Big announcement. Big deal. A few little asterisks.

Last week, business and political leaders gathered at the White House to announce a new $15 billion steel mill planned for Iowa that they said will support 1,750 jobs and process up to 10 million tons per year. With supply from neighboring Minnesota’s first new iron-ore mine in half a century, “the largest steel plant ever built in the United States” would represent a new day for US industry. But, on Friday, legislators in the Hawkeye State were told the mill is “not a done deal” after they were called in to rush approval for $1.4 billion worth of incentives because the company behind the project insisted on receiving more than what Iowa law allows.

A Bright Future and an Underwhelming Past

Building the new plant is Minnesota-based Mesabi Metallics, a subsidiary of Indian conglomerate Essar Group. The White House proclaimed the company’s plans to set up shop in Iowa are a sign of America’s domestic steel industry “roaring back to life” under the 50% tariffs on steel imports the Trump administration introduced last year.

Iowa Economic Development Authority Director Debi Durham told state senators Friday that she was “surprised” by President Donald Trump’s statement that the project was a “done deal” because, as she plainly put it, “this is not a done deal.” In fact, officials learned after the announcement that Mesabi would not build in the state unless Iowa lawmakers awarded it double the state’s 5% tax credit for capital projects of at least a billion dollars. Legislators were called in for a special session to approve a $1.4 billion incentive package, the largest in Iowa history, over the objections of Democrats and some Republicans. “In my area, we have John Deere, which has served Iowa and the nation and the world for over 100 years, and they don’t get a $1 billion tax break,” Dave Sires, a Republican state senator who said the subsidy shouldn’t be given to foreign firms, told WHO 13 Des Moines. Other critics worried about Mesabi’s own complicated history and megaprojects of the past that have underdelivered for the region:

• Mesabi emerged from bankruptcy in 2017, a year after failing to complete a planned iron-ore mine in Minnesota and ending up $1 billion in debt. KCCI Des Moines reported Friday that it’s still repaying the North Star State for public funds it received for that project.
• And Midwesterners haven’t forgotten, during the first Trump administration, when Wisconsin was promised a $10 billion Foxconn factory, lured with local and state tax incentives, that was to transform the town of Mount Pleasant into an economic powerhouse and bring 13,000 jobs. Only a small fraction of that, a $1 billion investment and 1,400 jobs, ever materialized.

Owners Away: The plant would not be the first major victory for a foreign firm in US steel production in recent years. Last year, Trump greenlit the takeover of US Steel by Japan’s Nippon Steel, which he opposed during the 2024 election, after negotiating a “golden share” that gave the US government a veto over major decisions like plant closures.

It’s not a cancer cure, but it’s an oncological Holy Grail in its own right.

Shares of biotech firm Grail jumped more than 11% last week after a US Food and Drug Administration advisory panel endorsed its groundbreaking blood test technology that is capable of screening for dozens of different cancers. How about that? Some potentially honest-to-God good news. One final step will be sorting through the insurance question, which is exactly as complicated as it sounds.

That Scans

Under present FDA-approved guidelines, only prostate cancer can be screened for via blood testing. Grail’s testing mechanism, dubbed the Galleri, can test for 50 different cancers from a single blood draw. The hitch? As one study of 140,000 patients showed, Galleri is much better at catching late-stage cancers (detecting 60% of Stage 4 cancers) and not so good at catching early-stage cancers (the tests detected just 14% of Stage 1 cancers). That means that Medicare and major insurers are far less likely to approve coverage of the test, at least in widespread early-use instances; the FDA must also still give a final stamp of approval, but the advisory panel endorsement is a critical first step.

Securing Medicare coverage or not dramatically changes the financial math of the medical breakthrough:

• According to Grand View Research, the multi-cancer early detection market size stands at around $1.5 billion this year and could grow to $4.6 billion by 2033. Gabelli Funds Portfolio Manager and Research Analyst Elena Meng told Reuters last week that widespread insurance coverage for the tests could create a $60 billion market in the US alone by 2030.
• Meanwhile Wolfe Research, which initiated coverage of Grail’s stock earlier this summer, says the value of the company could thus change greatly depending on which way the coverage question goes; with coverage, the stock would be worth around $150, roughly on par with its current price, without coverage, it may be worth just one-third that much.

Cautious Optimism: Grail isn’t the only cancer-focused biotech company enjoying some Wall Street overexuberance. Last week, Citi analyst Geoff Meacham downgraded Moderna’s stock to a sell rating, arguing that the roughly 200% share price surge since the company announced positive news for its melanoma vaccine back in August may have been “outsized.” Shares of Moderna slipped nearly 3% last week.

Meta’s AI bot Muse told The Daily Upside it’s not fair that Amazon only allows its own AI bot to shop its site and thought banning all bots was a lazy move. But still, Muse refused to shop on the site.

Amazon has tried to put up a hard wall around its shopping experience over the past year, most recently blocking Muse. Perplexity’s AI agent was also temporarily blocked from Amazon earlier this year. While eBay and Kohl’s also kept out bots to varying degrees, Shopify, Walmart and Best Buy have welcomed third-party agentic commerce.

While Amazon has cited security and privacy concerns as reasons to ban bots, agentic commerce products besides its own could also threaten Amazon’s lucrative ad business — a business built on browsing, not just buying.

Alexa Shops Alone

Amazon is expected to make nearly $84 billion this year from advertising sales, much of which will come from merchants paying to boost their products in the search rankings. But paying for a light gray “Sponsored” sticker won’t matter if everyone starts buying products through AI agents, instead of searching and browsing themselves. Besides, Amazon has its own AI offering:

• Amazon launched Alexa for Shopping this spring, replacing Rufus in an overhaul of its AI shopping strategy. Alexa uses a shopper’s data to personalize recommendations, compare products and so on in a separate chat box from the traditional search experience shoppers are used to.
• If shoppers used Alexa for Shopping more often, Amazon could risk cannibalizing its search-based ad business. Or it could just insert advertised products into Alexa for Shopping, too.

Waiting in Line: AI shopping hasn’t quite taken off. Take OpenAI, which used to let shoppers check out directly in ChatGPT but pivoted earlier this year to instead let retailers set up dedicated apps and recently launched new agents that could be used to shop called Dots. The switcheroo could be related to reportedly disappointing Walmart sales. If customers start to feel more comfortable letting AI agents swipe their credit cards for groceries and more, Amazon could risk losing business by blocking third-party agents. Rivals meanwhile could’ve already secured their spots in shoppers’ new checkout routines. CEO Andy Jassy said about this time last year that Amazon does expect to partner with outside agents. It could be a waiting game.