Constellation Energy Buys Rhode Island Gas Plant From Shell for $715 Million

Pinterest is opening selected display and video inventory to programmatic buyers in North America, South America and Asia Pacific. The October 26 launch makes Programmatic Deals generally available in those markets and gives enterprise advertisers a way to buy Pinterest placements through systems they already use for broader campaigns.

Programmatic advertising uses software to plan, buy and measure ad placements across multiple publishers. For Pinterest, the attraction is access to demand that might not flow through its own advertising interface. The company says the new channel complements rather than replaces its first-party ads business.

Pinterest is connecting inventory to supply-side and demand-side partners. Google Ad Manager supports supply in Latin America and Asia Pacific. Buyers can access inventory through Amazon DSP, Google's Display & Video 360 and Yahoo DSP. Pinterest also describes itself as the primary supply-side platform for the program, which should help it retain control over inventory and audience signals while broadening distribution.

The commercial argument rests on user intent. Pinterest says it reaches 640 million monthly active users and handles more than 80 billion searches each month. More than half of those searches are commercial, and 96% of text searches are unbranded. In practical terms, many users are deciding what to buy before choosing a brand, which can make the platform valuable for both awareness and performance campaigns.

Pinterest cited a Moroccanoil campaign run through Magnite and Amazon Ads as an early example. The company said the campaign produced 15% incremental reach with 1% audience overlap, 63% more new-to-brand buyers among users whose purchase paths included Pinterest, and a click-through rate 20% above the brand's average video and display campaigns.

Those figures are encouraging, but they are company-reported results from one campaign. They do not establish the average performance of Programmatic Deals, the fees Pinterest will retain or whether programmatic demand will be incremental rather than shifting spending from direct purchases.

The investment case therefore depends on two separate outcomes. Pinterest must make its inventory easier to buy without weakening pricing, measurement or advertiser relationships. It must also convert the platform's commercial search activity into higher revenue per user, especially outside the United States and Canada where monetization has historically been lower.

General availability removes friction for large advertisers, which is a credible positive step. The next evidence should come from management's disclosure of programmatic demand, advertiser growth and any effect on average revenue per user. Until then, Programmatic Deals expands Pinterest's sales channel, but does not by itself prove a material earnings contribution.

Axon's new Body 5 camera is more than a routine hardware refresh. It is designed to move the company's AI software closer to officers in the field, turning the launch into a test of whether Axon can convert a large installed camera base into faster software adoption.

Axon introduced Body 5 and Body 5 Pro on October 24. The company says both versions have 25 times the AI processing power of earlier Axon cameras, better low-light video, improved connectivity and 20% longer battery life than Body 4. The Pro model adds a close-focus lens that can scan a driver's license, four front-facing speakers for clearer real-time translation and more memory to run several AI tools at once.

The camera is also built for 5G and Wi-Fi 7. Axon says the Pro model can switch Wi-Fi connections without interrupting a livestream, adapt video quality when coverage weakens and report an officer's location more precisely. Compatibility with existing Body 4 docks, mounts and cables could lower the operational cost of an upgrade for agencies.

The most consequential feature is Guardian, a new group of AI safety tools in Axon's AI Era Plan. Escalation Detection can analyze audio from an active Body 4 or Body 5 recording and create an alert in Axon Fusus. A person then reviews the livestream and decides whether support is needed. Approach Alert can warn an officer when the door of a stopped vehicle opens. Axon emphasizes human review and agency control over data, which is important because a false alert or missed event could affect both officer and civilian safety.

Axon has not announced pricing or general availability for Body 5 or Guardian. That omission prevents investors from estimating the immediate revenue contribution. It also means the launch is not yet evidence that agencies will replace cameras on a particular schedule.

The financial context is stronger than the product announcement alone. Axon reported second-quarter revenue of $904 million, up 35% from a year earlier. Software and Services revenue grew 36% to $398 million, annual recurring revenue increased 39% to $1.6 billion, and AI Era revenue grew nearly 700% from a small base. Those figures show rapid adoption, but they do not reveal how much Body 5 will accelerate software revenue or what incentives Axon may use to encourage upgrades.

Hardware can pull software forward if agencies use the added processing capacity to buy the AI Era Plan and related tools. It can also pressure margins if higher memory and component costs arrive before software revenue. The first evidence will be a ship date, pricing and agency commitments, followed by whether Software and Services keeps growing faster than connected devices.

Body 5 strengthens Axon's integrated platform, but the launch should be judged by paid usage rather than processing power. The camera matters most if it expands recurring software revenue while keeping human oversight credible in high-stakes public-safety work.

AT&T and T-Mobile are using more artificial intelligence to manage their networks, but the same technology is forcing them to spend more on cybersecurity and keep people in control of critical decisions. Executives from both carriers told Reuters that fully autonomous telecom networks remain years away.

The risk is not simply that an AI model makes a bad recommendation. AI agents can act across several software systems with limited human input, which can expand the damage from a stolen credential, weak vendor control or poorly designed permission. A mistake could expose customer data or interrupt network operations rather than remain inside a test environment.

T-Mobile Chief Technology Officer John Saw said the carrier is increasing monitoring and threat response and working with vendors to close vulnerabilities faster. That caution follows major breaches in 2021 and 2023 that affected more than 100 million customer accounts in total. AT&T Chief Technology Officer Jeremy Legg pointed to risks outside a carrier's direct control, including third-party software and uneven security standards in open-source projects.

AT&T says its controls include encryption, multi-factor authentication, human review of data access and checks to verify that an outside AI agent is acting for a real customer. Those measures matter because an agent that can retrieve account information or change a service setting needs stronger identity and authorization checks than a chatbot that only answers questions.

Both companies already use AI in live network operations. AT&T applies it to review trouble tickets and sometimes carry out repairs. T-Mobile uses a self-organizing network to adjust cell-site antennas when another site fails. During Winter Storm Fern in January, T-Mobile said that system restored coverage to about 66% of affected customers within an hour and 90% within eight hours, often before electric power returned.

That result shows the business case for automation: a carrier can restore service faster and reduce the amount of routine work handled manually. It also explains why management is unwilling to remove people from the loop. Unusual failures, conflicting data and third-party dependencies are precisely the situations in which an automated action can spread a problem.

For investors, the development is primarily a cost and execution issue rather than a new revenue catalyst. AT&T spent $5.6 billion on capital investment in the third quarter, while it also works to extend fiber and wireless capacity. Security spending must compete with those priorities, yet underinvestment could create much larger costs through outages, customer losses, regulatory penalties and remediation.

The useful measure is not how often management uses the word AI. It is whether automation improves service reliability without increasing breach frequency or weakening accountability. AT&T and T-Mobile are moving toward more automated networks, but their expanding guardrails show that the safer path is controlled delegation, not unattended operation.

Boeing and Coca-Cola report third-quarter results before the market opens on October 27, but investors are asking almost opposite questions. Boeing must prove that a production recovery can generate dependable cash. Coca-Cola must show that strong brands can keep lifting volume and profit after a 35.7% year-to-date share gain.

Boeing's second quarter offered real progress without completing the turnaround. Revenue rose 8% to $24.6 billion as the company delivered 171 commercial aircraft, according to Boeing. Free cash flow, a non-GAAP measure, improved to $631 million from a $200 million outflow a year earlier. Yet Boeing still posted a $0.76 core loss per share, and a $280 million charge on the VC-25B presidential aircraft program showed how defense execution can absorb gains from higher jet deliveries.

The third-quarter cash result therefore matters more than a single earnings-per-share comparison. Boeing has said it expects free cash flow in the low hundreds of millions of dollars despite a roughly $700 million Justice Department payment. A positive result would suggest that more deliveries and customer advances are beginning to outweigh penalties, development spending and factory investment. A miss would raise doubts about the pace of recovery even if revenue grows.

The 737 MAX 10 adds a second test. The aircraft received certification on October 20 after a long delay, and the source report says airlines have ordered more than 1,400. Investors now need a credible timetable for first deliveries in 2027 and evidence that Boeing can increase 737 production toward 47 aircraft a month without weakening quality or disrupting suppliers. Certification removes one regulatory obstacle, but it does not by itself create cash flow.

Coca-Cola enters the morning from a stronger operating position. In the second quarter, net revenue rose 7% to $13.4 billion, global unit case volume increased 5%, and comparable earnings per share rose 11% to $0.97. The company also raised its full-year comparable EPS growth outlook to 9% to 10%. Coca-Cola Zero Sugar grew volume by 16%, while the World Cup campaign supported demand across several brands.

The harder comparison now begins. Management has warned that the second half faces tougher year-over-year volume comparisons and that the fourth quarter has six fewer selling days. Asia Pacific price and mix fell 9% in the second quarter, so investors should separate broad volume strength from regional pricing pressure. The useful benchmark is Coca-Cola's full-year organic revenue growth target of about 5%, not just whether reported earnings exceed the consensus estimate.

Both companies can beat expectations for very different reasons. Boeing needs production and certification milestones to become cash. Coca-Cola needs resilient volume and pricing to support a premium valuation. The October 27 reports will show whether each company is converting its strongest recent headline into the result that shareholders actually need.

Amazon's AI investment is now large enough to push companywide free cash flow below zero, even as AWS grows at its fastest rate in years. That combination makes the spending debate more precise: the question is not whether demand exists, but whether Amazon can turn contracted demand into cash before the cost of capacity keeps rising.

Amazon's official second-quarter release shows trailing 12-month operating cash flow of $161.4 billion, up 33% from $121.1 billion. Free cash flow, however, fell to an outflow of $7.6 billion from an inflow of $18.2 billion a year earlier. The company attributed the change primarily to a $66.1 billion year-over-year increase in purchases of property and equipment, net of proceeds and incentives, largely for artificial intelligence.

Those figures correct the source article's $11.6 billion free-cash-flow deficit. The official definition and period are the safer basis for comparison. They imply that net property and equipment purchases exceeded operating cash flow by about $7.6 billion over the trailing year, before considering other financing choices.

The business funding the build remains strong. AWS sales rose 37% to $42.2 billion in Q2, while AWS operating income increased to $16.6 billion from $10.2 billion. Companywide operating income reached $27.5 billion. Amazon also reported $62.6 billion of net income, but that figure included $53.4 billion of pre-tax nonoperating income, primarily from its Anthropic investment. Operating performance is therefore a better gauge of the recurring business than the headline net-income jump.

Demand visibility has improved alongside spending. Reuters reported that AWS backlog reached $496 billion at quarter end, up from $364 billion three months earlier. Amazon also said its AWS AI business and chips business had each exceeded a $25 billion annual revenue run rate. The scale of the backlog supports management's decision to build, but backlog is not current revenue. Contract timing, cancellations, customer ramp schedules and available power can all delay conversion.

The cash-flow risk lies in synchronization. Servers, networking gear and land can be paid for before a data center has enough power or customers are ready to use it. That creates a gap between capital spending and revenue. If AWS capacity fills quickly, today's outflow can support years of higher cloud sales. If deployment slows, Amazon carries depreciation, financing and operating costs while assets wait to produce revenue.

The next earnings report should answer three questions. First, did trailing free cash flow improve from the $7.6 billion outflow? Second, did AWS backlog convert into revenue without sacrificing the 39% Q2 segment operating margin? Third, did management change the pace or composition of capital spending as memory, power and construction costs moved?

Amazon has recovered from deeper cash deficits before, but the current build is larger and more infrastructure-intensive. The investment case rests on the speed of backlog conversion and the returns on each new unit of capacity, not on the size of spending alone.