BUSINESS

Felix Rosenqvist’s eye for detail and precision helped him win the 110th Indianapolis 500. He discovered those qualities in his “Forever Face” for the Borg-Warner Trophy.

The Green Bay Packers are hoping castoffs Laken Tomlinson and Mekhi Becton can help fix their brutal offensive line play.

A few years ago, quantum computing was the kind of technology that scientists talked about at conferences, and investors mostly ignored. The machines were too error-prone, the use cases too theoretical, and the timelines too fuzzy to build a serious investment thesis around.

That’s changing, and Bank of America just put a number on how quickly.

Sept. 28, analyst Vivek Arya initiated coverage on IonQ (IONQ) with a Buy rating and a $60 price target, according to a note shared with TheStreet. 

Arya ranks 147th out of more than 12,500 Wall Street analysts on TipRanks, with a 59% success rate.

The stock was trading around $45 at initiation, meaning Arya sees 33% upside from current levels.

Arya argues that IonQ’s recent acquisitions could accelerate the timeline for the technology to become genuinely useful.

Also Read: IonQ Inc. Latest News and Stories

Why two IonQ acquisitions might matter more than any earnings beat

IonQ made two deals this year that, individually, might look like routine corporate development. Together, according to Bank of America, they could meaningfully compress the timeline to commercial-scale quantum computing.

In January, IonQ acquired Seed Innovations, an artificial intelligence (AI) and software development company that will help the firm manage and scale complex quantum workloads.

In July, it closed the acquisition of SkyWater Technology, a semiconductor foundry that gives IonQ direct control over chip development and manufacturing to support its next-generation Superion systems.

More AI:

Nvidia just made a move Wall Street wasn’t ready for

Microsoft just took sides in AI policy fight

OpenAI just disclosed something genuinely alarming

That second deal is the more structurally significant one. Quantum computing has a hardware problem: the qubits that perform calculations are error-prone, and improving their fidelity while scaling qubit count requires extremely precise chip manufacturing. 

By bringing that manufacturing in-house, IonQ eliminates a critical dependency that could otherwise slow its roadmap.

Arya described it as a “semiconductor-enabled path to scaling qubit count” that supports IonQ’s development of Superion 256, its next-generation platform targeting launch in Q2 2027. 

The Superion 256 will be the first to integrate electronic qubit control, which Arya called “a significant architecture shift that could support substantial logical-qubit scaling,” according to the note.

Logical qubits — error-corrected qubits that can perform reliable computation — are the bridge between today’s error-prone machines and systems useful enough for real commercial problems in chemistry, materials science, optimization, and cryptography.

The IonQ’s revenue trajectory that makes BofA’s 2030 thesis legible

I’ll be honest: quantum computing revenue forecasts extending to 2030 require a significant degree of trust in trajectory rather than current performance. That’s a reasonable objection to any initiation in this space.

But IonQ’s recent numbers are harder to dismiss than most early-stage technology stories. 

Q2 2026 revenue reached $80.05 million, up 287% year-over-year (YoY), beating analyst consensus, according to IonQ‘s statement. IonQ has now delivered five consecutive quarters of record results.

IonQ announced that following the SkyWater acquisition, IonQ raised its full-year 2026 revenue guidance to a midpoint of $455 million, up from a prior midpoint of $285 million.

Related: Nvidia once rattled IonQ stock. Now it plans to install IonQ tech

That guidance lift reflects both organic momentum and the foundry revenue stream IonQ is now building. Remaining performance obligations grew 297% YoY. That’s the clearest signal that customer commitments are expanding beyond one-time contracts into longer-duration relationships.

Bank of America’s valuation framework applies a 7.0x enterprise value-to-sales multiple to $2.5 billion in estimated fiscal 2030 quantum hardware and services revenue and a 2.0x multiple to $1.2 billion in estimated foundry revenue, according to the note. 

That blended 5.4 times fiscal 2030 EV/S multiple implies a 69% compound annual growth rate in sales from fiscal 2026 to fiscal 2030.

My read is that growth rate requires everything to go right — engineering milestones hit, commercial adoption accelerating, and no significant competitive disruption from IBM, Google, or well-funded startups. 

Arya acknowledged as much, noting that delays in machine delivery or commercial adoption could put the forecast at risk.

IBM uses superconducting qubits. IonQ uses trapped-ion technology.Shutterstock

What the IBM quantum computing comparison tells you about the timeline

I’ve covered IBM CEO Arvind Krishna’s quantum computing roadmap before. Krishna said in July that quantum computing would have “a measurable impact” on IBM’s top and bottom line by 2028 or 2029, with a potential $1 trillion in value created by the end of the 2030s.

IonQ is operating on a broadly similar timeline, but from a different architectural approach. IBM uses superconducting qubits. IonQ uses trapped-ion technology, which generally offers higher fidelity but faces different scaling challenges. 

The semiconductor foundry acquisition addresses one of trapped-ion’s historical limitations — the ability to manufacture ion trap chips at scale with the precision required for higher qubit counts. IonQ CEO Niccolo de Masi has been unambiguous about the ambition. 

Second quarter revenue of $80.1 million again exceeded our guidance, reflecting continued customer demand across our expanding quantum platform.

IONQ shares are up roughly 20% over the past month but have gained only 5% year-to-date, compared to the S&P 500’s 13% gain, according to Yahoo Finance.

It has underperformed the broader market in 2026, which is precisely why BofA sees entry value at current levels relative to the 2030 target.

Related: IBM CEO sends blunt message on quantum computing

McDonald’s (MCD) has spent years experimenting with automation.

Its newest plan makes the financial objective much clearer.

As part of its new McDonald’s > NEXT growth strategy, the fast-food giant is rolling out an artificial-intelligence-powered restaurant operating system, ArchIQ, that it says can eliminate roughly 50 labor hours per restaurant every week from order-taking alone, CNBC reported.

ArchIQ has an AI ordering assistant, “Archy,” that can take customer orders in English and Spanish. The system is also designed to manage inventory, schedule employee shifts, and use scales to check order accuracy.

McDonald’s is not saying it has plans to cut a specific number of jobs. But executives are promising fewer labor hours inside restaurants and more corporate efficiency from AI at a time when McDonald’s is trying to push operating margins substantially higher.

The automation strategy also comes as restaurant operators struggle with persistent inflation, rising labor costs, and softer customer traffic.

Those pressures are no longer expected to be temporary, said McDonald’s CEO Chris Kempczinski.

“We need to stop talking about that being a difficult environment and just say that is the environment,” Kempczinski told CNBC.

McDonald’s is giving AI more restaurant jobs to do

ArchIQ is central to McDonald’s restaurant modernization strategy.

According to McDonald’s, Archy can take orders in both English and Spanish, saving about 50 labor hours each week.

But taking orders is just one job.

ArchIQ can also manage inventory and employee scheduling, and other technology will use scales to make sure restaurant orders contain the correct items. McDonald’s will also implement AI-driven revenue-management tools and use Archy to suggest additional products to customers.

More Restaurants:

52-year-old international restaurant chain closing all locations

46-year-old casual dining chain closes underperforming locations

Classic burger chain has closed down all its restaurants

Those capabilities should help increase the average amount customers spend over time, CFO Ian Borden said.

The labor-hour figure is of special interest.

Fifty hours per week translates to roughly 2,600 hours per restaurant over a full year if the savings remain consistent.

The material from McDonald’s didn’t say that those hours would directly lead to layoffs. Restaurants could also hire fewer people, leave positions unfilled, move employees to other work, or cut their hours.

But the economic motivation is clear.

McDonald’s and its franchisees are struggling with higher labor costs amid pressure on restaurant traffic.

Its most recent U.S. same-store sales grew a mere 0.8%, and traffic at its restaurants in the U.S. declined.

Kempczinski said inflation remains “sticky” around the world, with restaurants facing higher costs for labor, construction, and ingredients. Beef costs in McDonald’s largest markets have nearly doubled during the past five years, according to the CEO.

That makes labor-saving technology increasingly valuable.

If McDonald’s can automate repetitive jobs such as taking orders, checking bags, managing stock, and building schedules, franchisees potentially need fewer employee hours to generate the same revenue.

The strategy also goes beyond individual restaurants.

McDonald’s wants AI to lower costs at headquarters.

McDonald’s is quietly taking work away from employees.Chicago Tribune / Getty Images

McDonald’s wants AI to cut corporate costs, too

McDonald’s isn’t just applying automation to cashiers and restaurant operations. Administrative costs are also on the company’s radar.

McDonald’s now expects general and administrative expenses to be around 2.2% of systemwide sales in 2026. It wants that number to be around 1.9% by 2030, and AI is expected to help reach that goal.

“At the company, AI will help enable a step-change improvement in corporate G&A,” Borden said.

The material provided did not detail which corporate jobs could be affected or whether those savings will come at the expense of layoffs.

But the language makes it clear that executives expect AI to significantly alter the company’s cost structure.

McDonald’s has targeted an operating margin in the low-to-mid 50% range by 2030 versus 46.1% in 2025.

Some of that improvement will come from higher revenue. Some of that will come from efficiency. And management says AI will be part of both.

McDonald’s plans to offer franchisees as much as $8.5 billion through 2036 to speed up upgrades to restaurants, equipment, and technology, with about $5 billion of that support expected through 2030.

McDonald’s also plans to spend an additional $1.5 billion to $2 billion on capital expenditures to accelerate NEXT from 2027 through 2030, in addition to about $3 billion in annual capital expenditures.

In addition, individual restaurants will face major costs.

A typical U.S. drive-thru lobby remodel costs a franchisee between $400,000 and $450,000. McDonald’s estimates the new plan will cost about $800,000 per restaurant in additional technology, kitchen, and operational improvements, but the company will provide some financial assistance.

The investments are expected to generate roughly $100,000 in additional annual cash flow for the average U.S. restaurant and pay back franchisees’ investment in about four years, McDonald’s said.

That’s why automation is not a side project. It’s becoming part of the fundamental economics of running a McDonald’s.

Related: McDonald’s launches exclusive state-themed meal

Everything you need to know about this week’s UFC 332 fight card: Date, time, location, how to watch or stream and early fight week betting odds.

Steph Curry teams up with L.A.B Golf to benefit the UNDERRATED Golf Tour. Proceeds will help golfers from underprivileged backgrounds.

The first phase of the AI boom was defined by GPU horsepower and raw compute. The next phase is being driven by agentic workflows, and it is quietly shifting the bottleneck inside data centers back to general-purpose CPUs and server memory.

While investors have spent two years obsessed with Nvidia and High Bandwidth Memory (HBM), the rise of autonomous AI agents is creating an unexpected squeeze in plain-vanilla DDR5 DRAM.

On Monday, Sept. 28, Baird analyst Tristan Gerra raised his price target on Micron Technology (MU) to $1,520 from $1,280. He kept an Outperform rating and argued that agentic AI will drive Micron to new heights, CNBC reported.

The new target implies about 40% upside from the Sept. 25 close. Baird lifted its target to $1,280 from $500 in June, so its number has roughly tripled in three months. Baird is right, and Micron’s own filings show why.

Micron is the only American company among the three firms that dominate DRAM, the working memory in phones, PCs and servers.

The $1.2 trillion company could overtake Nvidia as the top driver of S&P 500 profit growth, MarketWatch reported on Sept. 27. Index fund owners already hold this bet.

Agentic AI puts the humble CPU back in charge

Agentic AI is software that plans and completes multistep tasks, such as comparing flights or writing code. While GPUs handle raw parallel processing, these sequential steps require a general-purpose CPU to coordinate the work.

Because agents must hold working memory and track their state across long workflows, they create a continuous demand for high-capacity server DRAM (DDR5) wired directly to the CPU.

Next-generation AI servers are moving from one CPU per eight GPUs toward one per four or fewer, TrendForce found in May. Each extra CPU brings its own bank of server memory.

Gerra expects AI-related CPU demand to grow about 40% in 2027, TipRanks reported. On June 24, Micron said agentic AI pushes memory demand beyond accelerator racks into CPU and storage racks.

Baird raised its Micron price target to $1,520 from $1,280 on Sept. 28, betting that agentic AI will lift server CPU and DRAM demand in 2027.vzphotos / Getty Images

Ordinary server memory now rivals HBM on margins

Micron’s fiscal third-quarter results, released June 24, show where the money is going. Its core data center unit, which sells to server makers, posted $11.5 billion in revenue, up more than sevenfold from the prior year.

That unit earned an 87% gross margin. The cloud unit, home to Micron’s HBM business, came in at 83%. Per sales dollar, the plain-memory unit already out-earns the HBM unit. Gerra expects HBM margins to climb past 80% in 2027, adding a second profit engine.

Gerra sees DRAM contract prices rising 20% in the September quarter and 10% more in the December quarter. Server DDR5, the memory beside every new CPU, could climb another 15% to 20%.

Supply is tightening, too. Gerra expects DRAM supply growth to slow to about 20% in 2027 from more than 30% this year. He also sees Chinese rival CXMT’s output growth slowing sharply from about 45% in 2026. Less new supply plus more CPUs is a squeeze that keeps prices firm.

More Micron:

Michael Burry increases his bet against popular chip giant

Bank of America doubles down on Micron stock after AI bombshell

Micron stock jumps as investors look beyond GPUs in AI chip trade

Micron stock heads into earnings near its highs

Micron closed at $1,082.28 on Friday, Sept. 25, according to stockanalysis.com. By 8:20 a.m. EST Monday, Sept. 28, shares were down 1.3% premarket at $1,068.20 as investors took profits before earnings, TipRanks reported.

The 52-week range spans $154.65 to a record $1,255 set on June 25, an eightfold swing. Yet the stock trades at about 7.3 times expected earnings, a discount that signals doubt that these profits will last.

As of Sept. 25, 36 of the 49 analysts tracked by stockanalysis.com rated Micron a Strong Buy, nine said Buy and four said Hold. None recommended selling.

The average 12-month target is $1,515, nearly identical to Baird’s. Targets span $361 to $2,200, a sign of deep disagreement.

The top target in last week’s notes was $1,625, from UBS on Sept. 23. Wells Fargo cut its target to $1,400 from $1,525 the same day, citing valuation debates, Investing.com reported.

Fiscal fourth-quarter results land Wednesday, Sept. 30, after the close. Analysts expect $51.19 billion in revenue, above the roughly $50 billion Micron guided to on June 24, TipRanks reported.

Memory is no longer the AI trade’s side bet

Micron’s own quarterly report warns that weaker HBM demand could shift supply into regular DRAM and drag prices down. Even Baird has an in-house skeptic.

Its investment strategist, Ross Mayfield, warned on CNBC in June that huge profits could push cloud giants toward alternatives.

Those risks look further out than 2027. SK Hynix CEO Kwak Noh-jung told Reuters on July 10 that 2027 will be the industry’s worst supply year ever. Intel CEO Lip-Bu Tan said on Feb. 3 that memory makers see no relief until 2028, Bloomberg reported.

The bigger shift is structural. Memory once boomed and busted with PC and phone sales. Agentic AI ties it to software that never clocks out, and buyers are responding with multi-year supply deals.

Wednesday’s guidance on gross margin and spending could drive the stock’s next move, BofA said, according to Barron’s. In the agent era, the chips that remember may matter as much as the chips that think.

Related: Bank of America doubles down on Micron stock before earnings

Shoppers in several states are being urged to check their spice cabinets after a product was recalled due to elevated lead levels.

Galil Importing Corp is recalling Lior Cinnamon Ground Seasoning because the product may contain elevated levels of lead, according to the U.S. Food and Drug Administration (FDA).

The recalled cinnamon was distributed to retail stores, grocery stores, delis, and supermarkets in New York, New Jersey, Pennsylvania, Texas, Florida, and Illinois between Nov. 18, 2025, and Sept. 7, 2026.

No illnesses have been reported in connection with the recall.

Shoppers should check recalled cinnamon

The recall affects Lior Cinnamon Ground Seasoning sold in 90-gram transparent plastic containers.

The affected product can be identified by:

UPC: 794711005484

Lot code: GAP11304

Size: 90 grams, or 3.2 ounces

The lot code can be found on the side of the container underneath the ingredient line.

More Recalls:

Popular Walmart frozen meal recalled over possible listeria

Grocery giant recalls applesauce over potentially harmful toxin

Walmart, Harris Teeter recall fruit bars over glass risk

The recall was initiated after a sample collected and analyzed by the Maryland Department of Health found elevated lead levels in the product.

Galil Importing Corp said it has begun investigating its supplier and has implemented additional testing and supplier verification measures.

Lior Cinnamon Ground recalled over lead risk.istetiana / Getty Images

Lead exposure can pose serious health risks

Short-term exposure to very low levels of lead may not cause noticeable symptoms, according to the recall notice.

In some cases, an increased blood lead level may be the only apparent sign of exposure.

The effects of lead exposure depend on several factors, including the amount of lead, duration of exposure, and a person’s age and body weight.

Children face particular risks from prolonged exposure.

If a child is exposed to enough lead over a period of weeks or months, permanent damage to the central nervous system may occur, which can lead to learning disorders, developmental problems, and other long-term health effects, according to the FDA notice.

Consumers are advised not to eat the recalled cinnamon.

Galil Importing Corp is advising shoppers to either throw the product away or return it to the place of purchase for a refund.

The company has also instructed its customers to stop distributing and selling the affected product and to place any remaining inventory on hold.

Related: Walmart, Aldi, and Kroger follow Costco’s lead

Every app on your phone has memory. Whether you save a playlist, pay a bill, or ask a chatbot a follow-up question, that information has to land somewhere it can be found again in a split second.

MongoDB Inc. (MDB) sells that memory. Its document database helps companies build applications faster, and its free Community Server has topped 500 million downloads since 2009. That reach matters because every new AI app needs a place to store its data.

On Monday, Sept. 28, 2026, the company that helps businesses hold on to everything lost the one asset it could not back up.

CEO Chirantan “CJ” Desai stepped down, effective immediately, to pursue a senior role at Meta Platforms (META), according to Reuters. He had held the job for less than a year, CNBC reported.

The market did not wait for an explanation. MongoDB shares crashed more than 26% in early trading, according to Seeking Alpha.

The opening gap erased roughly $8 billion in market value, based on CNN’s opening price and Morningstar’s share count. That is more than three times the $2.46 billion in revenue MongoDB booked last fiscal year, according to Stock Analysis.

Desai will run Meta Enterprise Platform, a new unit selling Meta’s AI models, agents, and infrastructure to businesses, according to Bloomberg. In a Meta statement, he set the goal: to make Meta “the place enterprises come to scale their businesses.”

That goal should worry MongoDB holders. The executive hired to chase enterprise AI spending for MongoDB will now chase it for a far richer company.

The board knew 4 days before investors did

The exit looked sudden from the outside. The paperwork tells a slower story. Desai told MongoDB he intended to resign on Thursday, Sept. 24, 2026, according to a securities filing. Directors named Dev Ittycheria interim CEO two days later.

Ittycheria ran MongoDB from 2014 to 2025 and grew annual revenue from about $35 million to more than $2.3 billion, according to a MongoDB press release.

Chairman Tom Killalea said Ittycheria “knows this company deeply,” and MongoDB reaffirmed its third-quarter and full-year fiscal 2027 guidance.

Ittycheria pledged to “move quickly, execute with focus,” yet the calendar works against him. MongoDB hosts its Investor Day on Tuesday, Sept. 29, 2026. Investors wanted an update to a long-term plan that still assumes high-teens growth, even though MongoDB reported 30% revenue growth last quarter.

Now an interim CEO will pitch a plan drawn up under the executive who just left. That is a hard story to sell in one afternoon.

MongoDB shares opened at $310.94 on Sept. 28, 2026, $99.50 below the prior close, after CEO CJ Desai left to run Meta’s new enterprise AI unit.Bloomberg / Getty Images

MongoDB stock now trades far below analyst targets

Before the drop, the stock traded at about 12 times sales. That leaves little room for surprises.

MDB Shares closed at $410.44 on Friday, Sept. 25, 2026.

The stock opened at $310.94 on Monday, Sept. 28, 2026, a $99.50 gap below that close.

The 52-week range runs from $215.68 to $473.10, which puts the opening price closer to the low than the high.

The average rating from 41 analysts is Buy, with a 12-month price target of $456.88, according to Stock Analysis. That target sits about 47% above the opening price.

That gap is the story. Analysts are valuing a business that posted 30% revenue growth, while the market is pricing a company without a permanent leader. Needham called the timing “unfortunate” but stressed that MongoDB is running efficiently, according to TipRanks.

The pattern is familiar. Shares fell 13.5% on Wednesday, Sept. 2, 2026, after its Atlas cloud database grew 28.9%, below investor hopes of roughly 30%, despite a revenue beat. MongoDB keeps getting punished for doubt, not for weak results.

More MongoDB:

MongoDB investors have one number to watch

MongoDB missed one number and investors punished the stock

Piper Sandler names 5 software stocks cutting AI token costs

MongoDB shareholders paid for Meta’s latest hire

MongoDB tried to lock Desai in. His November 2025 offer included $32.5 million in stock awards, more than half tied to share-price targets through 2030, according to a 2025 filing. The filing also said the second half of his $2.5 million signing bonus was due after 12 months. He left about six weeks short.

The lesson for boards is uncomfortable. A pay package built to fend off rival software companies means little when the recruiter has Meta’s balance sheet.

Meta has done this before. In 2025, it paid $14.3 billion for 49% of Scale AI and hired its CEO, Alexandr Wang, according to Fortune. Scale walked away with Meta’s money. MongoDB’s shareholders were left with the bill.

For investors, the fallout reaches beyond one database stock. As AI giants hunt for executives who can sell to corporations, a software CEO is now a recruiting target.

Key-person risk usually reads like boilerplate. At MongoDB, it just came with an $8 billion price tag.

Related: MongoDB missed one number and investors punished the stock

The company has a lucrative opportunity in letting rivals tap its data centers for computing power.