in10x//theses
T3 // integration_gap//2026-08-04

The Integration Gap

Enterprise AI is hard; value flows to integrators and AI-literate services while models commoditize

today's evidence: strongly supports DOUBLE DOWN REVIEW 8 days running with the evidence behind it

the chip is what today's evidence says: green means the world moved the way this bet needs, red means it moved against it. it reads the day, not the money. the receipts quietly keep track of that.

today's readFresh reports show labs and cloud giants racing to hire scarce forward-deployed engineers, OpenAI putting $150M behind embedding them at clients, and multiple studies repeating that most AI pilots still die in messy real-world data, so outside help stays essential.

→ watch tomorrow: Watch whether the next big lab or cloud earnings call gives hard numbers on how many deployment engineers they are actually adding.

// the deal, written down first

a bet is only honest if it can lose. these two lines went down before the money did and are never edited: the first is what makes me admit the idea is wrong, the second is what makes me lean in harder.

KILL IF

Self-serve enterprise AI works; integrator bookings shrink 2 consecutive quarters

DOUBLE DOWN IF

Hyperscalers/labs keep hiring thousands of deployment engineers

// connected bets · where this one links to the rest

// the pick · the best way to own this idea today

$PLTR Palantir

The forward-deployed-engineer model at scale — enterprise AI that actually ships

PLTR still has the cleanest mechanism exposure as the AI platform that already runs on forward-deployed engineers, plus 85% revenue growth and 84% gross margins that dwarf ACN’s 6% growth and IBM’s 1%. ACN’s cheaper 11x forward P/E and strong 30-day run are nice but not enough to restart the track record after only five days.

what got picked, and when, is logged on its own on the receipts page, the misses alongside the hits.

one stock, re-picked every weekday from the data below · not investment advice

// earnings watch · what each report has to show for the bet to hold

$PLTR Palantir reported 2026-08-03 SUPPORTS THESIS

Actuals showed U.S. commercial revenue up 149% Y/Y and total revenue up 93% Y/Y, with FY26 guidance raised to 82% revenue growth and 134% U.S. commercial growth, crushing consensus. That directly matches the expectation of durable/accelerating U.S. commercial strength and guidance language supporting commercial revenue durability. Headlines are thin on AIP bookings, RDV, and FDE vs self-serve commentary, but nothing indicates lighter-touch shift or booking shrinkage, so the print supports the integrator-moat thesis.

what we said to watch for

To support the thesis that enterprise AI value accrues to forward-deployed integrators, PLTR must show continued acceleration in US commercial AIP bookings and remaining deal value, with commentary that customers still require heavy deployment engineering rather than self-serve model access. A sequential decline in commercial bookings or explicit management language that deals are shifting to lighter-touch/self-serve implementations would undercut the integrator moat and start the kill path of two consecutive booking shrinks.

US commercial AIP bookings and sequential growthTotal remaining deal value (RDV) and commercial mixCommentary on deal cycle length and FDE/deployment intensity vs self-serveCustomer count adds and expansion vs new logo mixFY guidance language on commercial revenue durability

// the companies · why these, and what their numbers say

PLTRThe forward-deployed-engineer model at scale — enterprise AI that actually ships
ACNThe world's AI integration workforce — its bookings are a read on enterprise AI demand
IBMConsulting arm + watsonx — legacy-enterprise AI modernization
company mkt cap rev growth gross margin fwd p/e fcf margin vs 52w high earnings
PLTRPalantir $301.2B +84.7% 84.1% 60× +33.6% -39.3% Aug 03
ACNAccenture $101.4B +5.6% 32.0% 11× +16.5% -43.1% Sep 24
IBMIBM $213.2B +1.1% 58.1% 17× +17.4% -32.0% Oct 21

in plain words: mkt cap what the whole company costs to buy · rev growth how much faster sales are running than a year ago · gross margin what is left of each sale before running the business · fwd p/e how many years of expected profit you pay for one share, and (t) means last year's profit where nobody forecasts next year's · fcf margin the slice of sales that ends up as spare cash · vs 52w high how far below its best price of the last year the share sits. these numbers refresh daily from Yahoo Finance.

// the last 90 days · these names against the market they trade in

0%+52%-38%05-0408-03
PLTR -13.8% ACN -6.9% IBM -0.7% QQQ +4.2%

// the prices · last close 2026-08-04

ticker close 1d 30d vs QQQ 30d
PLTRPalantir 125.91 +2.3% -2.6% -0.9%
ACNAccenture 165.72 -0.1% +22.1% +23.9%
IBMIBM 226.15 +1.1% -21.9% -20.1%

// the daily evidence · every signal scored from −2 to +2

a signal is one thing this bet depends on. the AI scores each one every weekday, from −2 (strongly against the bet) to +2 (strongly for it). what the world did counts triple, because a share price can move for any reason at all.

deployment_difficulty what the world did · counts 3× 2

news on enterprise AI rollout struggles, forward-deployed engineer hiring

integrator_bookings what the world did · counts 3× 1

earnings/bookings news for consulting+AI segments

enterprise_ai_roi what the world did · counts 3× 2

studies on enterprise AI ROI / project failure rates — evidence deployment is hard enough to need integrators (Gartner/MIT/McKinsey)

price_momentum what the price did · counts 1× 1

30d relative performance vs QQQ

strongly supports · supports · neutral · against · strongly against · one square per weekday

2026-07-222026-08-04

// the tape · this bet, day by day

08-04 Fresh reports show labs and cloud giants racing to hire scarce forward-deployed engineers, OpenAI putting $150M behind embedding them at clients, and multiple studies repeating that most AI pilots still die in messy real-world data, so outside help stays essential.
08-03 OpenAI's $150M push to embed forward-deployed engineers, Cognizant's new unit for failed agent pilots, Accenture AI revenue tripling to $2.7B, and fresh 95% pilot-failure studies all keep showing enterprises still need heavy outside help to make AI work.
07-31 Fresh OpenAI $150M DeployCo/FDE push, Cognizant EMEA unit for failed agent pilots, and multiple 95%/75% failure-rate studies keep deployment hardness and integrator demand at max support; ACN AI partnership headlines mildly lift bookings.
07-30 Fresh Cognizant/Tredence FDE launches and multiple 95% pilot-failure studies (MIT et al.) reconfirm deployment hardness; ACN hotel/AI bookings headlines and +5.2% day are mildly supportive but partly partnership-level.
07-29 Fresh Cognizant/Tredence FDE launches plus multiple 95%/80% AI pilot-failure studies reconfirm deployment hardness; ACN bookings/buyback news is mildly supportive but partly stale.
07-28 Forward-deployed engineering launches and Forbes pieces plus 80-95% AI project failure studies strongly reconfirm deployment hardness, while ACN record bookings are supportive but partly mixed/stale.
07-27 OpenAI Presence launches as non-self-serve agents with embedded engineers plus fresh 80% AI-project-collapse and 11%-goal-hit studies reinforce deployment hardness; ACN AI-demand/bookings headlines are partly stale/mixed so only mild +1.

// the headlines · the last 7 days the machine read for this bet

2026-07-27
OpenAI Presence sells enterprise AI agents with engineers attached www.artificialintelligence-news.com · deployment_difficulty
OpenAI Presence: Enterprise Agents You Rent, Not Own www.digitalapplied.com · deployment_difficulty
Streamlinefeed streamlinefeed.co.ke · enterprise_ai_roi
+1 more collected