Preview: A patent cliff set off a $106 billion buying spree, and the money is spreading through medicine. Even our algorithm, which never favored medical stocks, is buying, carefully.

What you're reading: We built Luniter because we love science and want to know what's coming next, and which companies are actually building it. We track about 200 of them. Every Monday we pick the one story from that world we think is worth your time and explain it in plain English. It takes about five minutes. If you're new here, welcome.

Medicine is now the warmest corner of the market we track. On the first of May, the top theme on our 20-theme temperature map was AI infrastructure, and three of the four medical themes sat at the bottom of it. Ten weeks later they have traded places: AI infrastructure has cooled out of the top spot, and medical therapeutics reads warmer than anything else on the map, with two more medical themes in the top ten. The climb was not smooth. It came in stairsteps, with real pullbacks, including a breather this week. But the direction has not changed in two months. This issue is about what is pulling money into medicine, why our own model portfolios, which have never loved medical stocks, started buying them anyway, and how we'll know whether the rotation is real.

The largest drug companies are staring at a hole in their future revenue: by one count, best-selling drugs worth more than $170 billion in annual sales will lose patent protection by 2032. When a bestseller's patent expires, cheap copies arrive and most of that revenue evaporates. You cannot invent replacements on that schedule. So the industry is buying them. Drugmakers announced $106 billion of acquisitions across 201 deals by early June, on pace for the industry's biggest year since before the pandemic.

A buying wave that size puts a price on everything nearby. Smaller drug developers get bid up as investors handicap who might be next. The software used to discover drugs gets a validation stamp of its own: Eli Lilly agreed in April to buy Ajax, a biotech built on Schrödinger's physics-simulation software, for up to $2.3 billion, and when the deal closed in late June, Schrödinger collected roughly $57 million for its stake, with more due if the science delivers. And the makers of manufacturing gear win no matter whose drug succeeds. Repligen, which builds the filtration hardware used to produce biologic medicines, gets more pipelines to serve with every deal.

Healthcare has become the market's second trade: where money goes when it wants growth that isn't AI. One of the big biotech index funds gained about 13% in June alone. If the buying keeps up, the companies selling the tools benefit, because every acquired pipeline still needs to be designed, tested, and manufactured.

Alongside the scores, we run a simulated portfolio picked entirely by our data, published in the open (we call it the Opinionated Machinery). Medicine has never been its neighborhood. Its internal ranking list, built from head-to-head comparisons across all 200 names, has not one medical stock in its top twenty, and that list is dominated by defense names and AI hardware, the things that have worked. For a medical stock to get bought anyway, the timing signals have to be strong enough to overrule the reputation. For the past two weeks, they have been.

Between June 29 and July 8, the machine bought into medicine seven times. It started with Repligen, the manufacturing-hardware name above, and the first purchase says everything about its mood: <$1,000 of stock inside a million-dollar portfolio. Then came two frontier bets, kept deliberately tiny. Relay Therapeutics designs drugs by modeling how proteins move rather than just what they look like; it is up roughly 144% this year on real clinical results, and priced accordingly. Beam Therapeutics edits genes by rewriting a single letter of DNA at a time. On July 6 the algorithm added Zimmer Biomet, the hip-and-knee implant maker grinding through a turnaround, cheap and just starting to move, and Schrödinger, the drug-design software company mentioned in the Ajax deal. The seventh buy, on July 8, was Eli Lilly, a week after Medicare expanded access to weight-loss drugs. (The machine doesn't read the news, but market prices do.)

Starter positions, spread across very different kinds of medicine. Today the three production books hold thirteen medical names, between 8 and 11 cents of every invested dollar. The machine has no opinion about medicine's future. The scores line up right now, so it is buying, carefully.

Scorecard — Globus Medical (GMED)

The machine's strongest medical conviction is none of the names above: it is Globus Medical, which makes spinal implants and surgical navigation systems. All three production books own it, and it ranks #27 of 200 on the machine's list, higher than any other medical stock. The market disagrees so far: the stock is down about 10% this year even though the company raised its own profit guidance this spring. Lately, though, its price and its business have started improving together: a profitable, unglamorous company in the calm middle of its price range, in a medical complex that has been warming around it for two months.

Two things worth naming plainly. Washington first: a 100% tariff on imported patented drugs takes effect July 31 for the seventeen largest manufacturers and in late September for everyone else, though drugmakers that agree to sell to Americans at the lower prices they charge abroad can escape it. The rules are changing mid-rally, and companies that manufacture overseas are more exposed than the device and equipment makers. Second, our own dashboard is flagging early excess. Illumina, the DNA-sequencing giant, has been heating up in price while its value reading deteriorates, the look of a stock running ahead of its business. The rotation is real. It is not uniformly healthy.

Across the Universe: the week in numbers

What's moving

The medical themes sit in the cooling column this week, and that is the breather from the story above: therapeutics peaked on July 8 and exhaled, but it remains the warmest theme on the map by a wide margin. The week's biggest warmers were in old-economy energy, hard oil and ocean power.

One promise kept from last issue: we said we'd watch whether space's recovery outlived the deal-week glow. It didn't. Space cooled hardest of all 20 themes this week, and Rocket Lab, the sector's leader, slid from the calm middle of its price range to out of favor, down 23% in a month and 56 spots on the machine's ranking list.

From the Machine

The simulated portfolio above also had a week outside medicine. It sold two losers: AeroVironment, the military-drone maker, at a 26% loss, and MP Materials, the rare-earth miner, at a 21% loss. Both were bought about six weeks ago and cut by the same cool-down rule: when a stock sits out of favor and its timing signal stays weak for three straight trading days, it sells. On the ranking list, the biggest fader was Frequency Electronics, which makes precision timing gear for satellites, down 79 spots in a week. And our test fleet's consensus name is still ATS, an automation builder owned by 62 of the 105 experimental versions we run.

Earnings season settles the next act. Neurocrine reports July 30, Lilly on August 5, and Globus on August 6; that stretch will tell us whether the businesses justify the warming. Nearer in, we're watching whether this week's breather is a pause or a top, and whether the machine's small medical positions grow into real ones or get sold by the same rule that cut its losers this week. Either way, you'll read about it here.

Research and educational analysis only — not financial advice. Every investment carries risk.