The following comp sheets offer PEG ratios (P/E divided by the next 2 years of compounded earnings growth) for various peer groups. Please note that “2026 Consensus Estimate Trend” calculates year-to-date consensus EPS estimate changes from sell-side analysts. It is not me expressing an opinion. It is me relaying data. Next, PEG ratios aren’t perfect. They’re just one valuation lens out of many and one piece of the puzzle. I use PEG ratios because I love that they contextualize P/Es by rewarding stronger profit growth. Faster, more durable bottom-line compounding is the thing that (all else equal) warrants a premium and this methodology acknowledges that better than others do, in my opinion. Finally, companies without earnings growth were excluded from averages to avoid issues.

This week’s Robinhood product event was AI-themed and geared towards the active trader. Starting with Robinhood Agents, the company debuted autonomous tools for research, investment strategy creation and execution. The launch includes strict guardrails for setting actions and builds on the 3rd-party AI agent connectors they built in May. These connectors are now used by 150K+ customers, showing Robinhood how popular the new tools already are and motivating them to build 1st-party alternatives. While Robinhood obviously doesn’t have the model building skills of an OpenAI or a Meta, that doesn’t matter. They can easily access these models and use their ocean of lucrative retail investor data to customize the highly capable tools, making them more accurate and relevant. Models including GPT-Luna are available for agent work and that specific product is actually free for customers to use through the end of 2026.

Customers are able to manually approve each action an agent takes or allow that agent to act in a purely autonomous fashion. While agent autonomy is a bit nerve-wracking when they have access to our fortunes, it could be highly valuable for an upcoming launch called “Loops.” This provides perpetual agent work on a customer’s behalf to automate tedious work like selling covered calls for income. While I’m sure there will be early takers, I question whether or not people are willing to completely check out of their trading strategies and allow sometimes unpredictable agents to transact with their accounts. I personally see this as a hard sell and think the more manual versions of this product will be the ones that become highly popular. 

To augment the information and context that Robinhood Agents have at their disposal, the company is adding 11 3rd-party data connectors from financial service companies. These include unusual options activity subscriptions from Unusual Whales and one through a Nasdaq partnership as well. The idea here is to add layers of unique and valuable data that Robinhood doesn’t already have to allow Robinhood’s platform to deliver more informed and capable agents. They’ll take a cut of these paid subscriptions as they come out. Big fan of this.

More Access & New Investment Types:

Pending regulatory approval, Robinhood will move beyond 24/5 equity trading to 24/7 trading across a selection of U.S. stocks and ETFs. While I’m not the biggest fan of this for U.S. retail investor health, my opinion doesn’t really matter and this is inevitable. Hopefully people make responsible decisions as they trade SPY at 2:30 AM on a Saturday. Along similar lines, Robinhood plans to extend options trading from 9:30-4:00 to 7:30-4:15. That means contracts will be open for trading after most earnings data is released (aside from companies like Apple and Nvidia that release data around 4:30).

In terms of new products, perhaps the most interesting announcement was Crypto Perpetual Futures (launching in the coming months). These are derivative contracts without any expiration date, allowing people to speculate on cryptocurrency price changes. It will offer up to 10x leverage. Again, not a product I plan on ever using (from any vendor) but I’m sure there’s a large retail cohort that wants it. And Robinhood is happy to give the people what they want. Robinhood sees this as the first “true perpetuals” product in the USA and plans on doing much more in this area across more asset classes over time. Other product announcements include earnings contracts that let people bet on things like revenue beats/misses as well as up to 4x intraday margin.

More Notes:

Quick Take:

A few things here. This company innovates at an impressive clip and tends to blaze the trail in terms of new asset classes and product iterations for its customers. Their interface is beautiful, their international expansion is going extremely well early on and they’ve built a large base of loyal customers. The top-of-funnel growth softness has been addressed and their financials look quite pretty. 

The only thing that keeps me away from owning this is the revenue mix. Retail crypto, options and margin volumes are instrumental parts of their growth engine and these product launches show they’re leaning into the strength wherever they can. The product diversification is coming from things like banking, which is great. But? It’s also coming from sports gambling (I’m sorry but sports prediction markets are the same thing as gambling), which I do not think should be offered in the same place that retirement accounts are. This feeds into the same theme of speculation and hyper-aggressive risk taking.

I actually think all of this can be a good business model for the company (and has been), but I do not think it’s a good business model for their customers. That’s why I choose to invest elsewhere, despite admiring the pace at which HOOD moves and their performance as a public company. Just me. For those who disagree, I totally understand.

NOW launched a new product called "Flow." This is the culmination of its Moveworks purchase. It provides a native, conversational one-click & actionable service desk right from within a Slack or Teams environment. There's no need to access it through a dedicated NOW portal or interface, and it can be easily onboarded in a single day without any added hardware or much of a learning curve. Flow starts at a $10K annual minimum, sold as in-product credits — while existing ServiceNow customers simply consume the Assist credits they've already purchased. Email and a Flow web app are in the mix too — Slack and Teams aren't the only doors in. This will be a more flexible and open front-end means of accessing its AI IT Service Management platform in a highly convenient, headless (access through chat channels like Slack and Teams rather than logging into ServiceNow) manner. It's built and backed by ServiceNow but runs on its own standalone architecture as a purpose-built service desk for teams that don't want the full platform commitment. We knew this was coming. No surprises here but good to see the roadmap advancing on schedule.

Meta’s long-rumored and telegraphed enterprise launch came this week with the new “Meta Enterprise Platform.” The launch packages all the momentum Meta is enjoying across models and agents. It includes Muse and the Meta Business Agent alongside the Muse API and Muse Code. The Muse API lets developers call Meta's Muse models directly, including through OpenAI and Anthropic interfaces, while Muse Code is its agentic coding solution that has been beta testing for two months. While there will be doubters and skeptics as there always are for Meta, I’m highly optimistic that this performs well. Before this new platform announcement, Meta had already rapidly grown business agents to 1M+ weekly active merchants before an Instagram debut. The company is already helping merchants like Movida boost WhatsApp bookings by 44% with these tools. Generally speaking, Business Messaging is a big reason why their “other revenue” segment crossed $1B/quarter last period. Finally, Meta represents 3B+ engaged users. What do merchants want? Access to those users. And how can they directly get it with tools purpose-built to help them most effectively monetize these users? With Meta Enterprise Platform.

Meta is playing with house money here. If this flops (which it won’t), they cut their losses, reallocate the compute to use cases that are already working and come out just fine. If it works? This provides a sharp increase to their total addressable market and a significant elongation in their growth runway. Great bet to make.

One of the most interesting parts of this news is the person who was named to run things. It was former MongoDB CEO CJ Desai who bolted from that company four days before their investor day and less than one year after taking the helm. I realize he probably got a large raise and a special opportunity, but still not an amazing look any way you slice it. He’s a world-class go-to-market talent in enterprise software and is an absolutely perfect fit for this role. He was ServiceNow’s President and COO for a few years and Cloudflare’s President of Product and Engineering before his short stint at MDB.

OpenAI’s Muse competitor announced this week (Dots) is only available for premium tiers. This highlights why it’s so great to be Meta. They don’t need to directly make money on providing Muse to consumers. OpenAI is a lot more reliant on that because they don’t have a fortress apps business that prints mountains of profits all year long. Meta can keep direct access free… use that to build massive scale… then layer on transaction-based revenue and inevitably paid add-ons down the road.

Snowflake announced & closed a $3.75B convertible note offering. The first tranche ($2B) matures October 2029 with an initial conversion price of ~$500. The second tranche ($1.75B) matures October 2031 with an initial conversion price of ~$484. Both carry a 0% interest rate. This represents ~3% dilution if all notes are converted into equity.

Amazon is hiking GPU prices by 15% next week. They hiked prices by 20% in July. Another good sign of ongoing pricing power and cycle strength.

Netflix Co-CEO Ted Sarandos said “we’re not growing as fast as I want us to." He blamed prioritizing live content spending, which leads to fewer watch hours but over-indexes in terms of driving new viewers. That really shouldn't be impacting top-line figures that much. They do take the initial watch hour hit, but the incremental sign-ups they get from this should materially offset that hit. I continue to think this is a byproduct of their competitive environment being daunting. Amazon, Google and Apple have all decided to prioritize their libraries. They don't need to make a lot of money on these libraries. And they have more to spend. That is hard to sustainably beat. I worry this company's best days are behind it.

There were inaccurate rumors about Nu buying Monzo to expand into Europe.

Alphabet’s Gemini 4 Argon is out and looking good across benchmarks. There's going to be a ton of leapfrogging. But people were again starting to think Alphabet couldn't stay at the head of the pack following recent delays. This should help quiet that emerging negative sentiment. Felt like a safe bet to make with Brin leading the charge. 

Jobs data was pretty underwhelming this week. JOLTS Job Openings for August missed, while Nonfarm Payroll data, Private Nonfarm Payroll data and the unemployment rate (4.2% vs. 4.1% expected) were all worse than expected. ADP Nonfarm employment change was strong, but that was the lone highlight. This led to some relief in bond yields and the probability of another 2026 rate hike fell due to this and the inflation data we got. The Core PCE for August rose 0.2% M/M vs. 0.3% expected while the PCE rose 0.3% M/M vs. 0.4% expected. This was good news despite a methodology change that helped the metrics a bit this month.

For output data, the Chicago Purchasing Managers Index (PMI) was 58.8 vs. 51.2 expected and the latest Q2 GDP reading was 2.2% vs. 1.5% expected. Finally, the Institute for Supply Management (ISM) PMI was 54.5 vs. 54.8 expected and 54.6 last month.