Week of July 13, 2026 · 10 stocks
AI infrastructure runs the table while a failed heart-drug readout wipes $30 billion off AstraZeneca and Ionis
AI infrastructure dominated the feed from chips to data centers, while a single missed trial erased roughly $30 billion across AstraZeneca and Ionis in one session.
Ten genuine names cleared the trending board this week out of 1,335 tracked and more than 1.2 million mentions, and the through-line is hard to miss: AI infrastructure ran the tape. Nvidia, Dell, Penguin Solutions, Pure Storage and Meta all trended on some version of the same story — chips, racks, memory and data centers being built out as fast as capital allows. The counterweight came from pharma, where a single failed heart-drug readout wiped roughly $30 billion across AstraZeneca and Ionis in a day, and from Circle, whose new federal bank charter turned the crypto-adjacent trade into a regulatory one.
Pure Storage tops the board this week — the context feed files it under “Everpure” thanks to an entity-name mixup, but the company on the tape is the flash-storage maker riding the hyperscaler build-out. The hook is the same AI-capex wave lifting everything else on this list: enterprises stacking high-performance storage next to the GPUs.
The crowd chatter here is thin — only a couple dozen posts — so this is a data story more than a sentiment one.
The numbers carry it: Q4 revenue grew 20% year over year to $1.06B at a 70% gross margin, and the stock is up 26% year to date and 62% over the trailing year, with the sell-side modeling a walk toward roughly $4.4B in FY27 revenue. One analyst’s “priced for perfection” line captures the tension neatly — the fundamentals are clean, and the valuation already assumes they stay that way.
Meta had the loudest week on the board — more than 34,000 posts, nearly doubling day over day — and for once the noise was about shipping, not scandal. The company launched Muse Image, its first in-house generator out of Meta Superintelligence Labs, broke ground on a CAD $13B, one-gigawatt data center in Alberta (its first in Canada), and floated renting out spare GPU capacity as a business.
The crowd flipped hard bullish: talk of a “Google-style re-rating,” BofA adding Meta to its US 1 list, and the stock reclaiming $600 and its 200-day for the first time since April.
The data mostly backs the enthusiasm — ad revenue grew 33% year over year to $56B last quarter, unusual at this size, at roughly 18x forward earnings. The overhang bulls wave off is the $1.4T New Mexico teen-mental-health case headed to trial in August.
Penguin Solutions delivered the print of its public life on July 7: revenue of $479M against a $414M estimate, up 48% year over year, with adjusted EPS of $0.84 versus $0.55 expected. Management raised full-year guidance to 22% growth and $2.60 in EPS, and the AI-driven segment grew 104% year over year to make up nearly three-quarters of sales.
Positioning went aggressively long — Stifel, Rosenblatt and Needham all lifted targets to $75, and the crowd treated the NVIDIA AI Factory partner status as a moat, with calls for $100-plus.
The data supports the ramp but flags a caveat: operating cash flow ran negative on a receivables build, and CFO Nate Olmstead resigned July 8 for Trade Desk, handing the inflection to an interim finance chief. October 6’s Q4 is the next check on whether the beat pattern holds and whether those receivables convert to cash.
Circle’s week turned on a single regulatory stamp: the OCC granted final approval on July 10 for Circle National Trust, the first federally regulated national trust bank for digital assets, sending shares up double digits premarket. The charter lets Circle custody USDC reserves under national-bank supervision — the exact framework the stablecoin legislation moving through Congress is expected to require.
The conversation ran heavily bullish, pointing to USDC monthly volume hitting an all-time high of $1.21T in June and a stack of targets from Citi, Bernstein and William Blair.
The tape carries a counter-signal the bulls glossed over: CEO Jeremy Allaire sold $4M on July 6, and two more insiders sold into the pre-approval rally. Revenue grew 20% year over year to $694M last quarter, so the business is intact — but the August 11 print and the pending bill leave two dated catalysts inside six weeks.
AstraZeneca is on the board for the wrong reason: on July 9 its Phase 3 CARDIO-TTRansform trial for Wainua (eplontersen) in heart amyloidosis missed its primary endpoint, and roughly $27B of market value vanished in a session. A prespecified monotherapy subgroup showed a nominal benefit, but with most patients already on a stabilizer, the headline read as a clean miss.
The crowd turned bearish quickly, with speculation the company might bolt on a smaller biotech to backfill the lost pipeline revenue.
The base business, notably, is fine — revenue grew 12.5% year over year to $15.3B last quarter at an 80% gross margin, and consensus still walks EPS higher through FY28. This is a pipeline-premium reset, not an operating one. The July 27 Q2 print, and the full CARDIO-TTRansform dataset due at the ESC Congress in August, decide whether the compression stays contained.
Ionis wore the same trial failure far more heavily than its partner. As co-developer of eplontersen, it took the CARDIO-TTRansform miss down 17-24% and shed roughly $2.8B in market cap — the ATTR cardiomyopathy indication was the second-most-valuable use for the drug, and its loss hands that market to BridgeBio and Alnylam, both of which rallied double digits on the read-through.
The mood was bearish and getting louder, with several “shareholder investigation” filings piling on and an officer selling $1.7M into the broken tape.
There’s a genuine offset the sellers skipped past: the FDA approved TRYNGOLZA (olezarsen) for severe hypertriglyceridemia — a real commercial launch — and the antisense platform still has a full pipeline behind it. Some posters framed sub-$60 as an entry ahead of the Lp(a) readout. August 5’s print, and any Lp(a) timeline, is where the “platform-not-broken” argument gets tested.
Dell got the most unusual catalyst of the week: President Trump’s “go out and buy a Dell” endorsement at the launch of the Trump Accounts program — to which Michael and Susan Dell donated $6.25B — sent the stock up 8-10% in a session and added about $22B in market cap. Since a February shout-out, Dell is up roughly 200%.
The crowd leaned aggressively long into the breakout, eyeing $500-plus and pointing to Michael Dell echoing Jensen Huang that memory is the biggest AI bottleneck.
Under the political noise the fundamentals are real: Q1 revenue of $43.8B jumped $10.5B sequentially, and consensus sees FY27 revenue near $170B against $112B this year. But Silver Lake keeps mechanically selling into the strength, and at 124% above its 200-day, the September 3 print has to keep pace with a very stretched chart.
Nvidia drew the most raw chatter after Meta — nearly 60,000 posts — on reports that Beijing will let Alibaba, ByteDance and DeepSeek buy a capped batch of H200 chips for AI training, ending a months-long blockade. Wall Street had zeroed out China contribution entirely, so even rationed access reads as straight upside into the August 26 print.
The tone was firmly bullish — posters flagged a forward P/E near a decade low and BofA’s reiterated Buy — though Michael Burry’s disclosed $187M short handed the bears a marquee name.
The scale is what the data can’t shrink: Q1 revenue was $81.6B, and analysts model $393B in FY27. The recent $25B raise was debt, not equity, at low coupons against 74% gross margins. The risk isn’t the moat — CUDA is intact — it’s a hyperscaler capex pause, which would take the whole trade down with it.
Barnes & Noble Education is the small-cap curveball on the list — just 44 posts, but up 179% day over day in mentions on a return-to-profitability quarter that popped the stock 15-20% intraday. Management raised fiscal 2026 guidance and guided to 16% EBITDA growth in FY27, with continued deleveraging.
The bull case the crowd is running is a valuation gap: 5-6x FY27 EBITDA against comps at 11-14x, plus 30% own-brand customer penetration with room to expand.
The data shows a real turn — revenue grew 10.5% year over year to $515M and EPS is set to swing positive — but also an insider cluster too big to ignore: three insiders sold a combined $1.1M-plus in early July, straight into the turnaround. The September 3 fiscal Q1 print is the tell on whether the profitability inflection sticks.
Bitfarms is trading on a rebrand: the Bitcoin miner is now Keel Infrastructure Corp. (soon $KEEL on Nasdaq), formally pivoting from BTC mining to AI data-center hosting — the same playbook Hut 8, Applied Digital and Cipher have run. A $458M convertible note at a 1.25% coupon locked in the buildout capital at near-zero cost.
The crowd frames it as the cleanest miner-to-AI story going — one thread cites +109% from a $2.87 call and “most undervalued AI data center play,” with the bear voice essentially absent.
The catch is the last actual print was pre-pivot ugly: Q1 revenue fell 45% year over year to $37M with deeply negative free cash flow — the mining-margin story the rebrand is trying to outrun. The next earnings has to show the convertible proceeds converting power capacity into hosting revenue, ideally with a named customer attached.
The board refreshes as the conversation does — live mention velocity, crowd sentiment and the per-name data are on TickerTalks.
Want this live, not weekly?
See what's trending right now →TickerTalks is a research tool, not financial advice. This recap describes what accounts on X are discussing — it is not a recommendation to buy or sell any security. Social attention is not predictive of future price moves; investing involves risk of loss.
