4drant retweeted
UniQure $QURE announces FDA acceptance and Priority Review of BLA for AMT-130 for treatment of Huntington's disease
PDUFA May 1, 2027
"In its acceptance letter, the FDA stated that they are tentatively planning to discuss this application at an Advisory Committee meeting. The Company will be notified once the date and timelines have been finalized."
4drant retweeted
Great customer service @JustEatUK
4drant retweeted
Muse is one of those rare products that is crypto adjacent even though it doesnt know it is. It allows you to setup a monero miner which yields a few $/m.
I would wager an extremely large percentage of these downloads are from bot farms taking advantage of subsidized compute
Muse Spark is good but unusable for me
input tokens need fixing, it is 1) not cache hitting as much as it should, 2) cache hit pricing is too expensive 3) tokens dont add to $10
A workflow that costs me $0.1-0.15 on DeepSeek costs me $10 on Muse Spark
@alexandr_wang @finkd
4drant retweeted
If the model is open source, you better own the harness. $GTLB
4drant retweeted
Replying to @bindureddy
I’m convinced Google is just benchmaxxing. Their models are TERRIBLE.
4drant retweeted
$IQE: Strategic Deep Dive – More Than Just a Funding Round
The market is buzzing about the £81M raise, but the regulatory filings (RNS) and the conclusion of the Strategic Review reveal a fundamental shift in IQE's operational DNA.
Here is the objective breakdown of today’s disclosures:
1⃣MACOM: From Customer to "Anchor Platform"
This isn't a passive investment.
MACOM’s entry as a strategic partner with Board representation signals a move toward a "Captive Foundry" model.
➡️Supply Assurance:
The long-term supply agreements signed today effectively lock in IQE’s capacity for high-end optical components.
This solves IQE’s historical struggle with "lumpy" revenue and low visibility.
➡️Validation:
MACOM is a leader in semiconductor solutions for the defense and data center markets.
Their technical due diligence and subsequent equity stake - effectively "de-risks" IQE’s III-V epitaxy IP for 1.6T applications.
2⃣The Capital Structure "Hard Reset"
The £81M fundraising (Placing, Subscription, and Retail Offer) is a surgical strike on the balance sheet:
➡️Debt Eradication:
A primary use of proceeds is the significant reduction of net debt.
By de-leveraging, IQE removes the "liquidity discount" that has suppressed its valuation compared to Asian peers.
➡️Institutional Access:
By clearing the debt overhang, IQE becomes investable again for institutional funds that were previously barred by risk-compliance mandates.
3⃣Independence as a Competitive Edge
The conclusion of the Strategic Review without a sale is the boldest signal.
➡️Why it matters:
Management has officially bet that IQE’s value as a "Neutral Hub" in the supply chain exceeds its current takeover premium.
➡️The "Bottleneck" Play:
As evidenced by LandMark’s (3081.TW) capacity constraints, the world is short on MOCVD-driven epitaxy.
Remaining independent allows IQE to serve the entire ecosystem (Lumentum, Coherent, MACOM) rather than being siloed within a single competitor.
4⃣Technical Readiness: 800G to 1.6T Transition
Today’s filings emphasize IQE’s role in the next-gen AI networking shift.
➡️The InP Factor:
The documents confirm that IQE’s Indium Phosphide (InP) platforms are now optimized for the 1.6T transceiver cycle.
➡️National Security Clearance:
The deal is subject to the UK’s National Security and Investment Act, highlighting that IQE’s Cardiff-based compound semiconductor expertise is considered "Critical Infrastructure" for the AI era.
⬇️The Bottom Line:
Last week’s drop to 44p was a "solvency scare."
Today’s RNS proves solvency is no longer the question.
The question is now about execution.
With MACOM providing the floor and the AI optical bottleneck providing the ceiling, IQE has transitioned from a distressed asset to a strategic cornerstone of the Western AI supply chain.
Do you buy the dip?
$IQE: The Great Reset & The Return to Fundamentals
The last few sessions for $IQE have been a masterclass in market psychology.
After a brutal 40% crash from 72p to 44p (Wednesday–Friday), the stock surged 11% today.
Why was this "dip" a gift, and why did the market get it so wrong last week?
1⃣The Sumitomo & Lumentum Trap
The primary catalyst for the sell-off was news regarding deeper ties between Sumitomo Electric and Lumentum.
Traders panicked, fearing IQE would be "boxed out" by these giants.
➡️The Reality:
This is a classic misinterpretation. Lumentum and Sumitomo’s collaboration on 1.6T transceivers is a massive validation of the sector's scale, not an elimination of Layer 1 providers.
➡️The Logic:
These companies are Tier-1 integrators. As they scale, their demand for advanced epitaxy and InP/GaAs substrates, the very materials IQE specializes in - will skyrocket.
This is a symbiotic relationship, not a zero-sum game.
2⃣LandMark: The Signal You Can't Ignore
While London was panicking, LandMark Optoelectronics (3081.TW) in Taiwan dropped Q1 2026 results that prove the AI optical thesis:
➡️Net Profit: Up 657% YoY.
➡️Performance: They delivered 74% of their projected full-year 2025 EPS in just three months.
➡️The Bottleneck: LandMark officially signaled they are hitting capacity limits for CW lasers and Silicon Photonics. If the Asian leader is "sold out," the overflow of orders has only one logical destination: IQE.
3⃣Anatomy of the "Washout" (72p -> 44p)
A 40% drop in 72 hours without a negative company-specific update was a textbook "liquidation event" driven by stop-loss hunting and thin liquidity.
➡️Valuation Gap: At 44p, $IQE was being priced as if the Silicon Photonics revolution had stalled.
⬇️Today’s Rally:
The jump back toward 53-55p is just the start of a valuation correction.
Compared to Asian peers trading at record-high multiples, IQE remains the most undervalued infrastructure play in the AI supply chain.
➡️Summary:
The sell-off driven by the Sumitomo/Lumentum news was a strategic gift for those who understand the Layer 1 supply chain.
LandMark has proven the pie is enormous, but the epitaxy providers are few. $IQE sits at the center of this bottleneck.
$IQE still undervalued
Thanks @BobbaPaddop for the info, and @seonu63038896 for LandMark report. 👊
4drant retweeted
Yesterday's all-time high occurred with only 2.4% of S&P 500 stocks making a 52-week high.
Out of 792 days where the index made a new 52-week high since 1998, this is the fewest number of stocks participating.
4drant retweeted
Out of curiosity. Why do Europeans hate their own markets?
If you look at all my core longs:
$IQE up 837% YTD
$SIVE up 385% YTD
$ALRIB up 258% YTD
$SOI up 208% YTD
$RPI up 107% YTD
It’s just endless salt coming from local analysts and reporters.
But they’re the national security gems in each country (aside from Raspberry Pi).
Locals end up selling all their shares at the bottom, then it just transferred to American investors and institutions.
Then they don’t get any of the upside.
If a company like Riber is used by $MSFT quantum and traded at a 26 p/e, it would be $1.6B+ in the US like $LWLG. But people are salty if it has a valuation premium at all.
The only appreciative community for foreign capital I’ve seen are Japanese, and most have been incredibly welcoming.
I feel like Europeans should be proud their leading frontier companies, are benefiting from Western capital.
So they can scale up production needed for American hyperscalers?
It’s a positive sum effect on the company, and the local economies as well.
This backward mindset phenomenon needs to change.
4drant retweeted
2/ Implication: if you previously bought anything from their $20 to the $200 max plan, you now must move to a seat price just for platform access AND pay API rates for every bit of usage. The result in many cases is that the $200 max plan can easily turn into $2000/user/month (or more!) usage fees.
Here is how GitLab is different and can help:
4drant retweeted
3/ Context is king with AI: with Duo Agent Platform, agents run on the repo side and have full SDLC context. Even in our early alpha testing of the GitLab Knowledge Graph, we see dramatic improvements in SWE bench score, with fewer tool calls and fewer tokens per operation: gitlab.com/gitlab-org/rust/k…
We are currently building a massive upgrade to the knowledge graph which adds full SDLC context to the semantic code graph, more coming soon...
4drant retweeted
8/ Governance is key: having worked 30+ years in this industry I can also say that for Enterprise it won't happen without governance, audit trails, trust. That is the difference between early adopters/start-ups/hype, and real scaled adoption. The stakes are too high. This is where GitLab wins today, and why we're pursuing our opportunity in the long game.
4drant retweeted
Replying to @RoundtableSpace
Technically possible ≠ practically usable. What's the inference latency at 128k?
4drant retweeted
For current and future IQE investors, very important weekend reading from '25 which has since been confirmed by Nvidia and others will join in to create major bottleneck for InP for rest of decade. stateofthefuture.substack.co…

































