Two underappreciated $KOPN angles emerged this week that changed the math for me.
Both came from Michael Murray, one on a $BULL presentation and one in a $FABC press release.
ANGLE 1: Soldier Borne Mission Command (SBMC) is what I see as a platform-agnostic supplier play.
SBMC is a US Army program with a $1B SAM. About 240,000 displays plus night vision goggle enhancements and thermal weapon sites.
Three primes are competing for the program:
@anduriltech (with $META as display partner)
Rivet ($PLTR -backed startup)
@ElbitAmerica $ESLT (partnered with Booz Allen)
Asked directly on the call:
"Does Kopin have the potential to supply all of the primes? The answer is absolutely yes."
That single line is what changed how I see this position.
The traditional defense supplier story is binary for me. You bet on one prime, you win or lose with them.
Kopin's color MicroLED, funded by a $15.4M IBAS award directly from the Department of War, is being designed in with the Army directly, not with any individual prime. So when their tech is ready for insertion, the capability exists to plug into all three.
Three primes competing. The potential for Kopin to be the display layer underneath all of them.
That's the kind of structural setup I look for when I'm trying to find asymmetric exposure in defense.
(2/3 👇)
Both came from Michael Murray, one on a $BULL presentation and one in a $FABC press release.
ANGLE 1: Soldier Borne Mission Command (SBMC) is what I see as a platform-agnostic supplier play.
SBMC is a US Army program with a $1B SAM. About 240,000 displays plus night vision goggle enhancements and thermal weapon sites.
Three primes are competing for the program:
@anduriltech (with $META as display partner)
Rivet ($PLTR -backed startup)
@ElbitAmerica $ESLT (partnered with Booz Allen)
Asked directly on the call:
"Does Kopin have the potential to supply all of the primes? The answer is absolutely yes."
That single line is what changed how I see this position.
The traditional defense supplier story is binary for me. You bet on one prime, you win or lose with them.
Kopin's color MicroLED, funded by a $15.4M IBAS award directly from the Department of War, is being designed in with the Army directly, not with any individual prime. So when their tech is ready for insertion, the capability exists to plug into all three.
Three primes competing. The potential for Kopin to be the display layer underneath all of them.
That's the kind of structural setup I look for when I'm trying to find asymmetric exposure in defense.
(2/3 👇)
ANGLE 2: Neural I/O finally has a concrete two-year roadmap, in my view.
Until this week, the AI infrastructure leg of the $KOPN story was directional for me. Real, but vague on timelines. That changed in two specific disclosures.
$FABC press release (May 18):
"Targets demonstration of Neural I/O MicroLED interconnect platform by late 2026."
Plus: $30M in cash, sufficient to fund operations through the demonstration milestone.
@MiTknlg on $BULL (this week):
"We hope to have a prototype in production this year and move to full production by the end of 2027."
The full sequence I'm now watching:
2026: $KOPN Neural I/O prototype in production
End of 2026: $FABC Neural I/O demonstration
End of 2027: Full production
Two NDAs with leading chipmakers are formally disclosed in the $FABC press release. The two NDAs are within Murray's "Big 5" universe: $NVDA, $AMD, $INTC, Samsung, $MU.
Silverman on the market size:
"We believe the market opportunity for next-generation AI interconnects could ultimately exceed $100 billion annually."
Even 1% of that is $1B. $KOPN 's TTM revenue is roughly $45M. That math is what keeps me long.
(3/3 👇)
x.com/PSInvestor/sta…
Until this week, the AI infrastructure leg of the $KOPN story was directional for me. Real, but vague on timelines. That changed in two specific disclosures.
$FABC press release (May 18):
"Targets demonstration of Neural I/O MicroLED interconnect platform by late 2026."
Plus: $30M in cash, sufficient to fund operations through the demonstration milestone.
@MiTknlg on $BULL (this week):
"We hope to have a prototype in production this year and move to full production by the end of 2027."
The full sequence I'm now watching:
2026: $KOPN Neural I/O prototype in production
End of 2026: $FABC Neural I/O demonstration
End of 2027: Full production
Two NDAs with leading chipmakers are formally disclosed in the $FABC press release. The two NDAs are within Murray's "Big 5" universe: $NVDA, $AMD, $INTC, Samsung, $MU.
Silverman on the market size:
"We believe the market opportunity for next-generation AI interconnects could ultimately exceed $100 billion annually."
Even 1% of that is $1B. $KOPN 's TTM revenue is roughly $45M. That math is what keeps me long.
(3/3 👇)
x.com/PSInvestor/sta…
HOW I SEE THESE TWO TOGETHER
For me, SBMC is the near-term defense revenue story with platform optionality. Whichever prime wins, the potential exists for to be the display layer underneath.
Neural I/O is the medium-term AI infrastructure story with timeline visibility. Demo, prototype, and production are now on a schedule I can model.
The thesis I wrote about back was Defense + IP + NeuralDisplay + MicroDisplay portfolio. The AI interconnect leg was bonus upside that wasn't even in the deck I bought against.
What changed for me this week: both legs got specific dates and dollar numbers attached.
For six months I was asking "could this work?"
For the next two years I'll be watching "execution against a known timeline."
A few additional facts I'm keeping in my own notebook:
Cash: $68M (per @MiTknlg ). No raise needed for now.
Capacity: "We do not need more footprint. We'll be growing into the footprint that we have, and we do not need more cash to do so." No surprise capex coming.
Government AI track: $KOPN is working with "multiple four-letter agencies" to bring Neural I/O to government AI data centers, a separate customer track from hyperscalers. The Department of War is the #2 buyer of AI infrastructure.
@MiTknlg 's track record: ran the playbook at Analog Devices (Hittite acquisition), then Ultra Electronics cyber (sold for $2.5B). Not new to transformation.
That's how I see the setup right now. Anyone else watching can draw their own conclusions.
It's not a sprint, it's a marathon.
NFA. DYOR.
x.com/PSInvestor/sta…
For me, SBMC is the near-term defense revenue story with platform optionality. Whichever prime wins, the potential exists for to be the display layer underneath.
Neural I/O is the medium-term AI infrastructure story with timeline visibility. Demo, prototype, and production are now on a schedule I can model.
The thesis I wrote about back was Defense + IP + NeuralDisplay + MicroDisplay portfolio. The AI interconnect leg was bonus upside that wasn't even in the deck I bought against.
What changed for me this week: both legs got specific dates and dollar numbers attached.
For six months I was asking "could this work?"
For the next two years I'll be watching "execution against a known timeline."
A few additional facts I'm keeping in my own notebook:
Cash: $68M (per @MiTknlg ). No raise needed for now.
Capacity: "We do not need more footprint. We'll be growing into the footprint that we have, and we do not need more cash to do so." No surprise capex coming.
Government AI track: $KOPN is working with "multiple four-letter agencies" to bring Neural I/O to government AI data centers, a separate customer track from hyperscalers. The Department of War is the #2 buyer of AI infrastructure.
@MiTknlg 's track record: ran the playbook at Analog Devices (Hittite acquisition), then Ultra Electronics cyber (sold for $2.5B). Not new to transformation.
That's how I see the setup right now. Anyone else watching can draw their own conclusions.
It's not a sprint, it's a marathon.
NFA. DYOR.
x.com/PSInvestor/sta…
• • •
Missing some Tweet in this thread? You can try to
force a refresh
