01
Bottleneck to Cash
NVDA · MU · STX · GOOGL · MSFT · AMZN · BWXT
Scarcity is already becoming earnings and cash flow. That is why these belong in Core.
Cold Open
More things became real.
SpaceX grew revenue 92%. D-Wave grew bookings 1,120%. Aeva found a new path into AI data centers. Oklo moved a reactor from groundbreaking to startup authorization — and then to criticality — in under a year. Serve now has more than 2,000 robots working in American cities.
And none of that settles the investment question.
That is what changed this week.
Space
Millions
Satellites connect real customers.
Quantum
Buyers
Commercial organizations are committing capital.
Robotics
2,000+
Autonomous robots are working in cities.
Nuclear
Criticality
A real reactor moved into operation.
AI Infra
Physical
Memory, storage, networking and power are consumed in the real world.
The technology argument is getting easier.
The economic argument is getting harder.
The economic argument is getting harder.
The future is arriving. Now we have to pay for it.
Main Thesis
The proof-to-payoff gap has a burn rate.
Issue 11 was about the cost of the build. Issue 13 is about the distance between proof and payoff.
That distance is different for every company in this portfolio.
Nvidia can sell another accelerator and recognize revenue.
Seagate can ship another drive.
Micron can sell another wafer of memory.
Microsoft can finance another data center from an enormous existing cash-generating business.
Those companies have crossed an important line. They are not merely proving that the future exists. They are getting paid for it.
Then there is the next group.
SpaceX can grow revenue at extraordinary rates while spending extraordinary amounts of capital to build AI infrastructure, Starship, and a global satellite network. Its first public earnings report showed 92% revenue growth, yet the stock initially sold off hard because the size of the investment required to sustain that growth was impossible to ignore.
D-Wave can report $35.5 million of first-half bookings while recognizing only $3.08 million of quarterly revenue. Aeva can sign a development agreement aimed at a hyperscaler deployment while generating $6.1 million of quarterly revenue against a $34.6 million operating loss.
These are not fake companies. They are companies living in the proof-to-payoff gap.
And that gap has a burn rate.
Proof Is Becoming Abundant
Receipts answer only one question.
For years the hardest question in frontier investing was:
Will this work?
Increasingly, that is no longer the hardest question.
This week delivered receipts.
SpaceX
Growth
First public quarter showed extraordinary growth; stock finished the week roughly 23% higher despite the first major IPO lockup expiration.
Oklo
Physical proof
Groves received DOE startup authorization and then achieved first criticality, moving from groundbreaking to operations in under a year.
Serve
Scaling
More than 2,000 robots across 44 cities with a 99.8% delivery completion rate.
Aeva
New vector
Expanded beyond lidar with an optical-connectivity business aimed directly at the AI-data-center bottleneck.
D-Wave
Lumpy
Bookings show real organizations committing real money, but revenue conversion remains uneven.
Receipts tell us somebody wants the product. They do not tell us what the shareholder ultimately earns.
The New Hierarchy
Four economic stages.
The portfolio increasingly falls into four stages — not based on how exciting the technology is, but on how far the economics have traveled.
02
Proof to Scale
RKLB · PL · MDA · SOFI · ACHR · JOBY · ASTS
Operating proof exists. Scale still has to improve the economics. That is why these belong in Growth.
03
Proof to Payoff
SPCX · OKLO · AEVA · SERV · INTC · CENX
The story is real; the return profile is not settled. That is why these stay Speculative.
04
Optionality
QBTS
The upside can be enormous and the evidence remains incomplete. That is what a Moonshot looks like.
D-Wave · Stage Four in one quarter
$35.5M
Bookings
$3.08M revenueThe difference between demand and recognized economics.
+1,120% bookings growthReal interest, still lumpy.
−9.3% stock reactionBoth bull and bear can point to real evidence.
Thesis Board
What actually changed.
AI Compute
Firing
Nvidia demand remains exceptionally strong.
Memory & Storage
Firing
Physical bottlenecks intact.
AI Networking / Optics
Emerging
Aeva created a new angle into the build.
Power & Nuclear
Strengthening
Physical and regulatory proof accumulating.
Space
Strengthening
SPCX absorbed its first major supply shock.
eVTOL
Milestone Bound
Earnings Monday; certification remains the gate.
Robotics
Scaling
Deployment proof rising faster than economic proof.
Quantum
Mixed
Bookings real; revenue conversion lumpy.
SpaceX · The First Supply Test
The most important chart was Friday.
More than 900 million insider shares became eligible for sale after the first lockup expiration. The obvious fear was that supply would overwhelm the stock.
It did not.
SPCX rose Thursday and then surged roughly 16% Friday, finishing the week approximately 23% higher.
That deserves respect. The market absorbed far more potential supply than expected.
But it does not deserve extrapolation. Another unlock is expected later this month.
Can SpaceX turn breathtaking infrastructure spending into breathtaking free cash flow?
The one-share position remains enough. Participation without dependence.
Oklo · A Milestone Is Not a Megawatt
Execution improved. Economics remain unproven.
Oklo deserves credit. The Groves project moved from groundbreaking to DOE startup authorization — and then to first criticality — in under a year. That is meaningful evidence that the company can build, organize, staff, and move a nuclear facility through an accelerated federal process.
Precision matters.
Groves is a low-power isotope test reactor. It is not Aurora. It does not prove utility-scale electricity economics. It does not prove NRC licensing for the commercial fleet. It does not prove that Oklo can build dozens of powerhouses profitably.
What it proves is narrower — and still valuable: the company can execute.
That upgrades the thesis. It does not upgrade the position to Core.
Aeva · The Most Interesting New Bottleneck
Moving data becomes its own bottleneck.
Aeva may have quietly created the week’s most intriguing new thread.
The company launched an optical-connectivity business using its high-power optical-source technology and signed an agreement aimed at a hyperscaler deployment beginning in the second half of 2027.
This matters because AI infrastructure is not merely a GPU problem.
Compute→Memory→Bandwidth→Optics→Power efficiency
Aeva suddenly has a potential way into that chain.
Interesting? Absolutely. Core? Not remotely. Four shares remain enough.
Macro · Bad News Bought Duration
Friday delivered the old market paradox.
The economy unexpectedly lost 23,000 jobs in July. Stocks went up.
The reason was rates. Investors reduced expectations for another Federal Reserve hike, Treasury yields fell, and the Nasdaq finished its best week since April.
That is particularly helpful to this portfolio because so many frontier companies derive a large portion of their value from cash flows expected years into the future. Lower rates increase the present value of that future.
But bad economic data cannot remain bullish forever.
If weakening employment becomes weakening demand, the same companies celebrating lower discount rates will eventually face lower revenue assumptions.
Next week’s CPI and PPI matter because they determine whether the Fed actually has room to respond to weaker growth.
Portfolio Reflection
Good news does not automatically require more capital.
The portfolio structure worked this week.
C
Core
Already monetizing infrastructure.
G
Growth
Crossing from product proof into operating scale.
S
Speculative
Technological proof arrived before economic certainty.
M
Moonshot
Optionality, intentionally small.
SpaceX had good news. No add.
Oklo had good news. No add.
Aeva had good news. No add.
D-Wave had evidence of demand. No add.
Intel returned almost exactly to the original seed entry, closing Friday at $101.65. No add.
Sometimes intelligence is leaving the size alone.
Explicit Exclusion
Still no CXMT.
The company belongs in the research file. It belongs in the Micron competitive map. It belongs in the China-memory thesis. It does not yet belong in the portfolio.
The first few weeks after a scarcity-driven IPO are where curiosity should be highest and position size should be zero.
Catalyst Calendar
The proof cycle continues.
Mon · Aug 10
Rocket Lab after close — Neutron, backlog conversion, Space Systems margins, cash requirements.
Archer after close — certification, manufacturing, liquidity, path toward initial operations.
Archer after close — certification, manufacturing, liquidity, path toward initial operations.
High
Wed · Aug 12
July CPI · 8:30 a.m. Eastern.
Macro
Thu · Aug 13
July PPI · 8:30 a.m. Eastern.
Macro
Later Aug
Next major SpaceX lockup event. Then NVIDIA.
Watch
Hot Take
The frontier’s biggest risk is no longer that nothing works. A lot of it works. The risk is that investors pay for the destination before anyone calculates the fuel bill.
Capability
≠ cash flow
≠ cash flow
Booking
≠ revenue
≠ revenue
Milestone
≠ margin
≠ margin
Partnership
≠ PO
≠ PO
Working product
≠ good stock
≠ good stock
Buddy, Real Talk
This is the kind of week that can make discipline feel stupid.
SpaceX jumps. Nuclear moves. Nvidia rips. Aeva finds an AI angle. Quantum bookings explode.
Should I have more?
That is exactly when position sizing matters.
You already have exposure. You do not need maximum exposure.
The goal is not to own enough of every winner to feel brilliant. The goal is to remain solvent, curious, and positioned long enough to discover which companies actually become the winners.
16 QBTSdoes not need to become 50 because bookings sound exciting.
2 OKLOdoes not need to become 10 because a reactor milestone landed.
1 SPCXdoes not need to become 5 because the first lockup was absorbed.
4 AEVAdoes not need to become 40 because optics could become enormous.
Seasoning works because it is seasoning.
Discipline compounds. Excitement still does not.
Bottom Line
The investment question has moved forward with the technology.
The frontier had a very good week. Not because stock prices went up. Because more things became real.
AI demand became more visible. SpaceX survived its first major public-market supply test. A nuclear project moved closer to — and then into — operation. Thousands of autonomous robots are working in American cities. Quantum customers are signing contracts. A lidar company found a path into AI optical connectivity.
Those are meaningful developments.
But the investment question has moved forward with the technology.
We no longer ask only: Can they build it?
How much does it cost to build?
How much capital is required to scale?
How long until revenue arrives?
How long until revenue becomes cash?
What are we paying today for profits that may still be years away?
The future is getting easier to see. The burn rate is getting harder to ignore.
Own the bottlenecks that already get paid.Fund the frontier carefully.Let proof earn the next dollar.