"Speed Can Reindustrialize America" - @austinvernon
Why this is in the vault
Founder shared 2026-05-22 with explicit framing: "looking for an opportunity to join in on the reshoring of American manufacturing." Anchor essay for the reshoring-interest thread. Sits alongside the Hengsperger Apr 19 piece as the second major operator-mode framing the vault has on this topic, and the two disagree on policy mechanism in productive ways (Vernon: zero tariffs + software-eats-soft-costs; Hengsperger: founder-led vertical integration of capacity + tech). Both converge on the same operational insight: the moat is digital + integration, not capex.
The core argument
Vernon's frame: US manufacturing is not broken from inability, it is structurally mismatched. American firms are competitive at high-volume production but lose badly in the "Missing Middle" of low-volume, custom, intermediate goods, where soft costs (quoting, design back-and-forth, billing, scheduling, indirect labor) "dwarf the list price" of the part itself. Eliminating those soft costs through end-to-end digitization (instant quoting, one-click ordering, integrated CAM, automated scheduling) collapses customer cost by an order of magnitude and lets US shops compete on speed - the one dimension where geography (Gravity Model: economic activity clumps near customers) gives domestic producers a structural advantage.
The corollary policy view is anti-statist: tariffs and CHIPS-style direct subsidies misallocate scarce skilled labor and constrain access to inputs, so the right government move is regulatory streamlining (faster approvals, expedited certifications, drone/logistics permitting), strategic stockpiles for rare materials, and otherwise getting out of the way. Software-driven shops will compound on their own once the soft-cost layer is eaten, and consolidation in fragmented verticals (sheet metal, CNC, casting, PCBs, injection molding) is the predictable endgame.
Where Hengsperger says "vertical integration of capacity + tech in one founder-led firm" Vernon says "horizontal software layer eating soft costs across one vertical at a time." Both routes lead to the same place - digital + speed wins - but the entry shape is different.
Key data points or specific claims
- US manufacturing is ~10% of GDP, ~$3T, second-largest globally (so the "we don't make anything" framing is wrong on volume; the gap is in the Missing Middle)
- SendCutSend bootstrapped past $100M ARR with ~$275K revenue per employee - the case study for the whole thesis
- Low-volume shops typically run at 10-20% equipment utilization; digitized shops approach 100%
- US steel production exceeds 100M tons annually (more than commonly assumed)
- Customer-side cost of a low-volume part can drop "an order of magnitude or more" when soft costs are eliminated
Frameworks or owned terms
- The Missing Middle - the low-volume / custom-parts band where US shops lose; the strategic target
- Gravity Model applied to manufacturing - economic activity concentrates near customers because transportation friction is non-trivial; this is the domestic moat
- Soft costs dwarf the list price - operating principle for digital-manufacturing pricing; the soft-cost stack is where the order-of-magnitude wins live
- Order of operations for automation - software first (soft costs), then production integration, then machine utilization, then robotic arms last (only when justified); inverts the "robots = automation" default
Companies / sectors named
For future cross-reference against /investing-smart-money-watch 13F filings:
- Digital low-volume platforms (the case-study cohort): SendCutSend, Osh Cut, Forge Automation, NOX Metals, Blitz Panel, Digital Metal, Obie Industries, ASAP PCB
- Manufacturing verticals to scan: sheet metal, CNC machining, casting, welding, injection molding, PCBs, electrical panels
- Downstream industries pulling demand: aerospace, defense, capital goods, automotive
Most of these are private and bootstrapped (SendCutSend explicitly so) so they won't appear in 13Fs directly. Watch instead for: public capital-equipment vendors (Hurco, Haas via parent, Trumpf-adjacents), industrial software (Autodesk Fusion-Manufacture stack, PTC, Hexagon), and any IPO motion from the named cohort.
Mapping against Ray Data Co
Existing surfaces this attaches to
01-projects/digital-manufacturing-discovery/- already on the board as a discovery surface, currently scoped to CAD/3D/CNC YouTube channel learning (CNC Kitchen, Teaching Tech, Product Design Online). Vernon's essay reframes that project from "founder hobbyist learning track" to "scouting the operator stack of a viable Missing-Middle entrant." The Fusion-360 day-12 track ([[2026-05-15-product-design-online-fusion-day-12-screwdriver]]) is the right shape of skill build for this; the missing piece is the soft-cost-software layer Vernon highlights.01-projects/data-marketplace/manufacturing-cell-research- the CNC + 3D-printing cell research in Tampa Bay maps directly to Vernon's "low-volume digitized shop" pattern. That project was framed as custom furniture; Vernon's thesis suggests the more lucrative shape is intermediate-goods (sheet metal, PCBs, electrical panels) where buyers are aerospace/defense/capital-goods rather than retail consumers - higher AOV, less marketing-dependent, more digital-quoting-friendly.- [[2026-04-19-hengsperger-reindustrialize-america]] - the contrasting frame. Hengsperger says founder-led vertical integration of capacity + tech; Vernon says horizontal soft-cost software layer eating one vertical at a time. Worth holding both as live hypotheses, not picking one yet.
Shape options for "joining the reshoring"
Four distinct shapes the founder's interest could take, not mutually exclusive:
- Investment thesis - long industrial-software + capital-equipment vendors; track for SendCutSend-cohort IPO motion; cross-check against existing power/memory/hyperscaler-capex theses (manufacturing reshoring is downstream of grid + chip capacity, which the existing theses already cover - so this would extend the bench rather than replace it).
- Operator/employer pivot - join an existing Missing-Middle digital shop in an ops/data/software role. The cohort Vernon names is mostly private, mostly bootstrapped, US-distributed. Aligns with the phData W2 base + Snowflake/Anthropic cert escalator path the founder is already on.
- Advisory / consulting - help one existing shop build the soft-cost software layer. Lower commitment than employer pivot, higher learning surface than passive investment.
- Content niche for Sanity Check - the Missing Middle is a sharp, owned-term frame that the newsletter could occupy. Not a derivative restate of Vernon (the no-derivative-pieces rule applies), but Vernon's framework + RDCO's harness-thesis lens could produce original pieces on "why the soft-cost layer is AI-eats-software's first physical-economy beachhead." Pattern-match against the existing harness-thesis cluster.
Cross-references to active investing theses
- Power thesis - manufacturing reshoring needs grid + on-site power; Vernon flags this implicitly (Hengsperger flagged it explicitly in concept #2, modern metal mills with next-gen nuclear). Reshoring is bullish for grid-buildout names already in the power bench.
- Memory + hyperscaler-capex theses - less direct. Manufacturing reshoring intersects semiconductor packaging and fab tooling (per the Karpathy/Anthropic transcript line about packaging being as important as fab itself) but that is the high-volume end, not Vernon's Missing Middle. Adjacent, not overlapping.
- No conflict with existing theses; this would be a new bench candidate focused on industrial software + digital low-volume shops, with the power thesis as the closest neighbor.
Open questions for founder
- Shape priority - investment thesis vs. employer pivot vs. advisory vs. content niche? These have very different time commitments and very different success metrics. Picking one focuses the work; holding all four risks the un-anchored-shiny-object pattern.
- Time horizon - is this a near-term play (act inside 12 months while phData job is fresh and cert escalator is open) or a 5-year orientation (background reading + Sanity Check coverage + slow investment thesis build)? The answer reroutes everything else.
- Hengsperger vs. Vernon - do you want to hold both frames as live hypotheses for now, or do you already lean toward one (vertical-integration founder play vs. horizontal soft-cost software layer)? The two suggest different scouting targets.
Related
- [[2026-04-19-hengsperger-reindustrialize-america]] - the companion reindustrialization piece; contrasting integration vs. soft-cost-software framings
- [[../01-projects/digital-manufacturing-discovery/2026-04-19-channel-shortlist]] - existing discovery surface this essay reframes
- [[../01-projects/data-marketplace/manufacturing-cell-research]] - the Tampa-Bay CNC + 3D-printing cell, candidate to repoint at intermediate goods per Vernon's frame
- [[2026-05-15-product-design-online-fusion-day-12-screwdriver]] - operator-skill build track that feeds this surface
- [[../01-projects/foundational-knowledge-discovery/2026-04-19-mooc-shortlist]] - sibling discovery surface (foundational learning track)