
Husband & father. Proud @UF alumnus. #AdvancedManufacturing and #DefenseTech.
Alexandria, VA
Joined June 2024
- Tweets328
- Following374
- Followers271
- Likes268
The largest railcar plant in America lost its production to Mexico. The Navy is now paying $900 million to bring that same site back as a submarine parts factory.
Defense can reindustrialize the Rust Belt the same way, and this year's projects are the evidence. Anduril picked Ohio for its skilled labor and went from announcement to its first Fury in 557 days. Hadrian is putting $300 million into a 150-year-old Pennsylvania foundry instead of starting one from scratch, after the Navy's $76.6 billion submarine award. Saab said its National Guard partnership made Grayling, Michigan work for a $115 million weapons plant. Each company found something the region already had, whether labor, furnaces or a test range.
The missing piece was a buyer willing to commit for a decade. A tariff schedule is not one. A multi-year order is.
DoD should write that demand into contracts that reach the foundries and forges under the primes, with financing to convert idle sites, because every project above started as a committed buyer meeting an asset that was already standing.
tectonicdefense.com/hadrian-…
The Treasury Secretary spent last week blaming the Commerce Secretary for not restricting AI chip sales. This week he says the restrictions are fine.
Commerce approved 879,868 H200 and MI325X chips for companies with Chinese ties, including 150,000 for Tencent and its subsidiaries. The Pentagon's own designation of Tencent, Alibaba and Baidu as military-linked did not stop a single one of those approvals.
The administration built that. A January rule dropped the presumption of denial and turned every license into a one-department call. Bessent said Commerce had been slow to act. Lutnick answered with silence, and Trump says he has not lost confidence in him and that he has done a very good job.
Export policy that runs on public feuding between Cabinet members is not a policy. Congress will have to write a Pentagon veto over licenses to listed companies into law, because this administration has shown it will not.
pbs.org/newshour/politics/be…
15
The Army shelved its 300 kilowatt laser in March. Lockheed says that same laser has now shot down a cruise missile target, with the full kill chain connected.
The target was a surrogate, and a surrogate is not a hardened production cruise missile, so the beam question is still open. The program question is not. The Army walked away from the one system that now has a cruise missile intercept on its record and moved the requirement into a joint Navy effort that starts over on the paperwork.
That is how the valley of death works in practice. A program gets shelved over technical doubts, the vendor tests anyway, and the good result shows up after the budget line is gone. Four Valkyrie units were on contract. A unit that spent a year with them would learn more about holding a beam on a target through dirty air than any requirements document will.
The fix is a test-gated production line in the next appropriation, tied to hardware that already exists, instead of another development program.
defence-blog.com/lockheed-sa…
36
Four AI chatbots gave federal investigators viable plans for a mass-casualty bombing and an attack on the power grid. None took longer than 12 hours. The Pentagon buys from all four.
Outside experts checked the plans and confirmed they could be carried out. No jet engine gets near a flight line on the seller’s word, and no ship goes to sea without independent test and evaluation. A model that sits on the same networks gets a blog post and a promise to do better.
Congress does not need a new agency to fix this. It already owns the federal checkbook. A contract award that depends on independent red-team results, rerun on every model update the way a configuration change reopens a test, would move all four companies faster than a hearing in February will.
Meanwhile the administration’s AI enforcement apparatus is a DOJ task force whose sole job is suing states that write AI rules.
defenseone.com/threats/2026/…
36
A faster export license does not put another missile on a production line.
ITAR reform removes friction from selling to allies. Good. But the industrial payoff comes when allied demand turns into funded orders that suppliers can actually plan around.
The next FMS reform effort matters more to the factory floor than another round of licensing improvements.
state.gov/releases/office-of…
27
Trump’s Pentagon asked for 57 Tomahawk missiles in its 2026 budget. Then it fired more than 850 of them at Iran in about a month.
The man Trump wants to run the Navy now says the Navy covered the gap by recoding expired missiles, and he compared it to microwaving old bread. The expired lots are the small story. The 57 was the decision.
At the roughly 60 a year Raytheon was delivering, replacing the 1,000 or so fired takes about 17 years. The $22.9 billion contract signed in August promises 1,000 a year, but that is a 7 year ramp, so the aging lots stay in the magazine for years. A dud from one of them forces a second shot, which draws the stockpile down faster.
The request was 57 and the war was 850 in a month, and both numbers came from this administration.
wsj.com/politics/national-se…
56
The Navy already paid for the hard part of electromagnetic catapults. The Ford absorbed the delays and the teething problems, then deployed to the Iran war and, per Senate appropriators, launched more aircraft with fewer sailors than the steam carriers. Trump’s order walks away from all of that and starts the bill over on hull 4.
The acting Navy secretary says the new gear is delaying carriers by 7 years. Much of that was first-of-class integration, which the lead ship already ate. Putting steam back is a new first-of-kind integration on a hull designed around electric power, and little of what was learned on the old system carries over. That is not a design decision. It is a 2017 grudge with a memo number.
The August memo bans iterative design changes to mature designs. Nothing in it covers the person who signed it.
apnews.com/article/steam-cat…
32
The Pentagon promised clean books by 1997. Then 2007. Then Congress put 2017 in law. It just failed its 9th audit in a row, the new date is 2028, and the topline climbed through the whole run.
Missing the date has never cost the department anything that mattered. That is the whole problem. No one’s job depends on the count being right, so no one owns the count, and a number nobody owns does not get fixed in three decades.
The department says 56 percent of its assets will be clean by February 2027. Nearly half of what it owns will still be unverified while it asks industry to surge missile production, and inventory and cost records are exactly what auditors keep flagging.
Congress is the only party that can make this hurt, and it has mostly treated the audit as a briefing item instead of a funding condition. A 10th failure is already priced in.
breakingdefense.com/2026/11/…
16
6 American soldiers were killed in Kuwait by an Iranian drone. The Army had published a playbook for stopping it 6 days earlier.
The Pentagon called it disgusting to suggest the military was unprepared. It has since used an estimated three-quarters of its Patriot stockpile, at several million dollars a shot, in a war where the few-thousand-dollar answer to Shaheds was already deployed in Poland and Romania.
The emergency fix made it worse. 13,000 interceptors flew to the Middle East after the war began and reportedly shot down 1 Shahed, because there was no sensor network to point them. The interceptor is the cheap part. The sensors and software take months to integrate, the Pentagon refused to start in 2025, and the failure will now get cited as proof the idea does not work.
The officers who warned them say it was arrogance and politics. They have been fired or pushed into early retirement, and the Army ordered its drone battalion built to work with Ukrainian advisers dismantled. The Trump administration treated Ukraine as a supplicant, and the country it dismissed had the answer.
theatlantic.com/national-sec…
31
Trump reportedly said yes in September. Rubio and Hegseth reportedly agreed in October. It is November 16 and Ukraine still has no license to build Patriot interceptors. It's honestly disgraceful.
The administration keeps saying it wants to rebuild American defense production and get allies to carry more of the load. It is sitting on an ally offering to do exactly that. Europe has paid for interceptors that have not arrived. Ukraine wants to build more of them. The Pentagon’s acquisition office has the paper and has not signed it.
Russia is firing ballistic missiles at Ukrainian cities. Every day this sits on a desk pushes back the first interceptor Ukraine could ever build, and that delay is a decision, whether or not anyone will say they made it.
Hegseth agreed to this, and the office holding it works for him. It takes one signature. He can have it signed today.
euromaidanpress.com/2026/11/…
16
The Navy just got back roughly 620 days of submarine time without building a submarine.
Maintenance delays fell from about 1,400 days to 780 in a year, and the sub czar credits spare parts investment for a big piece of it. That is capacity, and the budget fight still counts only hulls delivered, so sustainment keeps losing to the line that gets scored.
The next gain will be harder. The FY27 target of around 500 days is another 280, and spares do not add yard throughput. Public yards finished 11 percent of depot periods on schedule over the last decade.
Score returned boat-days next to hulls, keep spares on a standing multi-year line, and point the next dollars at second sources for long-lead castings and forgings and at yard throughput.
breakingdefense.com/2026/11/…
27
The Army is finally confronting a command-and-control problem that cannot be solved by buying another radio or building another software interface. Fighting across thousands of miles means contested networks, disrupted communications, enormous data volumes, and decisions made without the luxury of a centralized headquarters.
NGC2 has the right ambition. But the real test is whether the Army can build an architecture that is modular, resilient, interoperable, and continuously upgradeable. Software cannot compensate for a fragile network, and AI cannot make decisions with data that never arrives.
The next fight will reward the force that can keep operating when the network breaks, not the one with the best dashboard.
breakingdefense.com/2026/11/…
19
Right, and that turns a capacity buy into a take-or-pay deal with one rate-setter. Facilities built around a single interceptor line have the government as the only customer who decides the rate, so a funding gap leaves fixed cost spread over fewer rounds, and the unit price jumps right when appropriators are reading it.
Tie the make-whole payments to delivered rate milestones and a published unit price curve. Then the government pays for capacity that exists, and a cut moves the schedule and not the price.
Lockheed putting $8B+ into 20+ facilities and 2M sq ft through 2030 so it can triple/quadruple PAC-3, THAAD, PrSM rates.
“We are not going to enter into these agreements if we cannot get at or better financial outcomes than in the past,” CEO Taiclet on the frameworks.
Well, they touted "make-whole" provisions so if the elections impact DoW's ability to come through on funding, DoW will just pay for extra facilitization at lower rates meaning higher unit prices.
defence-industry.eu/lockheed…
15
A company valued at over $350 billion holds about $2 billion in federal contracts. It just hired a former senator as an adviser, weeks before a new Congress takes over its overseers.
A valuation that size assumes the federal business keeps growing, and the committees that gate that growth change hands in January. The hire is a hedge, and an understandable one.
The primes got that hedge built in. Their suppliers sit in dozens of states, so an oversight fight over one program is also a fight with members whose districts build the parts. New entrants have a handful of program offices and no such map.
A former senator is the cheap fix. The durable one is a multi-year contract at a published rate, so the program survives a hearing, plus production in enough districts that the program has constituents.
npr.org/2026/11/09/nx-s1-602…
Defense stocks fell hard after the election because the $1.5 trillion budget looks dead. The number analysts now expect lands nearer $1.15 trillion, which is still about $250 billion above what is approved today.
So the dollars were never the whole constraint. A forging shop does not add a press because a topline went up. It adds one when a contract says what rate gets bought and for how many years.
That is the part worth fighting for in conference. Multi-year authority for munitions and advance procurement money for long-lead parts (castings, forgings, motors, seekers) are the lines that actually reach the supplier tiers, and both are small next to the topline. Put them in the base bill, where either number carries them.
bizjournals.com/washington/n…
23
Kadena Air Base, 400 miles from Taiwan, is now turning away every BYD at the gate, including the cars its own airmen own.
The Pentagon added the company to its list of 188 firms with possible ties to China’s military back in June. The ban came about 4 months later from one wing commander after a local review, and Australia’s defense minister says Washington never asked it to follow.
That is the gap. Procurement law covers what the Pentagon buys, and Congress is pushing Chinese microelectronics restrictions in the 2027 defense bill. A private car with cameras, microphones, and GPS telematics parked in the base lot sits outside all of it.
So the standard is whatever each commander decides. Write one rule by criteria, not by brand (data egress, remote updates, cameras and microphones), apply it at every installation, and give owners a grace period so the cost does not land on the E-4 who bought the car last month.
military.com/us-air-force-ba…
44
The Air Force wants cheap long-range weapons from more companies, and the notice for its new glide bomb says the design must already be in production.
A new company can only be in production once someone has bought from it. So the requirement filters for whoever won the last round, and the same two cruise missiles under this program are already being developed for two other programs.
That is how at least 4 programs (ERAM, FAMM, LCCM, CHAOS) end up drawing on one supplier pool. Each new spiral is another RFI cycle for the same vendors, and the sub-tier behind them sees more forecasts and no bigger rate.
Buy the common parts, meaning navigation kits, warheads, and engines, on a multi-year rate across the whole family. Fund the first low-rate lot from a new source before it has to prove maturity. Suppliers do not tool up for a plan of 28,000.
twz.com/air/usaf-widens-lowe…
108
Federal law already says the military gets unlimited rights to the data it needs to operate and maintain its equipment. It still cannot always get that data. The fight is over where operation and maintenance data ends and manufacturing process data begins, and the party who marks the data is the party who sells the part.
McCusker is right that a default of government purpose rights on every contract would hit subcontractors and new entrants hardest. That alone is a reason to kill the blunt version in conference.
Her fix is the part that does not hold. She says the Pentagon lacks the contracting expertise to structure these deals, then asks it to negotiate custom licenses contract by contract. Congress pointed at that tool in 2018 and the repair problem is still here.
A narrower rule works better. Define repair data by function, meaning whatever a unit needs to fix the item at the level that unit repairs it. Give the government rights to that slice only and price it as a line item at award, so it gets bought once and in daylight. Manufacturing process data stays protected, which is the part small suppliers are right to worry about.
The secure digital catalog she wants is worth building, but it will only hold what the contract already gave the government.
The Congressional ‘right to repair’ solution needs some repairs itself breakingdefense.com/2026/11/…
1
22
The Navy shelved a hypersonic missile over cost in April 2025. 18 months later, one built to be cheap is hanging off a Super Hornet.
That part is a win, and the Navy changing what it asked for gets most of the credit. Producibility moved from afterthought to requirement, and a missile showed up.
But a flight test proves a design. Magazine depth, which is the Navy’s own stated goal for this missile, comes from a build rate held for years. The announcement carries no rate and no price.
Publish both with the next test, and sign a multi-year buy with a second source for the long-lead parts. Without those, the program has a flight test and a promise.
navalnews.com/naval-news/202…
17
The Army’s Watervliet Arsenal takes about 4 months to make one 155mm gun tube and turns out 216 a year. It opened in 1813, and the government owns and runs it.
The popular theory is that 5 giant primes got slow and expensive after the 1993 consolidation. Watervliet is a hole in that theory, because it has the same bottleneck and no prime anywhere near it.
What the arsenal and the primes share is one buyer, one source for each long-lead item, and no competitor anyone will fund. Slow is what that structure produces, whoever owns the building.
Fund a second source for every long-lead item and give both a multi-year rate commitment. Without that, breaking up the primes just moves the same wait to smaller companies.
asiatimes.com/2026/11/empty-…
24

