Serenity (@aleabitoreddit)
Serenity Tracker
Unofficial, fan-built tracker — not Serenity, not affiliated. We archive his public X posts; every entry links to its source. We earn nothing from any link on this page.
Get every new call the second he posts it — live alerts in the Telegram group, free. Join free
Latest call Updated August 20, 2026
Humanoids have a benchmark. Marvell has a $12B warrant. "The keyword for AI optics in the coming years will be: shortage" · Aug 18–19

Unitree closed day one at +460% (~$48B) — the first public pure-play humanoid benchmark for his $CCXI bet. Google took a ~$12.2B warrant in $MRVL tied to TPU custom silicon. And a Taiwan laser packager's GM says the whole AI optical chain is out of stock for years. His view, not advice.

🔔 Get the next one live →
The AI build-out, as he sees it Tap a stage to see those picks
Public analyst · X: @aleabitoreddit · 1M+ followers (crossed Aug 11, 2026)

He doesn't buy Nvidia.
He buys the three obscure companies Nvidia can't build a GPU without.

Serenity is a former Reddit WSB trader and AI/photonics engineer who built a 1M+-follower following on X by mapping the AI supply chain to chokepoint nodes — obscure, vertically-integrated suppliers without which the hyperscaler AI buildout literally stops. Self-reported track record: +122% over the year to 04/2026, and +3,612% YTD as of June 11, 2026 — his own figures: leveraged, volatile, unaudited, and posted at different times; we can't reconcile the two numbers, so treat them as his claims, not verified returns. July 17, same transparency in the other direction: he disclosed a −49.4% drawdown for the month in the market-wide AI/memory/photonics crash (his post ↗), cut his leverage, and kept the thesis — and on Aug 15 posted (in Chinese, dismissing “blew up” rumors) that his YTD return still stands at +2,411.84% after the crash (his post ↗; self-reported, unaudited, same as every figure here) — leveraged-and-volatile cuts both ways, and he posts it both ways. No fund, no 13F — just public theses on X, validated by subsequent earnings and industry warnings.

Follow on X → Framework: Chokepoint Theory Sector: AI photonics + supply chain Style: concentrated, ~1.4x margin
THE CHOKEPOINT STACK what an Nvidia GPU can't ship without Nvidia GPU what everyone buys ↓ can't run without ↓ Optics & CPO $LITE $COHR CW lasers $SIVE $AAOI Epiwafers $IQE InP substrate $AXTI fewer suppliers ↓ more pricing power Own the bottleneck, not the brand. “The Strait of AXTI” · his core framework
1Y return
+122%
Self-reported · 04/2026 · unaudited
YTD 2026
+3,612%
Self-reported · 06/11 · unaudited — then a −49.4% July drawdown ↗, also self-reported
Chokepoints mapped
7
Layers of AI photonics
Names called
20+
Self-reported wins — see caveats below
The framework
"The Strait of AXTI"

Just as 20% of global oil passes through one strait, the AI photonics buildout depends on a handful of single-vendor or duopoly chokepoints. Find them, own them, hold them.

The edge
"Only buys what he's touched"

Former AI research scientist + RISC-V Foundation member + silicon-photonics engineer. Reads the substrate-level technical literature retail (and most analysts) won't.

The validation
"Frontran institutions"

Called $AXTI's InP substrate bottleneck a year before IntelliEPI's CEO publicly warned of an "indium phosphide supply shortage" in Q1 2026. Earnings have repeatedly confirmed.

The chokepoint map

Seven layers of the AI-photonics supply chain, each with a structural bottleneck. The ticker on each row is the publicly-stated chokepoint owner.
1
Raw materials — gallium, indium, arsenic
$AXTI · vertically integrated
Sourcing and refinement of the rare metals every InP wafer starts with. Concentrated in China; AXT processes its own.
2
pBN crucibles & growth equipment
Shin-Etsu Chemical · JP
Pyrolytic boron nitride crucibles that grow the InP boules. Single-vendor equipment chokepoint.
3
InP substrate processing — "the crown jewel"
$AXTI · + an unnamed monopoly/duopoly
Where the photonics supply chain pinches hardest. IntelliEPI's CEO publicly confirmed an "InP shortage" in Q1 2026 — a year after Serenity's call.
4
CW lasers — light source for Co-Packaged Optics (CPO)
$SIVE · Sivers Semi
The 2027–2028 inflection. CPO replaces pluggable optics; Sivers owns the laser-source choke at ~$290M cap.
5
Optical transceivers — assembly
$AAOI · $COHR · $LITE · Innolight
The visible end of the chain. Nvidia's $2B optical investment locked in capacity here ahead of supply shocks.
6
Testing & qualification
$AEHR · Aehr Test Systems
Photonics test equipment — "early $AAOI stage" before the volume ramp. Recently won qualification orders from a leading transceiver vendor.
7
Optical cable & fiber (HCF + traditional)
$GLW · Prysmian · Furukawa
The plumbing. Traditional fiber + new hollow-core fiber for low-latency interconnect between AI clusters.

Every name he's publicly called

Filter by category. Star the ones you want to track. Picks paraphrased from public posts — see sources.
Market:
Jargon, translated: InP = indium phosphide, the crystal wafer these lasers are built on · CPO = co-packaged optics, moving the light link right next to the chip · ELS/ELSFP = external laser source, the laser that feeds those optics · CW laser = continuous-wave (steady-beam) laser · SiPh = silicon photonics · HVM = high-volume manufacturing.

The three-phase rotation

His framework for how institutional money flows through the AI supply chain.
Phase 1 · Done
Memory / HBM
"Caught the tail end"

$SNDK, Samsung, SK Hynix, $MU. Institutions rotated in late 2025 / early 2026.

Phase 2 · Now
Optical transceivers
"Frontran institutions"

$AAOI, $AXTI, $LITE, $COHR. Many up 100–400% since first posts.

Phase 3 · Emerging
SiPh / ELS / CPO
"Adding heavily now"

$SIVE, $AAOI (as ELS player), $POET, Celestial AI. 2027–2028 CPO inflection.

The implication for retail: by the time the next phase is consensus, the asymmetric upside is gone. Watch what Phase 3 names get added to before institutions pile in.

Track record — calls that actually played out

Receipts. So you can calibrate.
$AXTI
Validated
Called the InP substrate bottleneck a year early

IntelliEPI's CEO publicly confirmed "InP shortage is a bottleneck for the entire AI infrastructure" in Q1 2026 earnings — exactly what Serenity called.

view on X ↗
$RPI
+72% over 2 sessions
Raspberry Pi demand-spike thesis

Called +55% revenue growth when sell-side consensus was +14%. Earnings printed +58%. Stock +44.76% on 3/31 and +27.43% on 4/1.

view on X ↗
$AAOI
Multi-bagger from entry
Entered at ~$84, called "way too cheap"

Full transceiver supply chain + ELSFP build for CPO. Said "10x revenue from optical transceivers in H2 2027." ~$500M cash on the balance sheet.

view on X ↗
$LNG
+52.74% YTD
"Most obvious" energy-war pick

Cheniere Energy as the cleanest US LNG export beneficiary of geopolitical disruption.

$CVX
+35.44% YTD
Energy-war supermajor

Bought as an Iran-tension hedge.

view on X ↗
$IREN
Flipped
Was long → went bearish after $6B ATM

A reminder that he updates his theses when the cap structure changes against him.

view on X ↗
Track record is self-reported and from public posts. The wins above are not a complete sample — losers and unconfirmed calls are not shown here. Calibrate accordingly.

How to follow live

This page is a snapshot. The actual stream of theses is on X.
Primary
@aleabitoreddit on X
Real-time thesis posts, replies, and chokepoint maps. Turn on notifications.
Long-form
Singularity Research Fund essay
The most thorough third-party synthesis of his framework and positions.
Performance
PhotonCap: +122% 1Y writeup
Independent breakdown of his returns and the "engineer who only buys what he's touched" angle.

Before you copy a single trade — read this

No fund, no 13F
Unlike a hedge fund manager, Serenity has no regulatory obligation to disclose holdings. What you see is what he chooses to post.
Never copy his sizing
He runs concentrated with ~1.4x margin; entry prices are sometimes disclosed, weights never. Follow ideas if you must — never sizing. His "Doomsday" percentages are hedges on his book, not a portfolio for you. July 2026 made it concrete: ~1.4x on a book that fell ~35% underneath = his self-reported −49.4% monthly drawdown.
He moves in real-time
His $IREN flip shows he changes theses when conditions change. By the time you read a tweet, he may have rotated.
Survivorship in track record
Self-reported wins; losers and stop-outs are not posted with the same visibility. Calibrate the +122% accordingly.

Sources

All picks on this page are paraphrased from public posts. Verify against primary sources:

Not affiliated with Serenity, AleaBito, or @aleabitoreddit. Not investment advice. Some picks may be out of date by the time you read them — always check his most recent posts before acting.

Free weekly digest

Get his week, in plain English.

One email a week: his newest calls, what actually changed, and the receipts — the same verified summaries you see on this page, sent to you so you don't have to keep checking back. No spam, no hype, unsubscribe anytime. Not investment advice.

We store your email only to send the digest — nothing else. Prefer real-time? The Telegram group is faster (needs a VPN inside China).