A $190 Billion Day: Viking's Drug, Meta's Agent and the Market's Overshoot

One Novelty, Three Market Shocks

The September 22 episode of Comparic Global Markets opens with host Marcin Wenus asking a single question that stitches the whole show together: what happens when one novelty reprices an entire industry in one session? The answer comes in three unrelated places on the same day — a small biotech in San Diego, a personal assistant from Menlo Park, and a housing bet from Omaha. Each touches a different market fear — durability of a drug, inertia of subscriptions, and the bottom of a housing cycle. My reading is that the episode's value is not listing headlines but connecting these one-day moves through a common mechanism of expectation versus inertia.

Start with Viking Therapeutics, where the sharpest move happened. The company released topline maintenance results for its obesity candidate VK2735 in the morning and closed up more than 35%, having traded up nearly 40% before the open. To frame the reaction, Wenus sketches the current standard: Novo Nordisk's Wegovy and Eli Lilly's Zepbound are weekly injections taken long-term, with weight typically returning after stopping. For the patient it is a weekly ritual; for the insurer a weekly bill. Viking's pitch is to thin that rhythm.

The design was two-stage. For the first five months everyone received weekly dosing and lost roughly one-sixth of body weight — about 16 to 17%. The cohort kept on weekly dosing stayed more than 20% lighter approaching eight months, with the curve still edging down. The real question came after month five: what if we cut frequency to every two weeks or once a month? The market latched onto that single question because in obesity persistence matters more than the initial drop.

The answer collapses to three numbers. On every-two-weeks dosing patients kept up to 97% of what they had lost; on once-a-month up to 90%. The group switched to placebo kept only 61% — effectively regaining about 40% of the lost weight in three months. Side effects in this maintenance window — nausea and vomiting — ran at placebo levels, which Wenus notes is the main reason people quit. Doctors increasingly flag a 75% threshold — keep three-quarters of the loss and cardiac and metabolic gains persist — and Viking's results sit clearly above it, on sparser shots. Oppenheimer's take follows: the next battle in obesity will not be who sheds the most kilos but who keeps patients on therapy; a monthly shot instead of four is an argument for patient and payer at once. The firm keeps an outperform rating and a price target around $100, roughly double the Tuesday close.

Context sharpens the picture. On Monday Novo Nordisk laid out its plan to 2030 and fell 8%, with the perception growing that it has ceded leadership to Eli Lilly. Viking's promise of sparser injections landed the next day. The obesity market is worth tens of billions annually, so any shift in force is instantly zero-sum. Wenus does not skip the honest caveat: the best dose group had 12 people, the maintenance window lasted three months and cross-trial comparisons are risky. The company is running a large Phase 3 and aims to start a tablet version this year; there is no product on shelves yet, the market priced an option on it working.

Meta's Agent and the Banks That Sold Off

The second story is a household name: Meta Platforms. On Monday it rose more than 11% and added roughly $190 billion in market value in one day — larger than most S&P 500 constituents outright. The trigger was the personal AI agent Muse released in early September. Wenus draws a crisp distinction: a chat answers questions, an agent executes tasks on your behalf — filling forms, booking appointments, writing and sending emails. The app passed 2.5 million downloads in under two weeks, outpacing ChatGPT's first 12 days. The jump was not isolated: on Monday Intel, AMD and Qualcomm each rose about 10% — the need for heavy compute behind agents revived the AI trade. Wells Fargo lifted its target toward $800, Evercore called it a genuine hit, JPMorgan wrote that Muse could become the most widely used consumer AI app since ChatGPT.

Is the rally still buyable? Wenus turns to InvestingPro and lines up three numbers: the after-hours price around $740, the analyst average target around $760, and the model-based fair value around $680 — about 7% below market. Two warnings sit alongside: a few analysts recently cut earnings forecasts and the RSI flags overbought. The message is balanced — everyone agrees Muse is good news, disagreement is how much of that good news is already in the price. Tuesday the stock faded slightly lower. The overnight catalyst is Meta Connect at 01:00 on the night into Thursday, where Mark Zuckerberg opens the conference; after an 11% jump the market will look for justification or extra fuel. Wenus leaves a small trust test for viewers: after years of debate over scam ads, would you hand this much personal authority to Meta?

The most telling part came a day later away from Meta. On Tuesday investors sold names with nothing to do with Meta: JPMorgan and Bank of America down about 3%, Wells Fargo near 4%, Charles Schwab 6%, Allstate beyond 5%, plus Progressive and travel platform Booking. The financial sector in the S&P 500 lost about 2% to its lowest since July. Bloomberg's one-word explanation: consumer inertia. Many businesses earn because customers do not bother searching for a better deal — an unused subscription kept, the same bank retained because moving accounts is a hassle, insurance renewed because comparing offers takes an evening. Wenus compresses it: what if an agent can scan your card statements and cancel what you do not use?

The cleanest litmus test was Planet Fitness — down more than 9% on Tuesday to a one-year low. A gym membership kept while unused and not cancelled is the textbook inertia case. The same pressure showed in Spotify and in carriers Verizon, AT&T, T-Mobile, all living off monthly bills. Extend the scenario and the picture sharpens: an agent that reads your statement, finds dormant subscriptions and cancels them itself, that each year compares insurance and moves the policy where it is cheaper, that books a hotel direct and skips the intermediary. A Goldman Sachs note warns — sectors living off recurring bills and human passivity could be hit hardest. Bloomberg analysts go a step further: assistants could become the new toll collectors, because the traffic that today goes straight to a store, bank or booking site would tomorrow pass through the assistant. Banks did not lose a single customer on Tuesday; the market priced the future and for the first time tallied who pays if AI starts shopping for us. Wenus flags this as a theme to watch through the autumn.

Buffett Buying the Dip and the Rest of the Tape

The third story is Berkshire Hathaway. After Warren Buffett stepped back from management and chair duties, the question was whether new leadership could deploy capital like him. The answer came in Lennar. In three sessions Berkshire bought more than $200 million of stock in America's second-largest homebuilder, lifting its holding from about 13 million at end-June to near 24 million and crossing the 10% threshold that forces prompt disclosure of every trade — which is why the news hit quickly. Timing matters: the day before the first purchase Lennar reported soft results — net income roughly halved, clearly fewer new orders, margins compressed and full-year home delivery guidance cut. Targets were trimmed to around $75, Raymond James keeping an underperform; the stock was at a one-year low, down close to 40% since December. Berkshire bought right there and Lennar rose more than 6% on Tuesday when the filing surfaced. Wenus frames it as classic Buffett — buy a good firm when everyone fears it; the US remains short of houses and mortgage rates will one day ease, the only question is when and how much margin Lennar gives up until then. The philosophy remains long-term.

A quick tour covers three opposite tails. Vicor, which supplies power delivery to AI processors, raised third-quarter guidance to roughly double the prior quarter's implied growth — four firms including large data-center operators will pay Vicor for licensing the tech that brings power to AI chips; the stock is more than twice its early-year level. On Holding rose 8% after an investor day in Zurich promised low-double-digit sales growth each year to the end of the decade plus a $1 billion buyback, yet January-to-date it remains more than a third lower after a weak year. On the downside, Quest Diagnostics fell 4% and Labcorp 3% on a different payer squeeze: the agency overseeing public health insurance for seniors, Medicare, calculated it had been paying about 16% more for lab tests than private insurers and plans to cut rates by up to 15% a year from next year, aiming to save about $1 billion annually — for labs simply less revenue from one of their largest customers. The broader tape: Nasdaq closed Tuesday at a second straight record, S&P 500 essentially flat, Dow slightly down, US crude WTI slipped below $90 for a fifth straight decline. Wenus ties the epilogue with one common denominator — 12 people in the best Viking group added more than a third to a market cap, one Meta app discounted banks and insurers, Berkshire bought what others dumped; the market reacts faster than ever to a single novelty and overshoots in both directions. What to watch next: whether Zuckerberg at Meta Connect adds fuel after the 11% jump, whether financials and subscription names retrace or extend, and whether Viking holds its gain once the first euphoria fades.