and at Apple's brick-and-mortar stores. With Apple raising prices across several devices due to soaring memory costs, Apple Upgrade could help ease sticker shock by offering lower monthly payments than current financing programs. The ability to spread out payments may also entice users who have held onto an iPhone for four or five years — or even longer — to finally upgrade to a newer device. This is important because Apple's newer, more powerful devices support its Apple Intelligence offering, which is on track to receive a much-needed makeover this fall. The additional capabilities and horsepower of newer devices also supports a broader range of apps, creating more opportunities to monetize each device sale. Lastly, a multiyear hardware subscription program further locks users into Apple's ecosystem, reducing the risk that they switch to competing devices. Alphabet's Google announced a new trio of specialized Gemini models . The models — Gemini 3.5 Flash-Lite, Gemini 3.5 Flash Cyber, and Gemini 3.6 Flash — are designed for different use cases and couldn't have come at a better time. For starters, they should help alleviate investor concerns about the pace of innovation following last week's disappointing news that Gemini 3.5 Pro was running behind schedule. The flagship model was originally due out in June . Alongside Tuesday's news, Google said that "Gemini 3.5 Pro is currently testing with partners and we plan to make it broadly available as soon as it's ready. In parallel, our team is already focusing on building the next generation of models. We have started our most ambitious pre-training run yet, for Gemini 4, and are excited by the progress." The arrival of these three specialized models coincides with a broader shift in the corporate world, as companies rethink their AI implementation strategies. After being hit with surprisingly large AI compute bills, companies have started to transition from "tokenmaxxing" — in which users threw everything they had at the most advanced models they could get their hands on — to a period in which the focus is on token efficiency. A token is the basic unit of data processed by an AI model; the more being consumed, the bigger the bill. One way to address this shift in thinking is to provide customers with specialized models for critical tasks — such as drug discovery or cybersecurity — and more efficient models targeted for high-volume, lower-complexity queries. Google certainly seems to be taking steps in this direction with Tuesday's updates, most notably with the introduction of Gemini 3.5 Flash Cyber. This evolving approach to model development is not just happening at Google. It's happening across the tech industry. Rather than design one model to rule them all, LLM developers are starting to think about models as the tools they are. Anyone who has ever assembled Ikea furniture, repaired an automobile, or worked on their home, knows one truth above all: having the right tool for the job changes everything in terms of how well and how efficiently the job gets done. The era of tokenmaxxing seems to have woken AI engineers up to this universal rule. As a result, we expect the efficiency of AI usage to increase, and with it, overall adoption. That makes Tuesday's news a positive for Google's AI strategy, and we expect to see more specialized models over time. This may be one of the topics discussed on Alphabet's earnings call Wednesday night. Headlining Tuesday night's earnings reports is Club name Capital One . We want to see the bank get back in the habit of hefty earnings-per-share beats through a more disciplined approach to its card and network related investments. Also reporting Tuesday night are Interactive Brokers , EQT Corporation , Chubb and Alaska Air Group . Fellow Club name GE Vernova reports before the opening bell on Wednesday. We want to hear the company about strong orders and pricing for its gas turbines, which are key to the AI buildout, with accelerating opportunities in its Electrification segment . Other companies scheduled to report on Wednesday include AT & T , PulteGroup , and Otis Worldwide . The only economic data release scheduled for Wednesday is weekly mortgage applications. (See here for a full list of the stocks in Jim Cramer's Charitable Trust.) As a subscriber to the CNBC Investing Club with Jim Cramer, you will receive a trade alert before Jim makes a trade. Jim waits 45 minutes after sending a trade alert before buying or selling a stock in his charitable trust's portfolio. If Jim has talked about a stock on CNBC TV, he waits 72 hours after issuing the trade alert before executing the trade. THE ABOVE INVESTING CLUB INFORMATION IS SUBJECT TO OUR TERMS AND CONDITIONS AND PRIVACY POLICY , TOGETHER WITH OUR DISCLAIMER . NO FIDUCIARY OBLIGATION OR DUTY EXISTS, OR IS CREATED, BY VIRTUE OF YOUR RECEIPT OF ANY INFORMATION PROVIDED IN CONNECTION WITH THE INVESTING CLUB. NO SPECIFIC OUTCOME OR PROFIT IS GUARANTEED.