Welcome back to Hitting the Bid Weekly!

Quick one before we get into it — I'm trying to make this newsletter more useful. 10 seconds:

Equities finished higher after another volatile week shaped by AI earnings, labor data, and shifting expectations around Iran.

Tuesday delivered the week’s biggest move. Strong results from Palantir and Caterpillar renewed confidence in AI-related spending, while falling oil prices and Treasury yields added support. The S&P 500 (SPY) gained 1.8%, the Nasdaq (QQQ) jumped 3.4%, and the Russell 2000 (IWM) rose 1.9%.

Momentum cooled Wednesday and Thursday as investors worked through mixed results from AMD, SpaceX, and several software companies. The broader market held up relatively well, but the reactions reinforced that strong AI demand is not always enough when spending and valuations remain elevated.

Friday’s jobs report shifted attention back to the Fed. The economy unexpectedly lost 23,000 jobs in July versus expectations for an 80,000 increase, while May and June were revised down by a combined 103,000. Equities rallied as traders reduced their expectations for a September rate hike.

Monday interrupted that progress. Crude oil rose 5% after Iran attached new conditions to reopening the Strait of Hormuz. Across the full Tuesday-through-Monday window, SPY gained about 2.0%, QQQ rose 3.0%, and IWM added 1.3%.

This week brings CPI, PPI, retail sales, and earnings from Cisco and Applied Materials.

Key market moves this past week:

Not an exhaustive list — just a few I’m watching closely for potential market impact.

A friend gave me the idea to take a look at Constellation Energy $CEG ( ▲ 2.93% ). I’ve been watching it for a couple of weeks, mostly because the chart was chopping around with nothing too compelling for me.

Last Thursday finally gave me something to work with, at least fundamentally. Constellation reported Q2 adjusted EPS of $2.55 against consensus estimates of roughly $2.34 and raised full-year guidance to $11.50 to $12.50 per share, up from its prior range of $11.00 to $12.00. Shares traded as high as $280 after the report before reversing lower and closing near $261.

The quarter had real substance behind it. Constellation signed an additional 920 megawatts of long-term nuclear power agreements, while results benefited from the Calpine acquisition, higher capacity revenue, and stronger commercial margins. Management also cleared two regulatory hurdles tied to restarting the Crane Clean Energy Center, which remains on track for a second-half 2027 return. The planned sale of the Brazos Valley Energy Center is expected to satisfy the final DOJ condition associated with the Calpine deal.

This isn’t a story about AI enthusiasm getting ahead of itself. It’s a power generation company converting demand for reliable electricity into actual earnings growth.

Here’s what keeps me from calling this settled, though. CEG is still down roughly 23% year to date and remains more than 30% below its October 2025 intraday all-time high of $412.70. One good quarter doesn’t erase a nine-month downtrend. The failure to hold Thursday’s early gains matters too, even if the underlying results were substantive.

What I’m watching is whether CEG can reclaim the $280 to $284 area and hold it. That zone has acted as resistance several times, so a close above it with follow-through would be the first real technical evidence that the downtrend is ending, not just being interrupted.

A break below the recent lows in the low $260s would make Thursday look more like a one-day reaction that faded back into the same range that has frustrated this stock for most of the year.

On July 30, a group gathered in Chicago for an event called Claude Conversation. The question on the table was blunt: What is AI doing to me, my kids, and my community?

I wasn’t there for that event. But I was there a week later when a different group of people showed up for the Claude Impact Lab to make parts of that question more concrete.

The format was simple. Pick one lane out of three: teach people AI, help them navigate the city, or help them get well. You get one lane, the whole day, and no changing your mind once you’ve picked.

Before lunch, you also had to name a real recipient. Not “the community.” A specific Chicagoan or organization that could realistically use the thing you were about to spend the next six hours building.

My team picked navigating the city. Chicago already has a way to report a pothole or a downed tree: a city portal with fields and dropdowns that assumes you have ten minutes and the patience to fill it out correctly. We built a version for someone who has neither. Take a photo, give a location, type one sentence. Submit.

It wasn’t clever. It was just smaller than the problem we started with.

That was the pattern across the room, not just at my table.

One team built a food-access platform that layered real food-insecurity data over real Chicago pantry locations. The example they used wasn’t “food deserts” in the abstract. It was whether someone in Hyde Park could find a halal grocery option.

Another team built a career assessment that used twenty questions to match people with sixteen skilled trades using government wage data. The results connected users with apprenticeship programs they could apply to, including actual deadlines. That detail matters because many programs accept applications only during specific windows.

None of us set out to solve “AI and the economy.” We solved for one pothole, one neighborhood, or one person deciding whether a skilled trade could be right for them.

The question got smaller before it got useful. AI helped us build working prototypes in a day. The constraints helped us identify questions worth answering.

I noticed that I don’t narrow questions by default. Left alone with an open-ended problem, I tend to circle it.

Is the newsletter working? Is HOA Basecamp the right thing to be spending Tuesday nights on?

Those questions feel urgent, but neither gives me much to do next. They are weather, not decisions.

The room on Friday didn’t give us that luxury. A deadline, a lane, and a real recipient did the narrowing for us.

The original question stayed with me too. My son is almost eleven months old, so “What is AI doing to my kid?” no longer feels entirely hypothetical. It’s the kind of question I can easily circle at odd hours without getting anywhere. I still don’t have an answer. But the hackathon offered a more useful place to start: What is one specific way I can help my family navigate this right now?

That doesn’t answer the big question. It gives me a smaller question I can actually act on.

I don’t have a hackathon’s worth of structure most weeks. But I can borrow its shape whenever I catch myself circling.

Pick the lane before picking the solution.

Name the person this is actually for.

Set a deadline that doesn’t allow for more circling.

That’s the whole trick, as far as I can tell.

Not more certainty. Just a smaller, sharper question than the one you started with.

Like what you’re reading?
Forward this to a friend or share the newsletter link to help grow the HtB community.

Hitting the Bid content is for informational and entertainment purposes only. The information contained is not, nor is it intended to be, trading or investment advice or a recommendation of any security, futures contract, digital asset or alike. I may hold a position in the trading vehicles discussed. Trading and investing contains risk. All investors should evaluate their own risk tolerance, financial situation, and investment duration before entering any trade or investment.