New Ask Claude or ChatGPT about any stock, using Bullpen. Bullpen in Claude or ChatGPT. 7 days free
Bullpen.News From Edwin Dorsey
Get Daily Recap Feedback

Wednesday, November 11, 2026

Back to the front page

Cat-Scratch Fervo

Fervo Energy raised $2 billion in its May IPO, then fell nearly 70% below issue price after a Morpheus short report, despite targeting 1 GW of geothermal power by 2030.

FRVO $2.3B
Doomberg 2 clicks

From X

@calvinblissett

Calvin Blissett on X

$MELI: I have done a decent amount of surface level research (reading smart people’s work) . No research I have seen is able to tell me with facts why this is Costco/Amazon scale economies shared being CHOSEN and not just competition FORCING them to cannibalize themselves in line with the innovators dilemma. The whole thesis rests on that. A thought I saw that I did like is they are building the logistics rails and in the future it will be a toll bridge that will get monetized as such for most e-commerce in regions they operate
@convequity

Convequity on X

Data centers need power delivered on schedule, not just efficient generation. $EROC builds on-site power systems around natural-gas reciprocating engines, a different technology from the turbines in short supply. These engines can supply AI data centers before grid connections or gas turbines become available. In our view, their role is broadening from temporary backup toward sustained operation as turbine shortages persist. 1. ERock has a 470MW equipment order from Anthropic. 2. It is delivering a 366MW generation project with El Paso Electric to support $META’s data-center campus, with initial commercial operation targeted for mid-2027. 3. Production commitments extend into 2028, giving ERock demand visibility while making assembly capacity and delivery execution more important. ERock sources engine assemblies from specialists, then integrates them with electrical equipment and control software. In our view, that ability to commission a complete power system quickly is the company’s main advantage. For a customer with computing equipment waiting for electricity, earlier delivery can outweigh the higher fuel efficiency of a combined-cycle plant. Delays or cost overruns would weaken that case. In our view, the next test is whether ERock can convert these commitments into operating systems on schedule and profitably.
@convequity 1 click

Convequity on X

$NVDA, $CBRS, and Huawei's Kirin 9050 Pro are three different bets on the same turning point: what the chip industry does when shrinking transistors stops paying. In AI chips, most energy goes into moving data, not into the calculation itself. In our view, as transistors shrink, calculations get cheaper much faster than the wires around them, so the industry's centre of gravity is shifting from ever-smaller transistors to shorter paths for data. That shift is why Huawei's logic folding — designing a circuit vertically to cut the distance data travels — looks potentially transformative and disruptive for the semiconductor industry to us. Here is the comparison in one line: one 16-bit floating-point calculation is the baseline, 1×, about 0.52 picojoules in NVIDIA's published GH200 energy model. Every figure below is the energy spent moving data, measured against that one calculation. 1. Moving data to local SRAM next to the computing units: Cerebras uses about 2× the baseline. NVIDIA's highly optimized local-memory path uses about 5–15× the baseline. 2. Moving data across the chip: we estimate Cerebras's core-to-core network at about 12–37× the baseline. NVIDIA's optimized path through L2, its shared on-chip cache, is about 60–90× the baseline. 3. Moving data off the chip to external memory: Cerebras's MemoryX path is about 330–800× the baseline. NVIDIA's optimized path to HBM, its external high-bandwidth memory, is about 230–290× the baseline. The energy cost is set by how far the data travels: a short hop costs single-digit multiples of one calculation, a trip across the chip costs tens of times, and a trip to external memory costs hundreds of times. If moving data is the real constraint, then shortening paths, which is what Huawei is doing with logic folding, matters more than another process node — for NVIDIA and Cerebras as much as for Huawei. These are engineering estimates, workload-dependent, not measurements of complete systems. The optimized block transfers behind NVIDIA's figures are already counted above; what remains is the physical cost of the wires and memory themselves. In our view, Cerebras's commercial case depends on whether its compiler can keep data close to computation across enough workloads. NVIDIA's advantage is a broader range of software, not a way around the energy cost of moving data.

Investor Letters

All letters

Airbnb 3Q26: Business Update

Airbnb’s 3Q26 revenue rose 19% to $4.9 billion as nights and seats grew 16%, with app bookings up 29% and Reserve Now Pay Later adding…

ABNB $111B
Speedwell Memos 2 clicks

Research update

S&P Global research will assess Platts, Kensho and ratings growth; Benefit Systems’ 72% YTD rally changed its valuation, while Netflix and…

SPGI $136BICE $96B
Global Outperformers 3 clicks
OSW · short

One Spa World

OneSpaWorld (OSW) faces about 37% downside as its valuation requires unattainable growth in average revenue per customer after major-market…

OSW $3.1B
Value Investors Club

💥Tead-ering💥

Teads’ revenue fell 26% and its stock dropped 38% as Google AI summaries cut publisher referral traffic, worsening losses and straining its…

TEAD $17MGUTS $11M
PETITION 2 clicks

The AutoNation Story

AutoNation shed rentals and megastores, focused on dealerships, and retired 91.1% of its shares through buybacks after its 1990s roll-up…

AN $5.9B
Cannibal Stocks 2 clicks

The 10-Q Fills In the Buyback

Empire State Realty Trust bought $86.1 million of stock through November 6, while corrected sum-of-the-parts values range from $9.12 to…

ESRT $1.3B
Orphan REITs 2 clicks

Publications we follow

Every publication on this site, hand-picked.

130 publications

Newsletters 110

YouTube 4

Blogs 15

Seeking Alpha 82 of 282

Nominate a publication

X accounts we follow

Every account on this site, hand-picked.

91 accounts
Nominate an account

Investor letters we follow

369 firms

Sectors

Ordered by how often our editor reaches for them.